REPORT RESUMES

ED 003 818

24

DEDUCTIVE REASONING IN ADOLESCNCE--CRITICAL THINKING

READINESS IN GRADES 1-12, PHASE 1.
BY- ENNIS, ROBERT H.
PAULUS, DIETER
STATE UNIV. OF N.Y.. ITHACA
REPORT NUMBER CRP-1680
PUB DATE
314P.
EDRS PRICE MF-$0.45 HC-$12.56

65

DESCRIPTORS- *CRITICAL THINKING, THOUGHT PROCESSES. *LOGICAL

THINKING, COGNITIVE PROCESSES, ADOLESCENTS, *COGNITIVE
DEVELOPMENT, ITHACA, NEW YORK
THIS STUDY DEALT WITH THE CONCEPT OF DEDUCTIVE LOGIC AND
ITS USE BY ADOLESCENTS, AGES 10-18. TWO EMPIRICAL QUESTIONS
OF CONCERN IN THIS STUDY WERE (1) THAT OF READINESS FOR
MASTERY OF LOGIC AND (2) THAT OF THE NATURAL-CULTURAL
DEVELOPMENT OF MASTERY OF LOGIC. TWO DEDUCTION TESTS WERE
DEVELOPED, ONE FOR EACH OF THE TYPES OF LOGIC TREATED. THEY
WERE THE CORNELL CONDITIONAL-REASONING TEST, FORM X, AND THE
CORNELL CLASS-REASONING TEST, FORM X. THE TESTING OF 803
SUBJECTS LED TO (1) THE SPECIFICATION OF A SET OF PRINCIPLES
OF TWO MAJOR TYPES OF LOGIC, CONDITIONAL LOGIC AND CLASS
LOGIC. (i) TWO LOGIC TESTS, (3) A THEORY OF OPERATIONAL
DEFINITIONS, AND (4) AN ANALYSIS OF PIAGET'S CONCEPTION OF
LOGIC. (GD)

.U. i. DEPARTMENT

OF HEALTH, EDUCATION AND WELFARE

OM* of Education

pylon
This document
or organ :Anionhas been reproduced exactly as received from the
Noted do not me pertly

originating N. Pokits of view or opinions

represent official Office of Education

PosWon or policy.

CRITICAL THINKING READINESS IN GRADES 1-12j

---afTRA/.-11DUCTIVE

REASONING IN ADOLESCENCE)

COOPERA'xiVE RESEARCH PROJECT NO. OE 1680

Robert H. Ennis, Principal Investigator

and
Dieter Paulus

Nev York State College o

Wiculture at Cornell University and

the School of Education, Cornell University
Ithaca, N. Y.

1965

The research reported herein was supported by the Cooperative Research

Program of the Office of Education, U. S. Department of Health, Education,
and Welfare.

0-1)3213

ACKNOWLEDGEMENTS

We are indebted to the school systems of Ithaca,

Newfield, New York, for their help and cooperation in
People from these systems who were particularly helpful are
Mr. Leonard Backer, Miss Dorothy Balbirnie, Mr. William Beischer, Mrs. Virginia
Bristow, Mr. Raymond Buckley, Dr. Leonard Buyse, Mrs.
Rhoda Curtis, Mr. Warren Currier, Miss Marion Davenport, Dr. John
Graves, Mrs. Florence Harley, Mr. Dotglas Hart, Mr. Robert Holgate
Mr. Thornton Jesdale, Mr. Robert Keilog, Dr. James Mason, Mrs. Martha Nelson, Mrs.
Margaret Paetow, Mrs. Theresa Parakh, Miss Natalie Payst,

Mk. Curtis Rohm,Mr. Philip Skeps, and Mrs. Ruth York.

The help given by the following faculty members

is very much appreciated: Professors Max Black,

Darlington, Keith Donnellan, Mauritz Johnson, Jason Millman, and

Frederick Stutz.

Project staff members, who gave far more than required of them,

are Professor William Gardiner, Mr. John Guzzetta, Mrs.

Mrs. Carmen Mix, Mr. Richard Morrow, Professor Lucille angel and

Miss Margaret Smith.

Our debt to the Cooperative Iesearch Branch of the U.S. Office

of Education is great.

From this group we received support and encouragement

without the limitations and specifications that are

sometimes found to accompany Federal

for Which support was available

grants.

The kind of research

was not specified in advance, and

changes made necessary by discoveries as the research proceeded were

accepted without comment.

RHE and DP

ii

TABLE OF CONTENTS
Acknowledgements

1211t.

ii

List of Tables

vii

List of Graphs
Chapter No.
I

Introduction

I-1

II

The Subject Matter: Logic

II-1

III

A. An Analysis of Critical Thinking

B. Types of Deductive ;ogle
1. Sentence Logic
2. Class Logic
3. Ordinal Logic
4. Other Types of Logic
C. An Elaboration of The Types of Logic
Investigated In This Study.
D. A Comparison with Piaget's Logic
1. Points of Correspondence
2. Points of Noncorrespondence
Summary
E. Chapter Summary
Basic Deta on Subjects

-1
-1
-3
-3
-4
-4
-5
-5
-7
-7

IV

A. The Groups with Which We Worked

1. The "LDT's"
2. The "LNDT-1's"
3. The LNDT-2's
B. The Sources of Our Deta
1. Readily Obtainable Information
2. Less Readily Obtainable Information
a. Socio-Economic Status
b. The Cornell Deduction Tests
C. The Figures
The Cornell Deduction Tests
A. The Structure of the Tests
1. The Logical Fbrm of the Items
2. CouWst Components
a. Concrete Familiar
b. Symbolic
c. Unfamiliar
d. Suggestive
3. Item lbrmat
4. Item Arrangement
B. Operational Definitions of 'Mastery of
X Principle'
C. Reliability
D. Validity
1. Content Validity

-2
-3
-8
-8
-8
-8
-9

iii

-1
-3
-4
-5
-6
-6

-7
-15
-16
-18
-26
-27
III-1

:1?
-19
-25
.25

iv
Chapter No.

2. Construct Validity
IV

a. Correlations With Familiar Measures

b. Item Analysis
c. Making Sense in the Context of the
Study
E. Suggested Further Development
F. Chapter Summary
The Natural-Cultural Development of Knowledge
of Logic
A. The Literature
1. Piaget's Formal Operations
a. The Real Versus the Possible
b. Kinds of Formal Thought
c. Stages
d. Questions Posed by Piaget's Work
2. Other Studies
a. Is there a development of logical
ability as children grow older?
b. Does the development come in stages?
c. Is conditional logic mastered by age
11-12?
d. Is class logic mastered by age 11-12?
e. Is the truth-validity characteristic

13211.

-26
-26
-29
-33
-33
-36

V-1
-1
-2
-6
-6
-8
-10
-10
-10
-11
-11
-11

by age ]1 -12?

f. Within each type of logic, is there a

development of one sort of thing
before another and are some things
more difficult than others at a given
level? If so what is the nature of
this differential development?
Summary

-12

48

G
g
The
......
Is there a development of logical
ability as children grow older?
2. Does the development come in stages?
3. Is conditional logic mastered by age
11-12?
4. Is class logic mastered by age 11-12?
5. Is the truth-valid:sty characteristic
achieved by age 11-12?
6. Within each type of logic, is there a
development of one sort of thing before
another? If so, what is the nature of
this differential development?
a. Differential Development of Knowledge
of Principles of logic.
1. The Fallacies
2., The Converse
3. The Contrapositive
4. Transitivity
5. The Comparability of the Two Types
of Logic.

-13
-14
-14

-17
-22
-25
-28
-29

-29
-29
-32
-33
-35
-36

Chapter No.

VI

Page,

b. Differential Development of the

Three Components: Concrete Familiar,
Symbolic, and Suggestive.
C. Chapter Summary
1. The Literature
2. Findings
3. Needed Research
The Development of Readiness to Master Logic

-37

-41
-41
-42
-43

A. What Is Readiness?
VI-1
-1
1. Capacity and Willingness
-2
2. Dispositional Terms
-3
3. A Double Disposition
-3
4. Elusiveness of Readiness
5. A Substitute for the Identical Person
-4
-6
6. Suitable Conditions
-8
7. Summary
B. The Literature
-10
C. Procedures
1. Assignment of Grade Levels to Class
-10
and Conditional Logic; The Pre-Test.
-10
2. Total Teaching Effort
-11
3. The Actual Teaching
-14
4. The Post-Test
D. Results
-14
-15
1. General Qualifications
-18
2. Whet Did Our Subjects Learn?
a. Conditional Reasoning
-18
1. Total Scores
-18
-19
2. Component Scores
-21
3. Item Group Scores
-24
4. &mar/ and Possible Import for
Teaching
-25
5. Learning Words vs. Learning Logic.
-26
b. Class Reasoning
1. Total Scores
-27
2. Component Scores
-27
-28
3. Item Group Scores
-29
4. Summary
c. Possible Explanations of the Difference -29
Between Conditional and Class Reasoning.
-30
3. Can We State What Others Are Ready To
Master?
-31
a. An &ample of the Use of a Multiple
Regression Equation for Making a
Prediction.
b. Making Use of the Generated Predictions -32
-37
c. Summary and Overview
E. FUrther Research
-37
F. Chapter Summary
-38

vi

VII

VIII

The Past
A. The
B. The
C. The

and Future of the Project

Readiness Study: The Past.
Tests
Readiness Study: The Future

VII-1

-6

Summary
A. Introduction: The Statement of the Problems VIII-1
-2
B. The Nature of the Subject-Matter: Logic.

1. Two Types of Logic

2. Principles of Logic
C. Basic Data on Subjects

-5

-8
-9

D. The Cornell Deduction Tests

1. Reliability
2. Validity

-11
-11

The-Natural Cultural.Development of Knowledge -15

of logic..
The Development of Readiness to Master Logic-18

1. What Our Subjects Were Ready to Master

2. That Which Fairly Comparable Students
Are Reedy to Master
G. /killed Further Research
H. Overview
References
Cited References
General References

-19
-21
-23

Roa
R-3

Appendix
A.

Tables A1-14 and Graphs A1-4

B.

The Cornell Deduction Tests

"The Cornell Conditional Reasoning Test, Form
"The Cornell Class Reasoning Test, Form X"

C.

Trial Answer Sheet for tower Elementary Students

Samples of Teaching Materials

Class Reasoning Exercises Used in 10th and 12th C-1
Grades, Consisting of Three Used Early in the
Instructional Period and OneUsed at the End.

Conditional Reasoning Exercises Used in 11th

Grade, Consisting of Three Used Early it the
Instructional Period and One Used at the End.
A Class Reasoning Exercise Used Toward the
End, of the Instructional Period in 4th and
6th Grades.
A Conditional Reasoning Exercise Used Toward
the End of the Instructional Period in 5th Grade.

C-8

C-15

C-17

vii
LIST OF TABLES

II-1

Illustrated Basic Principles of Conditional Logic

II-10011.

11-2

Illustrated Basic Principles of Class Logic

II-12, 13.

III-1

Percentage Occupational Make-Up of City, New

York State and the United States.

111-2

111-3

Basic Information on Ouil Subjects

III-80 9.

IV-1

Logical Fbrm of and Answers to Items in "The

Cornell Conditional Reasoning Test, Form X:"

IV-4, 5.

IV-2

Logical Fbrm of and Answers to Items in "The

e0,110 Class-Reasoning Test, Form X"

IV-6, 7.

IV-3

Test-Retest Reliability Estimates on

the Conditional Reasoning Test.

IV-20

IV-4

Test-Retest Reliability Estimates on the Class

Reasoning Test.

IV-21

IV-5

Standard Errors of Item Group Scores-Conditional

Reasoning

IV-22

IV-6

Standard Errors of Item Group Scores-Class

Reasoning

IV-23

IV-7

Correlations Between Total Scores on the Cornell

IV-28
Deduction Test and Certain Familiar Variables:
Chronological Age, I.Q., Socio- Economic Status,
and Sex.

IV -8

Mean Difficulty and Discrimination Indices For The

Class Reasoning Test, Based Upon Pre-Test
of LDT's, LNDT-1's, and LNDT-2's Combined.

IV-31

V-1

Mean Chronological Age; I.Q.; Estimated Mental

V-16

Age; and Total Component, and Item Group

Conditional and Class Reasoning Pre-Test
Scores; by Grade for all Subjects Grouped
Together.
V-2

Percent Meeting the Sufficient Condition and the

Percent Failing to Meet the Necessary Condition
for Mastery of Each Principle at Each Grade
Level on the Conditional Reasoning Test.

V-18

viii

ELIE
V-3

Percent Meeting the Sufficient Condition and

V-19

the Percent Failing to Meet the Necessary

Condition for Mastery of Each

Each Grade Level

Principle at

on the Class Reasoning Test.

Difference Between Lowest Grade and Highest

V-31

Grade Mean Item Group Difficulty Indices;

Item Group Answers.
Comparisons of Mean Difficulty Indices of

V-38

V-5
Three Components: Concrete Familiar,

Symbolic,

and Suggestive.
Rough it
Mitre Estimates of the Effort

VI-13

VI-1
Devoted to the Teaching of Each Principle

At Each Grade Level.

VI-2

Statistically Significant Post-Test Differences

VI-20

Between Students Who Were Taught Logic and

Students Who Were Not Taught Logic, Using

Analysis of Covariance with Pre-Test and

I.Q. as Covariates.
Conditional Reasoning Pre- and Post Differences

VI-22

for 11th Graders to Whom Logic Was Deliberately

Taught

VI-4

Conditional Reasoning Pre-and Post- Test Differences VI-23

for Baniciple #1, the
Affirmation-of-the-

Antecedent Principle:

VI-5

VII-1

A-1

Multiple Correlations, Cross

Validation Estimates, VI-33
and Standard Errors for

A Tentative Time Table

Prediction Equations.

for Fixture Phases of the

Critical Thinking Readiness Study.

Individual Item and Item-Group

Mean Difficulty Appendix

Indices by Grades on the Class Reasoning Test

for All Subjects on the Pre-Test and also for

LDT Pre- and Post=Tests.

A-2

Individual Item and Item-Group Mean Difficulty

Indices by Grades
Test
on the Conditional Reasoning
Fbr All Subjects
on the Pre- Test and also for LDT

Pre- and Post-Tests.

VII-7

ix
A-3

Individual Item and Item-Group Mean Discrimination

Page
Appendix

Indices by Grades on The Conditional Reasoning

Test For all Subjects on the Pre-Test.
Individual Item and Item-Group Mean Discrimination

II

A-4
Indices by Grades on the Class Reasoning Test
for All Subjects on the Pre-Test.
Conditional Reasoning Adjusted-Mean Post-Test

It

A-5
Comparisons of Students to Whom Logic Was
Deliberately Taught and Students to Whom Logic
Was Not Deliberately Taught.
Class Reasoning Adjusted-Mean Post-Test Comparisons

"

A-6
of Students to Whom Logic Was Deliberately Taught
and Students to Whom Logic Was Not Deliberately
Taught.

A-7

Pre- and Post-Test Mean Difficulty Indices for

Students to Whom Conditional Reasoning Was
Deliberately Taught, by Item Groups and Grades.

A-8

Pre- and Post-Test

II

Percentages of LDT's Who

Met the Sufficient Condition for Mastery, and

Percentages Who Failed to Wet the Necessary
Condition for Mastery of Principles of Conditional
Reasoning, By Item Group and Grade.
A-9

Pre- and Post-Test Mean Difficult Indices for Students "

to Whom Class Reasoning Was Deliberately Taught,
by Item Groups and Grades.

II

A-10

Pre- and Post-Test Percentages of Students Who Were

Deliberately Taught Class Reasoning Who Met the
Sufficient Condition For Mastery and Percentages
Who Failed to Meet the Necessary Condition, by

Item Group and Grade.

A-11

A-12

Conditional Reasoning Weights and Constants for

Multiple Regression Equations Making Use of PreTests.

A-14

II

Conditional Reasoning Weights and Constants for

Multiple Regression Equations Not Using PreTest Scores.

A-13

II

Class Reasoning Weights and Constants for Multiple

It

Regression Equations Making Use of Pre-Test.

Class Reasoning Weights and Constants for Multiple
Regression Equations Not Using Pre-Test Scores.

It

V-1.

LIST OF GRAPHS
Conditional and Class Reasoning Mean Total Score
Page
V421

V-2.

vs. Mean Chronological Age.

Conditional and Class Reasoning Mean Total Score

V-23
VI-1

vs. Mean Mental Age.

Empirically Derived Relationships Between Mean

VI-35
Difficulty Indices and the Percentages
Meeting the Sufficient Condition and Failing
to Meet the Necessary Condition.
A-1 Plotted Points: Conditional Reasoning Mean Difficulty
Index vs. Percentage Meeting the Sufficient
Condition

A-2 Plotted Points: Conditional Reasoning Mean Difficulty

Index vs. Percentage Failing to Meet the

Appendix

ti

Necessary Condition.

A-3 Plotted Points: Class Reasoning Mean Difficulty Index

vs. Percentage Meeting the Sufficient
Condition.

A-4 Plotted Points: Class Reasoning Mean Difficulty

Index vs. Percentage Failing to Meet the

Necessary Condition.

tt

tt

Chapter I - Introduction

Many curriculum reforms have in the history of our schools been

accompanied by confusion about just what they were supposed to accomplish
and by ignorance about the feasibility of trying to accomplish whatever it
was they were doing at the level at which they were trying to do it.
In instituting the Critical Thinking Readiness Project it was our hope
to do what we could to help avoid these two dangers in the attempts to
introduce instruction in critical thinking. Our primary goals in this
project are to contribute to knowledge about what critical thinking is and
to knowledge about when it can be taught. This report, covering the first
phase in our attempts to achieve these goals, deals with deductive logic and adolescents (age 10-18).
A more specific list of the purposes of this phase follows. Some of
these purposes are subsidiary and some are complementary to the abovemetioned pair of goals, as limited for this first phase.

TO become more clear about the nature of deductive logic, as

used in ordinary reasoning and to compare this analysis
Piaget's to see whether we were talking about the same thing.
students

are

ready

to

master

with

To find out when

deductive logic.

principles

To be clear about the concept, readiness to master a principle.

To build logic tests suitable for use in this study.

To provide an operational interpretatieri ofimastery of a principle'

TO investigate the degree of mastery of principles

logic currently found among students of ages 10-18.

of

of

In this report there are no definitive answers to the questions implied

by those purposes, but two general principles have clearly emerged as a
result of our efforts:

In this area of study there are many conceptual pitfalls for

the unwai

1-2

Answers to the questions implied by the above list of purposes are of

the 'that depends' form. For example, the answer to the question
about when students are ready to learn deductive logic depends not
only upon the type of logic, but also upon the principle of logic
within that type. In his developmental studies of knowledge of logic,
Piaget distinguished only among types of logic, making rather broad
statements about these types.

This report of our efforts to achieve these purposes is divided into

eight chapters.

Each chapter is fairly self-contained, but is best under-

stcod in conjunction with the others. Probably the best order of reading
is to start with this, the introductory chapter; then to read the last, the

summary chapter; and then to read the other chapters in order (Chapters

II to VII).

The middle chapters, which make

up the body of the report,

are in a plausible sequential order. Here is a list of the questions dealt

with in Chapters II through VII, and a list of the corresponding chapter

headings:

Questions

What is to be found in the report?

What is the nature of the logic content
we were trying to convey and how does
this compare with Piaget's notion of logic?

IV.

Who were our subjects and how did we secure

information about them?

What is the nature of the tests used to

measure knowledge of logic, and how did
ve operationally interpret 'mastery of a
principle of logic'?

What sort of development of knowledge of

logic occurs without deliberate instruction in
school, that is, as a result of what we
have called "natural-cultural influences"? How
does this development very for different
principles of logic, and how does it very with
the content of the propositions used in
reasoning?

Chapter Headings
INTRODUCTION
THE SUBJECT MATTER:
LOGIC

BASIC DATA ON SUBJECTS

THE CORNELL DEDUCTION

TESTS

THE NATURAL-CUIMMAL
DEVELOPMENT OF
KNOWLEDGE OF LOGIC

I-3
Questions
VI.

Chapter Headings

What principles of logic were our subjects

ready to master, given 15 daily periods of
instruction in logic? And what sorts of
content were they ready to reason with? What
are other students ready to master, given
comparable instruction? What is the meaning
of 'readiness for mastery of a principle'?

THE DEVELOPMENT OF
READINESS TO
MASTER LOGIC

VII. What happened in

prior years that led up to

the work directly reported on here. What is
next?

VIII. Briefly what

THE PAST AND FUTURE

PLANS OF THE PROJECT.

appears in this report?

SUMMARY

In organizing this report

statement of the problem,

we considered having a chapter on the

one on related research,

on analysis of data, etc., but decided

fragment the focuses of concern.

one on procedures, one

against this approach because

it would

Instead each chapter contains, if

appropriate,
a section clarifying the problem,

a section on related literature, a section

analyzing key concepts, etc.

than it otherwise would have

This we believe gives the report greater unity

and makes for less page turning on the part of

the careful reader.

In much of the work reported

and found that we

on here, we were exploring virgin territory,

were spending a good deal of time on questions prior to

the actual standard experimental

activities of doing and observing.

prior

questions involved clarification

of questions and concepts and their

operational interpretation (including

empirical findings

test construction). Although our

are important, perhaps the. greater

this study lies in the

contribution of

suggestion of ways of approaching the readiness

question, and in the clarification

of concepts basic to the

empirical work.

These

Actually this study only

opens the door to a vast array of investigations.

Chapter II.
In this chapter

the role played by deductive logic; then we shall

examinedeductive logic itself,

types, and comparing the logic

Throughout we shall attempt to

discussing the nature and importance of various

we taught with that which Piaget investigated.

avoid decisions that commit us to a stand on some

of the

the field of logic. Although we do have opinions

about many of these issues,

items) should be

LOgic

we shall first examine the nature of our general concern,

critical thinking, indicating

contemporary issues in

The Subject Matter:

the actual criteria that we use (especially our test

acceptable to the major contending positions in contemporary

logical theory.

A. AN ANALYSIS OF CRITICAL THINKING

Our analysis of critical

by B. Othanel Smith:

"Nov if we set about to find out what...( statement means

and to determine whether to

Which, for lack, of

thinking springs from the basic notion put forward

accept or reject it, we would be engaged in thinking

a better term, we shall call critical thinking."

Note that this notion of Smith's

creative thinking.

(1953, p. 130).

implies a distinction between critical and

In critical thinking, the item to be evaluated has already

been produced.
This distinction between critical
parts the problem of

and creative thinking enables one to divide

in two

teaching people to be good thinkers. Each part is of

course quite important and in practice

fbr purposes of careful

(critical thinking),

the parts are quite interdependent.

But

investigation, it is helpful to separate out this part

for the judgment of which there are already fairly well

established criteria.

A minor difference between

Smith presents can be found in

our notion of critical thinking and that which

the fact that we have included the idea of pro-

ficiency in ours, whereas Smith

does not require that the determination about

11.2

whether to accept or reject be done

thinking take place.

properly or well in order that critical

We believe that our notion is more in accord with everyday

usage, but feel that so long as one is clear about Which notion he is using and

makes sure that the rest of the

theory fits the notion selected, either notion

is workable.

Accepting this basic notion

as revised, the principal investigator has been

developing and refining a list of abilities which characterize the critical

thinker.

The following list represents the current state of this investigation.

Although it is fairly self- explanatory,

one might want to seek clarification in

the selection, "A Definition of Critical Thinking" (Ehnis, 1964a), or in the more

theoretical analysis, "A Concept of Critical Thinking" (Ennis, 1962).

A critical thinker is characterized by proficiency in judging Whether:

A statement follows from the premises.

Something is an assumption.

An observation statement is reliable.

A simple generalization is warranted.

A hypothesis is warranted.

A theory is warranted.

An argument depends on an ambiguity.

A statement is overvague or overspecific.

An alleged authority is reliable.

Although the basic notion,

as revised, calls for the inclusion of proficiency

in judging value statements, for the time being that proficiency is excluded from

our list because value statements constitute one area where fairly well estab-

lished and agreed upon criteria do not exist.

Deductive logic, the subject of the current

study, is a central part of

critical thinking, as analyzed above. First of all, since it deals with the
question of whether a statement follows necessarily from another statement or

statements, it is at least central to the first of the nine aspects listed above.

II-3
However, though opinions may differ on this matter, deductive logic does

not seem to us to ekhaust this aspect of critical thinking. Although logic is

a guide, the validity of most real arguments appears to us to depend on other

considerations.

The connection between the premises and conclusion does not

seem as strict as that in the model provided by deductive logic. The correct-ness
of this claim about the relation between premises and conclusions and the
nature of the relationship that does exist are important topics for philosophi-

cal research.

But regardless of the outcome of this research it must still be

granted that deductive logic plays at least a major role in tht first aspect

of critical thinking.
Secondly deductive logic plays

an important role in the other aspects as

well. Among other things, it is

a constituent of the application of the

aria teria and/or principles of the other eight aspects.

Primarily because of its centrality

its criteria seemed to be in

aspects of critical thinking,

be investigated.

in critical thinking, but also because

more suitable shape than the criteria for other

we picked deductive logic as the first thing to

It is our hope to perform and/Or to stimulate parallel and

complementary investigations of other aspects of critical thinking.

TYPES OF DEDUCTIVE LOGIC

For purposes of this study the

'deductive logic' will

following explanation of the meaning of

suffice: Deductive logic is concerned with whether a

statement follows necessarily

follows necessarily, if, and

from one or more other statements. A statement

only if, its denial contradicts the assertion of

the other statement(e).

Three recognised types of

and ordinal logic.

pending sometimes

deductive logic are sentence logic, class logic,

These types are often called by other names, the

name de-

on one's philosophy of logic, but these names will serve to

identify three types of

logic Which are often

so grouped.

11.4

Sentence Logic.
Sentence logic is concerned with arguments in which the basic units are

sentences.

That is, distinct sentences, often connected or modified by such

logical connectives as 'if', 'only if', 'then', 'and', 'or', 'not', and 'both',
appear essentially unchanged throughout the course of the argument.

Here is an

example taken from "The Cornell Conditional Reasoning Test":

Suppose you know that
Tom may use paints only if he has cleaned up his clay work.

Ling
Tom may use paints.
Then would this be true?
Tom has cleaned up his clay work.
In that series the two sentencea, "Tom may use paints" and "Tom (or he) has

cleaned up his clay work", each appear twice.

In the first of the two given

statements they are joined by the logical connective, "only if", but these two
basic sentences appear essentially unchanged throughout the course of the argu-

ment.

Hence this is a case of concern for sentence logic.

Sometimes sentence logic itself is broken up into parts, depending on the

logical connective which is used.

'if, and only if', or

When the connective is 'if', 'only if',

or

any synonyms of these, we have whet is sometimes called, and

what we shall call, a 'conditional statement'. Arguments which contain only

conditional statements and simple sentences

conditional arguments.

or negations thereof shall be called

Reasoning associated with such arguments shell be

called

conditional reasoning.

Other types of sentence logic

are called by different systems of names.

consistent set labels as 'alternation statements'

those which make use of the con-

nective, 'or'; 'disjunctive statements' those which use the connective 'not...both';

One

11-5
and 'conjunctive statements' those vthich contain the conjunction 'and' (Cohen

and Nagel, 1934).

A more common set of labels at the present time

applies the

word 'disjunction' to statements containing 'or' instead of those containing

'not... both', and considers the latter simply denials of conjunctions.

though one did not have to make

Al-

a choice for the present research project, a

choice will have to be made in the continuation into these areas. The choice

is not simply an arbitrary

logic.

one, since it depends to some extent on one's Philoso-

phy of

Our current inclination is to recommend the former system to

..)eople investigating these

in a 'not...both' statement

areas in the future, because the connection implicit

appears to be more fully recognized with the

sepa-

rate name 'disjunction', and because the names of the former system seem more

natural.

Class Logic.
The familiar traditional syllogisms are arguments in class logic form, but

they do not exhaust class logic arguments,

Bo a more general description is

necessary.

The basic units in class logic

predicates.

The sentences do not

are parts of sentences, subjects and

reappear essentially unchanged; instead the

subjects and predicates are separated from

each other and rearranged. Here is

an example to which the criteria of class logic

are to be applied. It

is from

"The Cornell Class- Reasoning Test":

Suppose you know that

All the people mho live

on )4ein Street were born in Milltown.

None of the students in Boom 352 live on Main Street.

Then mould this be true?

None of the students in Boom 352 were born in Milltown.
Pbr purposes of simplification

and ease of teaching, the subject and predicate

are revised in order to form classes.

Fblloving this procedure the classes

involved in the above example

are 1) the people Mho live on Mein Street, 2) the

people mho were born in Milltown,

and 3) the students in Boom 352. The two

11.6
given statements present relationships between the first and second classes and
the third and second classes respectively. The statement about Which one must
decide suggests a relationship between the third and first classes. Thus the
subjects and predicates as represented by the classes are the basic units in
this kind of reasoning.
Admittedly the representation of subjects and predicates by classes is a
simplification which can result in philosophical problems. However, as long as
one is cautious enough to avoid thinking that the class relationships somehow
more suitably capture the meaning than the original statements, the transformations can be convenient in explanation and teaching.

CTdinalLogic.
A third type of logic deals with size relationships, such as, greater than,

equal to, leEs than, not greater than, etc. Here is an example taken from Bunt's
"Graded Reasoning Tests" (1919):
Tam runs faster than Jim; Jack runs slower than Jim. Who is the
slovest
Jim, Jack, or Tom?

Other TWI of losaa.

No successful comprehensive classification system of types of logic has

even been prepared, so the best we can do here is to list some other types with
an explicit disavowal of comprehensiveness.
Mathematical reasoning is such that the conclusion is supposed to follow
necessarily from the premises and would therefore be classified as logic. Since

we have no desire to enter the jurisdictional dispute (about *lob is a branch

of which) between some logicians and some mathematicians, we simply point out
the consequences of our definition, being ready to accept an amendment, if
desired.

Our definition, however, should at least serve to alert us to the

similarities between the fields of mathematics and logic.

Several other branches of logic with which at least same workbag been
done by logicians are, using the terminology suggested by G. H. von Wright

II-7
(1957, p. 58), alethic logic (traditionally called 'modal logic'), epistemic
logic, and deontic logic. Alethic logic is concerned with statements of possibility and necessity; epistemic logic with knowledge statements; and deontic
logic with statements of obligation. Each of these branches of logic is important in critical thinking, but no one of them has yet been sufficiently worked
out for us to do research on them of the sort which is being described in this
report.

Hopefully the current trend among philosophers toward increased in-

terest in these fields will continue, so that some satisfactory criteria will
be available for use in research like ours.
Presumably other types of logic could yet be identified, because no one
has successfully presented a rationale for the exhaustiveness of some list,
and it has been our experience that numerous examples of inferences that seem
in a way necessary do not clearly and =controversially fit any of the types
mentioned.

Applications of broad value statements and inferences to and from

statements containing such words as 'probably' and 'because' are the sorts of
thing we have in mind.
There is thus still much philosophical spadework to be done, both in
identifying types of logic and in determining vslid patterns of inference for
the relatively unexplored types. Until this philosophical spadework is done,
the investigation of patterns of development, learning, and learning capacity
in the types of logic must remain unfinished.

AN ELABORATION OF THE TYPES OF LOGIC INVESTIGATED IN THIS STUDY

Though an investigation of all types of logic is not yet possible, In were

able to work with two very common and significant types of logic, class logic
and conditional logic. We picked these for several reasons:

The criteria for judging arguments in these forms of logic are fairly

well developed and agreed. upon.

Where controversy does exist,* we essentially

avoided it.

*See P. P. firagille Introduction to Logical Theory (1952) for an indication of

areas of controversy that do exist.

2.

These two types of logic, together with the other kinds of sentence logic,

are the things most commonly taught under the heading, "deductive logic", are thus
well associated with the term, 'deductive logic'. We felt that we were forced to select
from among the branches of sentence logic because of time
pressures, and picked conditional logic because the if-then relationship is

fundamental in all logic and because

with conjunction and negation enables

one to do other types of sentence reasoning. This

selection of conditional logic from sentence logic should be kept in mind as

you read the next two reasons for our selection of certain types of logic, because
they argue for the inclusion of sentence logic instead of only conditional logic.
3.

An informal investigation that

we performed of reasoning in newspaper

editorials, U. S. Supreme Court opinions, and an auto mechanics handbook showed

that class logic and sentence logic included most of the deductive reasoning that

occurred there.

The only notable exception wss deontic logic, the logic

of

obligation.
4.

In the literature on people's reasoning there is some precedent for

the selection of class logic and

sentence logic. Most of the studies

that we have located deal with one or both of these types.

Although ordinal logic is well worked out, is already being taught in

schools, and is considered in the

literature, we omitted it.

We did include

it

in our pilot study, but felt that

year 1963-64, we had to reduce the

to two.

in order to do a more satisfactory job in the

scope of our project from three types of logic

Paray because ordinal logic is inevitably receiving some readiness

attention now; partly because

our pilot study showed that major advances in

capacity for mastery of the principles

grades (ages 6-12), especially

of ordinal logic occurred in the primary

early primary (ages 6-9) for most of the students

we studied; and lastly because work of

personnel for younger children,

the type we were doing requires more

we decided to omit ordinal logic in the study

11-9

being reported on here.*

Given class and conditional logic

is in order.

as our concerns, a fuller account of each

In what follows we shall set forth and discuss what we consider to

be the basic principles of each.

These lists of principles are not in the

most
elegant form (for the more elegant they are, the less easy they are for most

people to understand),

nor are they in the most easily understandable form, but

are in a form that is something of a compromise between the two.

In our teaching of logic we did not

use the language of these statements of

principles, because we felt much of it to be

This is so especially for class logic,

system of circles which provided

too formidable for teaching purposes.

in the teaching of which we made use of a

a model for class relationships.

With but one exception the principles are tested for in our tests, which will

be described later.
given in Chapter IV, which describes the

tests.
These combinations are not listed here.

ciples are also tested for in our tests.

They will be indicated in Chapter IV.

In the hope that they will be self-explanatory,

as follows:

these lists are arranged

On the left the principle is stated in Ehglish. In the center and on

right arguments appear.

the

In each case the principle justifies a judgment of

valid (conclusion follows necessarily) or invalid (conclusion does not follow

necessarily) with respect

to the argument. The argument in the center is in

symbolic form. The argument on the right is an example taken from one of our

tests and modified for the

purposes of this method of presentation.

The conditional logic principles

classroom presentation.

are ordered in our recommended order of

This ordering is based upon our experience teaching

logic, but has not been subjected to experimental test.

* Professor Lucille Bingel

of Fenn College, Cleveland, is preparing a report of

the work that we did do with ordinal reasoning.

TABLE II-1. Illustrated Basic Principles of Conditional

Principle
1. Given an

Symbolized Argument
if-then sentence,

the affirmation of the if-part

implies the affirmation of the

then-part.

2. Given an if-then sentence,

the denial of the if-part does
not by itself (as a result of

its being an if-part) imply

If p, thenq.

P.
Therefore q.
Valid.*

If p, thenq.

Concrete Argument

If the hat on the table is

blue,

then it belongs to Joan.

The hat on the table is blue.

Therefore the hat on the tablebelongs to Joan.

If Tom lives in the white

house,

Not p.

then his last name is Smith.

Invalid.

Therefore Tom's last name is not Smith.

Therefore not q.

the denial of the then-part.

3. Given an if-then sentence, If p, then q.

the affirmation of the then-

Logic

q.

Tom does not live in the white house.

If Mary lives in the whitehouse,

then her last name is Brown.

part does not by itself (as a

result of its being a then-

Therefore p.
Invalid.

Mary's lest name is Brown.

Therefore Mary lives in the white house.

4. Given an if-then sentence,

If p, then q.

If the car in the parking lot is Mr.

the denial of the then-part

implies the denial of the if-

Not q.

Smith's, then it is blue.

part) imply the affirmation

of the if-part.

Therefore not p.

part.

Valid.

5. The if-then relationship

If p, then q.

is transitive.

If q, then r.
Therefore, if p,
then r.
Valid.

The car in the parking lot is not blue.

Therefore the car in theparking lot is not Mr.

Smith's.

If Sam misses the bus, he will walk toschool.

If Sam walks to school he willcross the bridge.

Therefore, if Sam misses the bus, he willcross

the bridge.

The
validity status of the arguments will only be specified for the symbolized ones.
It is the same in each
case for both the symbolized and concretearguments.

TABLE II-1 (cont.)

6.

An if-then sentence implies

its contrepositive.

7. The if-then relation is

non-symmetric.
8.

Given an only-if sentence,

the denial of the only-if part

implies the denial of the major

part.

9. Given an only-if sentence,

the affirmation of the major
part implies the affirmation

of the only-if part.

10. The denial or affirmation
of one part of en if-and-onlyif statement implies respectively
the denial or affirmation of the

other part.

11.

Given an only-if sentence,

the affirmation of the only-if

pert does not by itself (as a
result of its being an only-if
part) imply the affirmation of

the major part.

12.

Given an only-if sentence, the

denial of themajor pert does not by
itself (as a result

of its being themajor

part)
imply the denial of the only-if

part.

If p, thenq.
Therefore, if
not q, then
not p.
Valid.

If p, thenq.

Therefore, if
q, then p.
Invalid.

If Mrs. Smith entered the flowershow, then

she entered her roses.

Therefore, if Mrs. Smith didn't enter her

roses, then she didn't enter the flower show.

If the chair is

green, then the table is black.

Therefore, if the table is black, then the chair
is green.

p only if q.

John is in the kitchen only if thereis food in

Therefore notp.

There is no food in the kitchen.

Not q.

Valid.

p only if q.
p.

Therefore q.
Valid.

p, if, and only

if, q.
Not p.

the kitchen.

Therefore John is not in the kitchen.

Harry is on the football teem only if he has his

mother's permission.
Harry is on the football team.

Therefore Harry has his mother's permission.

Bill will see Audrey thisyear, if, and only if, he

goes to Montreal this year.

Bill will not see Audrey this

year.

Therefore notq.

Therefore, Bill is not going to Montreal this

p only if q.

Dick is using the classroom dictionary only if the

Valid.
q.

Thereforep.

Invalid.

library is closed.

The library is closed.

Therefore Dick is using the classroom dictionary.

P only if q.

Jane went to the park

Not p.

her friend Pet yesterday.

Therefore notq.

Invalid.

year.

yesterday only if she saw

Jane did not

go to the park yesterday.

Therefore Jane did not

see her friend Pat yesterday.

TABLE II-2.

Illustrated. Basic Principles of ClassLogic

Principles

Symbolized Arguments

Concrete Arguments

1. Whatever is a member of

All A's are B's.

All of John's pencilsare blue.

of that class and vice versa.

that at least some A's

John's pencils are not blue.

All A's are B's.

All the cars in the

a class is not a non - member

2. Whatever is a member of
a class is also a member of
a class in which the first
is included. (Thisimplies

Therefore it is false

are not B's.

Valid.

All B's are C's.

Therefore all A's
are C's.

Therefore it is false that at leastsome of

garage are Mr. Smith's.

All Mr. Smith'scars are Fbrds.

Therefore all the cars in thegarage are Pbrds.

that class inclusion is

Valid.

3. Whatever is a member of a

All A's are B's. All the red booksare John's.

transitive.)

class is not (as a result of

that relationship) necessarily

a member of a class included in

that class.

4. Class exclusion is
symmetric.

Therefore all B's

are A's.

Therefore all John's booksare red.

Invalid.

No A's are B's

(and there are B's).

Therefore no B's

None of Jane's dolls havehats.

Therefore none of the dollsthat have hats

are Jane's.

are A's.

Valid.

5. Whatever is a member ofa

All A's are B's.

class excluded from the first.

Therefore no A's

class is not a member ofa

6.

Whatever is not a memberof

a class is not (as a result of

that relationship)necessarily

also not a member ofa class in

No B's are C's.

are C's.

Valid.

All A's are B's.

No C's are A's.

Therefore no C's are

B's.

All Frank's homework isdue today.

None of the homework duetoday is in history.
Therefore, none of Prank's homeworkis in history.

All the

papers in the box are torn.

None of John's

papers are in the box.

Therefore none of John'spapers are torn.

which the first is included. Invalid.

TABLE 11.2 (cont.)

7. Vhatever is not a member of

a class ic not (as a result of

that relationship) necessarily
a member of (nor a non - member
of) another class Which is
excluded from the first.
8. Whatever is not a member of

a class is also not a member

of any class included in the

first.

No A's are B's.

No C's are B's.

None of my shirts are wool.

some A's are

C's. Invalid.

None of the shirts hangingup in the closet

are wool.
Therefore, at least some of my
shirts are hanging up in the closet.

All A's are B's.

All of Joan's homework is due today.

Therefore at least

No C's are B's.

Therefore no C's are

A's.

Valid.

None of the math homework is due today.

Therefore none of the math homework is for Joan.

On the other hand the principles of class logic are not strictly in a recommended teaching order, since our teaching method for class logic makes

use of

the circle system mentioned earlier. But we think that the first four of these
principles are grouped roughly in teaching order anyway.
A few words of summary of these principles are in order.

The conditional

principles cover affirming and denying the antecedent ("if-part" and "major part")
and consequent ("then-part" and "only-if-part") in both if-then form (1-4) and

only-if form (8,9:1142). Transitivity (5), contraposition (6), non-symmetry (r),

and the combination of 'if-then' and 'only if' (10) are the other things covered by

the conditional logic principles. A person familiar with conditional logic will
recognize these as the basic elements of conditional logic, though he mightof
course feel some redundance if he seeks logical elegance.

It is our view that in

spite of this sort of redundance, the principles must at least for teaching purposes be spelled out to the extent that we have done so. And then there is the
very difficult question of whether they really are redundant anyway, because the
different forms (as we have specified them) are used under different circumstances.

librtunately we do not have to settle the question because teaching requirements

force the large number of principles on us.
The class principles cover first of all the basic meaning of being and not
being a member of a class (1), and secondly the basic notion of two classes' being
excluded from one another (4).

Then the main body of principles falls into two

groups: those starting with the assumption that something is in* a class (or
part of a class), (2,3,8) and those starting with the assumption that something
is not in a class (or pert of a class), (5,6,7). Each group is then subdivided
into three parts according to whether that class (or part of a class) is
included in a

third (8,7).

third

(2,6),

b) includes a third (3,5) or

.0

c) is excluded from a

Thus the possibilities are all covered.

* Deliberately vague to cover both membership and inclusion.

1I-15
We regard class principles numbered 2 and 3 as the two mast important to
teach to someone if he does not know them. Number 2 catches the transitivity
of inclusion while number 3 catches its non-symmetry.
A good way to grasp the impact of the class principles, if the arguments
do not suffice as explanation, is to use circles inside one another or separate
or
from one another to represent the inclusion at exclusion relationships.* A
conclusion inescapably diagrammed by the diagramming of the premises follows necessarily.

If not inescapably diagrammed, then it does not follow necessarily.

That completes our direct characterization of the logic that we taught.

The next section, which compares this logic with Piaget's, indirectly provides
some further characterization.

A COMPARISON WITH PIAGET'S LOGIC

Jean Piaget is the leading figure in the study of the development of

children's knowledge of logic, though not their capacity to learn logic, which
was in addition one of our major concerns. Since a comparison with his conclusions is inevitably called for, it is important to try to be clear about
whether What Piaget calls 'class logic' and 'propositional logic' correspond
to what we have called 'class logic' and 'sentence logic', because only to
the extent that they correspond does the question of agreement or disagreement
about the facts of development arise. In some ways the logics seem to correspond and in some ways they do not seem to do so. In fact it is difficult
to be sure of the extent of correspondence. This might be explainable by
Piaget's difficult style; and it might be explainable by changes over time
in his concepts to adjust to the facts or his own developing interests. Regardless of the explanation, the points of correspondence and lack of correspondence, as we see them, follow. We must confess that we have not been able
to construct with confidence what we consider to be a coherent consistent
account of Piaget's logic, but we shall do the best we can. Discussions of
*This is a simplified account, as you might expect.

ix-16
Piaget's logic have also been presented by Parsons (1960) and Revell (1963).
Points of Correspondence.
First of all Piaget did talk about logic. Secondly he does distinguish
between two types of logic, and thirdly this distinction basically seems to
correspond to the distinction between class logic and sentence logic that we
drew' earlier in the chapter.

We shall expand on each of these points.

Piaget seems to have adopted some of the basic moves and definitions in

contemporary propositional logic (1958*, pp. 293-303). His logical operators,

'1,4,

'3', 1.', and '-', correspond to the operators of traditional propositional

logic, and stand respectively for 'or',

then', 'and', and 'not'.

He

makes use of the same letters as are traditionally used to represent propositions,
'p', 'q', and 'r'.

And he generally relates all these in a way that essentially

conforms to that of Rgy elementary text in symbolic logic.

Furthermore his class logic symbolism, definitions, and rules are in conformity with at least much of what is traditionally classified as class logic.
Among other things, he uses capital letters to represent classes and uses a
single quote after a letter which represents a class to represent the complement
of that class (1958, pp. 274-77).
Hence it is quite clear that Piaget was talking about logic, establishing
at least- some correspondence between his interests and ours.

Along with us, Piaget resists the contemporary trend toward the merging

of propositional and class logic.

This merging is accomplished by treating class

inclusion statements as modified conditionals.

For example, the statement, 'All

floating bodies are light', is transformed into the statement, 'For every x, if
x is a floating body, then x is light'.

Henceforth in this section on Piaget, all references made by the use of '1958'
shall be to Inhelder and Piaget's The Growth of ical Thinki From Childhood
to Adolescence. We believe that the parts to i
we re er were wr
en
Piaget and not Inhelder.

II-17
Sometimes the transformation goes further, putting the statement in terms
of existence, conjunction, and negation, making the above statement look like
this: 'It is not the case that there exists an x such that x is a floating body'
and x is not light.'
One of Piaget's reasons, with %thich we are in sympathy, is that to so
interpret class inclusion statements would be to use "a complex language for
describing phenomena which do not go beyond much simpler structures in the
subject's mind" (1958, p. 280).
3.

A third point of correspondence lies in the similarity of his basis for

the distinction between propositional and class logic and our basis for the distinction between sentence and class logic. Piaget says that although class logic

does deal with propositions, "decomposing and recomposing the content of propositions" (1958, p. 292), it does not deal with the combination of these proposi-

tions as independent units.

Propositional logic, on the other hand, does this.

In another place Piaget in characterizing the formal operational period

by a person's ability to do propositional logic,

says, "ibrmal operations,

therefore, consist essentially of 'implications'

established between propositions which themselves

seriations, etc."

... and 'contradictions'

express classifications,

(1950, p. 149).

Thus for Piaget, as for us, class logic is concerned with the internal
features of propositions which are not themselves composed of other propositions, whereas propositional (and sentence) logic is concerned with the relations between propositions (or sentences) which themselves remain essentially

unchanged throughout an argument.

We do not feel that the difference in

names ('propositional' vs. 'sentence')

constitutes a significant difference between

Piaget and us.

His term,

'propositional' is in more common use among logicians; for that reason we would have
chosen it if we had not felt that communication with teachers and students

will be

II-l8

facilitated by the use of the ter

'sentence', which already has meaning for

them.

The term 'proposition' in the sense in which it is used by logicians is quite

unfamiliar to most people.
Nov it is the case that the two terms can be used to mark different positions

in philosophy of logic, but the use of one term or the other does not necessarily
commit one to one of the positions, since the choice of one term or the other is
often simply a matter of convention or convenience. The positions that can be
marked by these terms are positions about the nature of the basic units that are
connected and modified by the logical connectives, 'or', 'if..., then', 'and',
'not', etc_

Are the basic units merely strings of words concocted by human

beings, or are they possessed of some sort of independent existence, regardless

of whether they are formulated by Wow beings?

We do not want to take a stand on this issue; we do not even want to take
a stand on whether or not it is a genuine issue. But we do want to use the
term 'sentence' because of its established usage. We do not know whether Piaget
takes a stand on the issue, but feel that as far as the teaching of the rules of
logic goes, it does not matter. His difference in usage does not represent a
substantive difference.
Points of Noncorrespondence.
There are two points of noncorrespondence which we would like to suggest:
First the judgment about whether propositional or class logic is in use seems at
times to depend for him upon whether there is a consideration of all the possibilities inherent in the situation.

the one we worked with.

Secondly Pieget's system is more simple than

We are less sure about the first difference than the

second.

According to Piaget (1958, pp. 272-333) an important feature of a

person's reasoning is the extent to which he works within a system of all the
possible combinations of the variables and views what he

sees as one of these

11-19

possible combinations. Associated with this, according to Piaget, is the subject's

ability to separate out the variables. Presumably if a person separates out
variables and holds certain ones constant, then he is eliminating certain of
the possible combinations of variables as possible causes of the phenomena being
investigated.

There is a strong suggestion that a defining characteristic of ability at

propositional logic is ability to work within a system of all possible combinations.

FUrthermore a defining characteristic of the use of class logic is the failure to

work within a system of all possible combinations.

These are not defining

Characteristics of our sentence and class logics.

He says, in searching for a way of telling "in which cases the subjects
reasoned through arrangements of classes and relations and in which cases they
used propositional operations" (1958, p. 279), "It is fruitless to look for an
exclusively verbal or linguistic criterion -- eaal considering all statements
containing the words 'if...then' as implications while regarding the statements
which do not contain them as inclusions or correspondences, etc." (1958, p. 279).
He then says that a better method of making the judgment about whether class or
propositional logic is in use is look at all the actions of a subject and see:

whether he tries to separate out the variables, Lihicginplies both

hypothetico-deductive reasoning and a combinatorial system; when they
appear, we have to interpret the stated judgments as propositional
ex-pressions ... (1958, p. 279).
Thus separating out the variables seems to be a criterion of the use of propositional logic.

But the "surest method of differentiation", according to Piaget is to see if

the subject interprets a given correspondence as the result of any one of several
possible combinations:
If...the subject interprets a given correspondence as the result of any
one of several possible combinations, and this leads him to verify
his hypotheses by observing their consequences, we know that propositional
operations are involved (1958, p. 279).

11-20
It is because of these statements and statements like them that we suspect
that Piaget holds that a definitionally necessary condition for the use of
propositional logic is that of working within a system which contains all the
possible combinations of the variables, and that a definitionally necessary
condition for the use of class logic is the lack of working within such a system.
What it is to work within such a system still remains to be clarified and un-

fortunately Piaget dces not make himself fully clear. Here is the best account
that we can work out:

According to Piaget, given two propositions (or classes), there are sixteen
possible ways in which they can be grouped. For propositions he calls these ways
the "sixteen binary operations" (1958, p. 293) and for classes he simply lists
them as "sixteen possible combinations" (1958, p. 277). The two lists correspond,
as he indicates in a series of footnotes in his discussion of the sixteen binary
operations (1958, pp. 293-303), so one will present only the list of the sixteen
binary operations.
In this list 'p' and 'q' stand for propositions. A denial of the proposition
is represented by the symbol with a line over it, e.g.,

A conjunction of two

propositions is shown by putting a dot between them e.g., 'p.q' which means
'Either p or q' is represented by 'p v q', and 'If p, then q' is

'p and q'.

represented by
The sixteen binary operations:

v pq vv

The negation of #1.

6.

14.

pq
The negation of #3:
p v q

The negation of

p)q

The negation of

p.q

#5: 5.i

#7:

ci,P
The negation of #9: p.q
(p

q)(q,p)

The negation of #11:v

p(q v q)

The negation of
q.(p v

16. The negation

of

#13:

v -q")

#15: "ci.(p v

11-21

A person acquainted with symbolic logic can see that (given the symbolic logic
interpretation of these symbols) this list contains the affirmation and denial
of each of the possible groupings of p and q and their denials. For him who is
not well acquainted with symbolic logic the exhaustiveness of the combinations
that Piaget sees perhaps can be shown thusly:
Let 'a' stand for 'p. q'

Let 'b' stand for

Let 'c' stand for
Let 'd' stand for lie.%-oP

These are the four possible conjunctions of the assertion and denial of
'q'.

and

Then there are sixteen possible ways that 'a', 'b', 'c', and 'd' can be

grouped:

c
d

A.
B.

ab
ac
ad

1.

be
bd
ad

2.

abc

abd
acd
bcd
abed
0 (that is, none of them)*

If one accepts the assumptions, one can see the exhaustiveness of the gyitem of

combinations of the assertion and denial of the two propositions.

Nov we suspect that for Piaget a

person is not doing propositional logic unless

he is operating within a framework that will take account of all of these

possibilities.

If this suspicion is correct, then Piaget's logic does not corre-

spond to the logic that we taught,

temporary interpretation.

nor does it correspond to the standard con-

The latter might say that a person is not fully pro-

ficient if he does not operate within such a framework, but it would not deny the

his system of explanation is suggested by naval (1963, pp. 213-14).

II-22
application of the tem 'propositional logic', to the following argument,

if

advanced by a person who was not aware of nor working within a complete combinatorial framework:
If p, then q. But not q.

Therefore not p.

We cannot be sure whether this interpretation of Piaget is correct, because

it does not seem compatible with the distinction between propositional and class
logic which we presented under "Points of Correspondence". If the interpretation
is correct, then there is a vast difference between the two logics.
2.

A second point of noncorrespondence, one of which we are confident, is

the greater simplicity of the system that he advocates. His system is more simple
in several different ways:
a.

The list of sixteen binary operations together with the interpretation

given to the symbols shows that Piaget is accepting the reduction. of the various
logical operators to two: conjunction and negation.* FOr example, for him, 'If
T1, then q' means the same as 'It is not the case that both p and not q'. Symbolically,

:11 =

tpce

The acceptability of this reduction is a contemporary issue in logical theory.

Common practice is to accept the above symbolic equivalence, but to deny that
'Ip:)q! captures the meaning of 'If p, then q'.

so far as we can tell.

Piaget makes no such reservations,

In case any of our readers would like to pursue this issue

we mention the fact that prominent contenders are P. F. Strawson and P. H. Grice,

both of Oxford, University.

Strawson opposes the reduction and Grice tends

toward
supporting it.**

* Another way of looking at it is to say that the reduction is to 'v' and negation,
.

but it amounts to essentially the eame thing.

** Strawson's views appear in his Introduction topical Theory (1952); Grice's

views are as yet unpublished. Strawson'sboc4:1111 get one started on this
issue.

11-23

Our system differs from Piaget's in that we do not accept this reduction and

the consequent simplification of logic.

We have not attempted this

simplification
in our teaching of logic, but in the tests we built to determine a person's

knowledge of logic we avoided using

any items that would be answered incorrectly

by a person following the simplified system.

b.

A second simplification which Piaget has adopted is the merging of propo-

sitions and what are called 'propositional functions'. This simplification is

not one that is accepted in the field of logic. First we must explain what a
propositional !Unction is and then we shall indicate the possible significance

of this merging.
Consider the statement, 'If a body floats, then it is light'. Although

this might at first appear to be

(or propositional) logic,

a standard conditional statement in sentence

note that neither of the units can stand alone.

Neither

of the units can meaningfully be affirmed or denied by itself, if it retains the

meaning that it has in the Whole.

In the original statement, the words 'a

body'

do not refer to any particular body,

stand alone, it does not make

so if we try to make the group, 'a body

floats',

any assertion that can be called true or

false. (Try to imagine how

you would prove it true or false.)

Since it does

not

say anything, it cannot stand alone.

The groups, 'a body floats' and 'it is light', are propositional functions

in the original statement.

They are not propositions or sentences, given the

meaning they have in that original statement, because they cannot stand alone

without being given a change in meaning.

Piaget in his analysis of Inhelder's

floating bodies experiment, treats the

statement, 'If a body floats, then it is light',

as a statement in propositional

logic, and is enabled to do

tions.

so by merging his propositional functions and proposi-

II-24
Nov why

might

he want to be able to treat such a statement as one of propo-

sitional logic? This is a difficult question and we can only hazard a guess, a
guess that does fit in with an earlier-suggested point of

noncorrespondence, his

casbinitorial criterion.
Note that the statement, 'If a body floats, then it is light' means roughly
the same as 'All floating bodies are light', the latter clearly being a statement
in class logic.

Given these

roughly

equivalent statements, a person is free to

use either one he wishes as an interpretation of the thought that all floating
bodies are light.

Nov if for deciding whether the logic in use is propositional

or class reasoning, his criterion, is, say, the combinitorial criterion, then he
can adjust the form of the sentence selected to fit the decision called for by
the criterion.
To be more specific, suppose that on the basis of the subject's total behavior, we

decide that he is operating

within a framework

that includes all the

possible combinations. Men we decide that the logic is propositional and we

use the

if-then form to

represent the thought that all floating bodies are

light,

which he presumably is entertaining as an hypothesis. We say that he is con-

sidering the

view, 'If a body floats, then it is light'.

Suppose on the other hand that the subject is not operating within a framework that

considers

all the possible combinations, then we judge the reasoning to

be class reasoning and

are light.

pick the class form of the thought that all floating bodies

We say that the subject is considering the view, 'All floating bodies

are light'.
Perhaps Piaget did not merge propositions and propositional fUnctions for

this reason, but so merging does

enable him to use the combinatorial criterion

and still roughly conform to established usage of the words,

'propositional' and

'class'.

This merging of propositions and propositional functions is something we

avoided in theory and tried to avoid in practice. Hence his simplification
mates a difference between the sentence logic that we

taught

and the propositiona3.

11-25
logic that he investigated.
c. A lest simplification is Pieget's apparent

page 11-7) to ordinary

reduction of elethic logic

(see

propositional logic. He stretches his symbolism to

mean
cover moves in alethic logic. Fbr example he wants to let
that each conjunction, 'p.q' and
represents a possible state of affairs.
The standard symbolic logic interpretation is that

that each is possible.

That Piaget also holds the

Shown in his explanation of

v q':

at least one is true, not

standard interpretation

"Disjunction kvisignifies that

true or q is true or both are true." (1958,

is

E. is

p. 296). He does not in this

interpretation say that each is possible.

In his explanation of the floating bodies experiment, however, he does so

interpret What he calls disjunction.

Otherwise the argument that he

endorses
is invalid.

the argument symbolized goes as follows:

Premise:

p.q. v
Conclusion:

It is false that p:)q.

In the standard interpretation of the symbols, that argument is invalid. But

if the premise is to be interpreted

as saying that p.q is possible and that

is possible, then it is false that

p:>q.* That is, if it is possible that

p and not q, then p does not imply q, other things being equal. So Piaget has

avoided the complexities of

an additional symbolism and set of rules by trying

to include alethic logic symbolism and rules in his basic set.

This treatment on his ;art is

exemplified, among other places, in his

analysis of the results of the floating

bodies experiment. In the

particular

discussion we want to present he

the assertion that the bodies will

assigns the variables as follows:

float and

1 Lril

"I

... any factor associated

...

with 2.

-- for example lightness (absolute)...." (1958, pp. 39.40).

* Assuming that

and 'q' stand for propositional Ainctions.

11-26
The statement being disproven in this selection is represented by tiv:)ce.

Presumably this is to mean, 'If a body floats, then it is light'.*

Novrwith all that introduction, we hope it is clear that in this passage
he is attempting to let his symbolism fit alethic logic:
The subject may note the two possibilities combined (p.q) v
i.e, of p.(q v q) -- which constitutes the operation we may speak of as
the affirmation of p independently of the truth or falsity of q. But this
operation contains p.q and amounts to discarding This is what Fran, for
example, says when he declares that "the wood can be heavy (or light) and
it floats"... (1958, p. 40).
The reasoning that Piaget is describing here goes like this:
It is possible for something to float and be light; it is also
pos-sible for something to float and not be light. Therefore it is not
the case that if something floats, it is light.
This argument seems legitimate, but it is not an argument in propositional
logic.

Arguments dealing with possibilities in this way are to be judged by

the rules of elethic logic (which are not yet well worked out, by the way).
In subsuming elethic logic under propositional logic, Piaget has done some

simplifying which we have not done in our presentation of logic. Instead we

have avoided elethic logic.
Summary.

Thus there are points of similarity and points of difference between the
logic with which we worked and the logic reported on by Piaget.

Similarities

are that we both seem to be dealing with logic, that we both draw a distinction
between two types of logic, and that the distinction

appears similar. A pos-

sible point of difference is his use of the combinitorial system criterion in

drawing the distinction between class and propositional logic. A definite

point of difference is the greater simplicity of the system he uses. This
simplicity appears in the reduction of propositional logic to two basic
* Let us neglect the difficulties in his assignment of the variables, i.e., his
merging of propositions and propositional functions, and his neglect of the difference between propositions and concepts. ('p' is here assigned by him to a

proposition (or sentence),

to a concept or term)).

11-27
operators, in the merging of propositions and propositional fUnctions, and in

the merging of alethic logic and

propositional logic. We are of the

opinion
that our results can probably be compared with Piaget's, but want to retain

as qualifications the points of non-correspondence specified above.

A further difference, not in logics,

but in intensity of analysis,

can be

found in our attempt to study separately various principles of the types of

logic with Which We worked.

Piaget, on the other hand, tends to speak

globally about his two types of logic.

CHAPTER SUMMARY

In this chapter we have outlined the conception of critical thinking under

thich we are working, have suggested the significant role that deductive logic

in this conception of critical

plays

thinking, have sketched out a number of

types of deductive logic, have elaborated on two very basic types, and have

compared these two types with Piaget's

conception of similar types.

Primarily

in order to work within a small

organization and thus maintain control of that

happens, and in order to work with types of logic that are both significant and

fairly well 'worked out, we picked class

and conditional logic

as the two on

which we would concentrate.

Conditional logic is in

our view the most important part of that we call

sentence logic and that is often

judged
Piaget's class logic and

We have

propositional logic similar enough to our class logic

and sentence logic to declare

tions, te feel that what

called propositional logic.

them roughly comparable,

so, with some reserva-

we have learned about the development of knowledge of

logic does have a bearing

on that Piaget has claimed.

CHAPTER III.
So that our readers will,

Basic Data on Subjects

in reading about the results, have an idea about

from Whom these results were secured,

we are presenting basic information about

the subjects before we attempt to present, discuss, and interpret the results
in detail. In this chapter we shall indicate the institutional nature of the groups
with whom dealt, shall comment Where necessary on the instruments we used (except for
the logic tests, to which we devote the entire next chapter), and then shall
provide summary information about our subjects. In order that

the grouping and categorizing be intelligible,

we shall sketch in reasons for

some of the decisions.

THE GROUPS WITH WHICH WE WORKED

The "LDT's".

At each of the grades 4-12

work intensively.

we selected one class-size group with Which to

With each of these groups one of the members on our staff

worked for one period per day (around 40-50 minutes) for four weeks (20 instruc-

tional days).

Approximately three quarters of this time vas devoted to the

teaching of logic.

The other quarter was in general devoted to advancing the

subject matter that was ordinarily scheduled for that time.

Because we discovered in our pilot study that it was not possible to teach
thoroughly even one type of logic in the time we had available, we decided to

years between the two types

we selected. Class logic was taught in

grades 4, 6, 8, 10, and 12.

and 11.

We shall label these nine

alternate

Conditional logic was taught in grades 5, 7, 9,

groups our 'LIT' groups, indicating groups

to which Logic was Deliberately Taught.

Our LDT's were selected from
student population of about 8,000

an Upper New York State school system with a

in grades K.12. The district is called a 'city

school district' but also comprises suburban and rural communities which were

added in various centralization

proceedings over the last thirty years. About one

half of the students live in the city.

The city itself has a population of about

111-2

14,500, plus about 14,500 college students. The ixn industry in this community
is college education, but there are a number of medium- and small-sized factories
as well.

No census figures that indicate the nature of the school district population
are available, since in New York State, the boundaries of school districts do not
generally coincide with political boundaries. But we can give figures that indicate some things about the occupational nature of thepopulation of the city.

Fbr

comparison purposes we are also supplying similar figures for New York State and the

United States.

TABLE III-1. Percentage Occupational Make-Up of City, New York State and the United
States*

2.
3.

United
States
10.8

25.7

New York
State
12.5

Clerical and Kindred

18.9

18.1

14.4

Service workers, except private household

16.8

9.3

8.4

Occupation
1.
Professional, Technical and Kindred Workers

City

4.

Craftsman, Foreman and Kindred Workers

7.1

12.4

13.5

5.

Operatives and Kindred Workers

6.8

18.1

18.4

5.9

9.0

8.4

6. Managers, Officials and Proprietors

7.

Sales Workers

5.3

7.3

7.2

8.

Unemployed

3.5

5.2

4.9

9.

Miscellaneous

13.5

13.3

18.8

*U.S. Bureau of the Census. U. S. Census of Population: 1960. VOL I, Characteristics of the Population. Part 34, New York. U.S. Government Printing
Office, Part I, United States Summary, Washington, D. C., 1963.
In this city there is a higher percentage of professional, technical and kindred
workers than in the state and the nation.
We tried to secure classes that would typify this school systewand, to the
extent that this school system is typical of the United States, that valid be more

111-3
broadly typical.

TLib was difficult to arrange, given the numerous pressures

that impinge upon the public school

administrator, but the people with whom we

worked were very helpful and did the best they could.

Our classes in grades 4-6 were from

broad distribution of backgrounds.

an elementary school which contained a

The classes were existing units which were

turned over to our staff members for the designated period.

In grades 7-8, the classes made available to us were arithmetic, and the

9th grade class composed of students

taking algebra, but who were not doing so for one reason or another.

Because it was not possible to take

in high school classes of

so much time from the course of instruc-

tion

average or above-average ability (the New York

State Regents examinations

were a factor operating here), our classes at the

high school grades were composed of volunteers from study halls.

The "LNDT-1's".

At each grade level from the

parable class in which Logic

same school system we requested a fairly com-

was Not Deliberately Taught.

Hence the

designation,

'LNDT'.

Since there was another set of students to which logic was not directly

we added the numeral '1' to

the designation. The LNDT-1's served two

purposes: to provide more subjects for

ment of knowledge of logic, and to

All the data that

taught,

our study of the natural-cultural develop-

provide a check on the efficacy of our teaching.

we have on the LED's we also have on the LNDT-1's.

The LNDT -2's.

Another group of students,

the LNDT-2's, to whom we did not teach logic was

selected from a nearby school

community of which the city is the

is a suburb of the city.

system.

It is in an area that is part of the larger

hub, and, although it is a rural community, it

We did not gather the

economic status index

same data on the LNDT-2's as on the others. .A socio-

was not secured, and the IQ scores were not from the same

III-4
test.

These subjects were used primarily for data on the natural-cultural develop-

ment of knowledge of logic.

We attempted to test all the

students in grades 4-12

in the school system with our logic tests.

THE SOURCES OF OUR DATA.

We have the following data on our subjects:

Grade level in school.

Chronological age.

Sex.

I.Q. score.

An index of socio-economic status (except LNDT-2's).

Pre -test and post-test

Deduction Test.

scores on the appropriate Cornell

In this section we shall discuss, where appropriate, the instruments and/or

techniques we used to gather these data. However, we shell not here be concerned

in any detail with the instruments

that we developed, the conditional and class

reasoning tests, since a separate chapter

is devoted to them.

Readily Obtainable Information.

The first three items, grade level,

available from school records.

Chronological age was determined as of the date

on uhich the pre-test was given, which is

tion at the given grade level started

occurred in February

chronological age, and sex are routinely

approximately the date on which instruc-

for the LDT's of that grade level.

This

or March of 1964, depending on the schedule for that

particular grade.

I.Q. score, although

on the school records, required more discretion in

system,
data gathering.
Fbr the LDT's and LNDT-1's, which were from the same school
we chose the latest available
Tests.

score on the Lorge-Thorndike Intelligence

Where both verbal and non-verbal scores were available, we took the

mean; where only

one was available we used that. Fburth graders were

arithmetic

III -5

tested

with level 2 Lorge Thorndike,

the rest with level 4.

fifth and sixth graders with level 3, and

All I.Q. testing had been done within the previous two

years except for that of the eleventh

tested in sixth grade and eighth

and twelfth graders, most of whom were

grade respectively.

Thus we were able to se-

cure fairly comparable scores from the school's records for the LDT's

and

LNDT-1's.

the use that we made of the

The school system from which
majority of the
but because

LNDT-2's did not require the

same I.4. test.

they were drawn used a wide assortment of tests.

scores of LNDT-2's are on the California Test of Mental

The

Maturity, 57S,

many are from other tests, the mean IQ's Which we

report should be understood to be rough estimates.

Less

day Obtainable Information.

Socio-economic status.

Because of our interest in the relationship

between socio-economic status (henceforth

we made a rudimentary attempt to

bears a relation to SKS.

parent, making

secure for each LDT and LNDT-1 a number that

This number is based upon the occupation of the

use of Warner's seven-place occupational scale (1949, pp** 140-41).

Occupations with the highest

those with the lowest

listed in the school

socio-economic status receive a rating of 1p and

socio-economic status receive a rating of 7.

Two raters separately

applied Warner's seven-point scale to the occupation

records for each student's father, or mother if she was the

principal means of support.

records, we made,pn effort,
did not, however,

When the occupation did not appear in the school

which was generally successful,

to obtain it.

We

approach the subject's parents and ask them, feeling that this

approach would be resented,

able.

called 'STS') and knowledge of logic,

In exactly 11

unless we devoted more time to it than we had avail-

cases out of 428 (the total number of LDT's and LNDT-143),

we were unable to determine the

assignment. In those cases

occupation with enough specificity to mac an

we made an arbitrary assignment of 6, because it

iii-6
has been our experience that lack
is a good indicator of low
hold in every

of information or information that is too vague

socio-economic status. Naturally this assumptton does

not

case, so we made every effort to keep our use of it to a minimum.

We made the following

scheme as presented

additional amendments NA/or additions to the Warner

on pp. 140-41 of Social Claes in America.

These changes

were

made on the basis of

our knowledge of the prestige structure of this community,

with which each rater has

Unemployed people are not

inevitably there will be

At this time in the

Although

economic life of our country, this assignment is more realistic

say, during the Depression.

People in the armed forces

assigned the number

given a number in the Warner scheme.

some errors in so icing, we assigned them the number 7.

than it would have been,

had considerable experience.

are not given a number in the Warner scheme.

We

3 to officers (rank unspecified) and 6 to enlisted men.

Ccilege teachers also

are not assigned a number.

To them we assigned the

number 1.

Warner assigned the number

Warner assigned the

number 5 to hardware salesmen.

Ass we interpret his

electricians who do not

since he assigned this

5 to dime store clerks.

We changed that to 6.
We changed that to 4.

chart, there was no number assigned to qualified

own their own businesses.

We assign0 the number

5,

number to carpenters and plumbers who do not own their

own businesses.
After each rater had

practice-rated a class from a previous study, the raters

discussed their differences

settled upon the above

of opinion, which were small in magnitude and number,

list of adjustments to the Warner scale, and proceeded

independently to rate all 428

LDT's and LNDT-1's.

reliability of our ratings, we computed

coefficient between the

For purposes of estimating

a Pearson product-moment correlation

two sets of ratings.

It was .95, indicating high scorer

reliability.

Since we were not able to

procedure of flipping

use both sets of ratings, we adopted the

a coin to select one set, which was used.

The Cornell Deduction Tests.

"The Cornell Class-Reasoning

There were two Cornell !-1duction Tests,

Test, Form X," and "The Cornell Conditional-Reasoning

The former was administered

teaching the LDT's for the

in grades 4, 6, 8, 10, and 12;

administered similarly to all

In this chapter we shall

use of a correction formula.

to plus99.

as a pre-test immediately before

grade, and as a post-test six weeks after the comple-

tion of teaching to all students

(R- W/2 + 27).

simple

the latter was

students in grades 5, 7, 9, and 11.

report only total scores, which were computed asking

The score reported is rights minus

wrongs plus 27

Each test had 72 items making the possible range run ftom minus 9

This scoring procedure and many other features of the test will be

discussed in the next chapter.

Other scores and their analyses will be presented

in other chapters.
That completes the presentation

the discussion of the two

of th.. sources of data,

logic tests which will follow in the next chapter.

Given the previous descriptions

ments for securing data,

except of course for

THE FIGURES.
of the groups and methods and/or instru-

the following table ptovides a general picture of the sort

of subjects with whom we worked.

211 LNDT-1's, and 375 LNDT-2's.

Altogether there were 803 subjects, 217 LDT's,

Mean IQ for all subjects is roughly 114* and

mean

SES rating for LDT's and

LNDT-1's combined is 3.7. Thus our subjects are above

average, as one might expect for

a college community in the Northeast.

*We say !!roughly"

because purity of IQ tests is not achieved because of the LNDT-2'3.

TABLE 111-3. Basic Information on Our Subjects

Chronological
IQ

Gradf

SES
SD Mean
SD

Age(mos.)
Group Male Female Total Mean
SD Mean

04 UT

13 12 25 116.6 5.o 109.43.8.3 4.1

05 LDT

13 14 27 128.2 4.4 111.616.0 3.6

LNDT-1
LNDT-2

LNDT-1
LNDT-2

11
19

11
28

14
26

12,

23

22
47
26

49

119.0 5.0 109.019.1 3.1

117.2 5.9 108.314.2

130.5 6.5 111.519.8

128.7 6.7 105.111.9

06 LDT 14 11 25 141.1 3.8

LNDT-1

LNDT -2

07 LDT

LNDT-1
LNDT-2

23
18

14
11
23

11
26

34
44

14
27

25
50

10

24 152.0 4.5

11

LNDT-2

28

09 LDT

LNDT-1
LNDT-2

10 LDT
LNLT-1

LNDT-2

11
12
26

13
21

o6
11
14

24
49

4o

7.9 3.1

120.4

6.9

1.9

2.0

110.013.5

55.4 11.2 64.9 13.7

08

11 LDT
14
LNDT-1 11
L2J1T-2 19

lo

17
12
11
11

18

35

26

22

1.9

4.4 1.8

52.3 20.0 58.6 16.5

1.9

189.3 5.2 131.57.7 2.7

191.1 8.4 114.315.7

1.5 77.2 8.6 80.1 9.2

68.5 12.7 64.1 19.2

30

204.07.2

206.5 6.2

102.113.6

40.5 9.7 45.2 13.6

43.0 13.o 46.414.8
42.9 10.5 47.0 9.5

52.6 10.6 55.0 10.3

54.1 10.7 56.6 10.7

49.813.0 53.6 12.8

2.0

48.9 9.o46.3 10.9

50.8

3.5 1.9

108.111.6 4.2

SD

2.5 58.8 12.2 61.6 13.3

179.06.4 109.413.8 4.2

182.3 8.1 104.916.6

199.84.3 116.011.9

SDI

Post-test

Mean

52.6 13.3 56.o 14.8

15 07 22 190.7 6.o 119.710.6 3.7 2.0 69.6 11.7 77.9 12.5

18

2.1 54.3 16.9 56.8 17.1

3.9 1.9 67.7 11.5 71.5 13.9

164.2 6.5 113.910.1 3.6

169.2 7.3 108.316.5

ConditionalReasoning

Poet-test Pre-test
Mean
SD Wan

1.9 46.4 14.3 45.8 16.0

2.3 44.3 14.9 46.2 17.5
43.4 16.3 47.2 15.1

152.4 4.8 118.78.32.8 1.7

153.7 5.4 113.713.5

17 183.49.o
23

122.2

Pre-test
S

rean I

4.6 2.1

141.4 4.0 110.514.1 3.8

142.6 5.8 113.615.6

08 Lin 11 16 27 163.5 3.9

LNDT-1

114.013.4

4.0

Claes Rea oning

11.655.0

13.1

52.3 11.3 55.1 10.6

63.710.5 80.716.4

16.3 58.5 11.6

53.5 12.9 56.411.7
5)..9

TABLE III-3. Basic Information on Our Subjects - cont.

Chronological
Age(mos.)

IQ

SES

TotalLDT

Grand
Total

08

19

09

12

17
31

117 100 217

LN1YT-1 109 102 211

LNDT-2 196 179 375

422 381 803

212.2 2.9 134.2 11.4 3.0

218.5 9.5 105.217.6 3.0

.4..

116.8

3.8

113.8

3.7

116.3
108.4

3.5

Conditional Reasoning

Post-test Pre-test Post-test

SD Mean SD Mean SD Mean SDMean SD

1.9 80.4 12.8 87.2 8.o

Grade Group 1Male Female Tbtal Mean SD Mean SD Mean

12 LDT 12 12 24 210.8 4.0 127.816.9 3.1
LNDT-1
LNDT-2

Class Reasoning

Pre-test

1.5 81.9 8.1 83.5 15.6

62.6 15.3 62.0 18.6

CHAPTER IV.

The.Cornell Deduction Teats

In order to measure our subject's knowledge of logic, we found it necessary to develop our own instruments, there being no instruments available that

suited our purposes.

Most existing instruments which are partly or wholly aimed

at testing for knowledge of logic do

attempt to secure separate

so on a rather global scale, and do not

measures of basic principles of logic. A key feature

of our approach is our interest in

that principles vary in the

specific principles of logic on the assumption

ease with which, and the level on which, they are

acquired.

Two tests Which do attempt separate

logic and which made at least

of mention here.

interpretation for specific principles

of

some attempt at comprehensiveness are vorthy

Unfortunately neither of these were usable, for reasons rich

we shall give.

In 1919 Cyril Burt published

for use with individuals

an instrument called "Graded Reasoning Tests"

as opposed to groups. He did in developing the instru-

ment attempt to make a comprehensive

types of logic.

This instrument provided Burt with data Which he used to

support statements about when children

1919).

coverage of basic principles of certain

develop certain logical abilities (Burt,

However we were not able to use his instrument

1. Time limitations required the

for

several reasons:

use of a group test.

Possibly Burt's

instrument is adaptable (Fairgrieve, 1921), but we did not try because of other

difficulties with it, given our purposes.

2.

It calls on the

(sometimes called

person being tested to invent and supply a premise

a "suppressed premise"). The trouble here is that any argu-

ment can be made valid by

supplying some premise.

tested individually it is difficult

Unless the individual is

to see how to find out just What premise

he was adding.
3.

The final product

is no score possible

vas held by Burt to provide an overall score. There

on a type or sub-type.

Often a given form of argument

is tested for by only one or two items.

IV-2
Granted that with individual testing, the

examiner can by asking the right ques-

tions often get a fairly good idea of the nature of a person's difficulty with

an item, but there is danger that something else will without detection affect

the response.

In a Ph.D. thesis completed in 1961,

ment that she developed in order to

among 6, 7, and 8 year olds.

investigate the degree of knowledge of logic

Her test appears to be satisfactory for adminis-

tration to groups, and is based

basic principles of valid arguments

Shirley Hill reported on an instru-

on an attempt at a comprehensive coverage of

in class and sentence logic, but it was still

not satisfactory for our purposes.

The test was built so that the most

important feature of the distinction

between a valid and invalid argumeat could not be tested for.

conclusions either followed necessarily or contradicted the premises.

Although arguments in which the

-class of invalid arguments,

conclusion contradicts the premises are a

sub

a more important sub-class is the group of argu-

ments in vhich the conclusion

does not follow, but also in which the conclusion

does not contradict the premises.

People are rarely trapped iv") thinking

that
an argument is valid in which the conclusion actually contradicts the premises.

The important distinction is

between a valid argument and between one which

soreone might be inclined to call valid, but which really is not valid.

rastery ,)f this distinction is

not tested for in this test.

The

Thus it to56 was

unsatisfactory for our purposes.

In this chapter we shall describe

into our instruments, and the

present information bearing
shall suggest some lines

the features that we deliberately built

rationale for having these features; we shall

on their reliability and validity; and lastly we

along which we feel further development is possible.

THE STRUCTURE OF THE TESTS

The two tests that

we developed are called "The Cornell Class-Reasoning

Test, Fbrm X" and "The Cornell

may

Conditional-Reasoning Test, Fbrm X".

They

be found in the Appendix to this report.

Each is a 72-item multiple- choice test

designed for use with any of the grades with which we were concerned (grades
4-12). Each item in virtue of its logical form was expected to play a role in

measuring a person's knowledge of

some principle or combination of principles;

in virtue of its content it was expected to contribute to one of the three con-

tent components that we built into the test. First we shall discuss the logical
form of the items; then we shall discuss their content.

The logical Form of the Items.

Each test was constructed in such

a manner that it tested for knowledge of

twelve principles or combinations of

each principle or combination of
set of six such items

principles. Six items were assigned to

principles. Henceforth we shall refer to each

as an 'item group'. There are therefore twelve item

groups in each test.

In the conditional reasoning test the list of principles or combinations

et.-bodied by the items corresponds

very closely to the list of basic principles

of conditional reasoning given in

combinations of principles

Chapter II. In order that we could have some

we left the twelfth one off, on the assumption

it is less often needed than

combination with another.

that

the others; and used the eleventh one only in a

Principles One through Ten, however,

groups exclusively assigned to them.

each had item

Table IV-1 gives the selection and ex-

tent of edtbination of principles of conditional reasoning.

In the class reasoning test each

assigned to it exclusively,

assigned to it.

except Principle Three which had two item groups

Item groups #3 and #4 embody symbolized arguments that are dif-

ferent enough to justify being

Principle Three.
class logic.

of the basic principles had an item group

distinguished, though they are both covered by

Thus nine item groups are assigned to specific principles of

The other three item groups are assigned to combinations of the

principles or to double applications

specific nature of the assignment.

of the principles. See Table IV-2 for the

TABLE IV-1. Logical Form of and Answers to Items in "The Cornell Conditional Reasoning Test, Ppm X"

Item

Basic Form, Which was used - for

the first two concrete familiar
items (CFI and Answer

the symbolic item

Princi- (SY and the suggestive

CFI:

Group ple(s) item (SU)

Answer

for

Basic

to

Fbrm Fbrm of CF3 CF3

If p, then q Yes

If p, then q No

: not q

: q

Form of CF4

(Same answer as basic

form)

not p

If not p: then q
:

If p, then q
not p

3
-.1-

1-4

4
5

6
7
8

3
4

5
6

7
8

basic form

: not q

If p, then q

Maybe

Same as

Maybe p

If not p, then q

Maybe

basic form

If p, then q No

If p, then q Yee

: p

: not q

If not p, then q

not q
: p

not q
: not p

not p

not q

If not p, then q

CF' CF2 CF3 CF4 sl* Su*

7 40 27 14

19

If q, then r
: if p, then r

11

24 32

37 3u 41

8 35 29 16 22 39

: not p

basic form

If p, thenq

Maybe

: p

No

p only if q Yes
P

Same as

basic form

: if q, then p

p only if q

basic form

if not p, then r

p only if q
p

61

: if not q, then p

Maybe If not p, then q

p only if q Yes
not q
: not p

73

If not p, thenq

If p, then q Yes Same as Yes If not p, then q 46 69 74 56

: if not q, then not p

31

9 13 26 18 34 23

If p, then q Yes Same as Yes If q, then r 45 55 66 52 49

not q

Maybe Same as

Item Numbers

If q, then notp

not q

not p only ifq

:

not p

No not p
not p only ifq

44 57 77 70

50

59 64

12 21 42 25 15 36

10

17 20 33 38

28

: q

: not q

: q

TABLE IV-1 cont.

Basic
Fbrm,
which
was
used
for the first two concrete
familiar items (CF1 and

Item Princi-(SY

the symbolic item

and the

CF23,

suggestive

Group ple(s) item (SU)

10

1&5 If p, thenq
or 1

Answer

for
Answer
Form of CF4
Basic
to
(Same answer as basic
Form Form of CF3 CF3
form)

No

If q, then r

: r

10 p, if and only if q No

p, if and only Yes

ltell If p, then q

Same as

pot p
: q

12

Yee If notp, then q 43 51 62 72

If q, then r

applied If q, then r
twice : not r

11

If p, thenq

Maybe

r only if q

if q
not p
: not q

basic form

not p
: not r

47

not p, if and only if

: q

Maybe If notp, then q

: r

Note: The symbol ':' is used to introduce

Item Numbers
CF1 CF2 CF3 CF4 SY SU
67 76

the proposed statement.

r only if q
not p
: r

* 'CF' stands for 'concrete familiar', 'SY' for 'symbolic', and 'SU' for

'suggestive'.

48

54 63 58

78 6o

53 71 68

65 75

TABLE IV-2. Logical Form of and Answers to Items in "The Cornell Class-Reasoning Test, Form X"

Basic
Form, which was used
for the first two concrete

familiar items (CF1 and Answer

Form of CF3 Form of CF4
Item Princi- CF2), the symbolic item for
(Sy) and the suggestive Basic (same answer as (same answer as
Form
basic form)
basic form)
Group ple(s) item (SU)

All As are Bs. No X is an A At

: At least some As are

not bs.

least some As are Be.

: X is not an A. : no As are Bs.

5
6
7

CF1* CF2 CF3 CF4 SY* SU*

8 35 29 16 22 39

2 All As are Bs. Yes All As are Bs. At least some AS are Bs. 7 40 27 14 19
All Bs are Cs
: All As are Cs.

X is an A.
: X is a B.

All Bs are Cs.

: At least some As are Cs.

As are Bs. Maybe All As are Bs. At least some As are Bs.
3 3 All
: All Bs are As.
X is a B.
: All Be are As.
4

Item Numbers

11

: X is an A.

24 32 37 30 41

All As are Bs. Maybe All As are Bs. At least some As are Bs. 9 13 26
3 All
Cs are Bs.
X is a B.
All Cs are Bs.
: At least some Cs are As.

: X is an A. : At least some As are Cs.

: No Bs are As.

X is an A.

4 No As are Bs. Yes No As are Be. No As are Bs, and there

As are Bs. No
5 All
No Cs are Bs.

31

18 34 23

10

17 20 33 38 28

All As are Be. At least some As are Bs. 12

21 42 25 15 36

are Bs.

: X is not a B. : At least some Bs are

not As.

: At least some As are Cs.

X is not a B. No Cs are Be.

: X is an A. : All As are Cs.

No Cs are As.
: No Cs are Bs.

X is not an A. No Cs are As.

: X is not a B. : At least some Cs are

6 All As are Bs. Maybe All As are Bs. All As are Bs.
not Bs.

44

57 77 7o 59 64

TABLE IV-2 cont.

Basic Form, which was used
for the first two concrete

Item Princi-

familiar items (CFI and Answer

Form of CF3 Form of CF4

Basic (same answer as (same answer as

CF2), the symbolic item for

(SY) and thesuggestive

Group ple(s) item (SU)

8
No As are Bs.
7
No Cs are Bs.

Form

Maybe No As are Bs.

: At least some As are Cs.

9

10
11

8
3.848

2
applied
twice

12

2 & 8

or 8

applied
twice
or

2,4&5

All As are B.

: All non-Bs are also

non-As.

All As are Bs.

All non-Cs are also

non-Bs.
: At least some As are
not Cs.

All As are Bs.

All Cs are De.

All Bs are Cs.

X is not a B

: X is an A.

Yes

All Bs are Cs.

No DO are Cs.

All As are Bs.

: No De are As.

basic form)
CFI* CF2 CF3 CF4 SY* SU*
At least some Asare 48 53 71 68 65 75
not Bs.

I.

At least some Csare

not Be.

: At least some As are Cs.

All As are Bs. All As are Bs, and there 45

X is a non-B.

are non-Bs.
: X is a non-A.
: At least some non-Bs

No

are non-As.

All non-As are At least some Asare Bs.

non-Bs.

X is a B.

: X is not an A.

No

All non-Cs are non-Bs.

Yes

55 66 52 4* 73

43 51 62 72 67 76

: No As are Cs.

All Bs are Cs. At least some As are Bs.

All As are Bs. All Cs are Ds.
X is an A.
All Bs are Cs.
: X is not a C. : No As are Ds.

: At least some As are

not Bs.

basic form)

All Bs are Cs. All Bs are Cs.

X is not a C. No Ds are Cs.

All As are Bs. At least some As are Bs.
: X is not an A. : At least some As are
not Ds.

47 54 63 58 78 60
46 69 74

is used to introduce the proposed statement.

Note: Thesymbol
* 'CF' standsfor 'concrete
familiar', 'SY' for'symbolic', and 'SU' for'suggestive'.

56

61

50

IV-8
An additional logical principle

Which is often invoked in both tests is the

principle of double negation: that

two negatives make a positive.

Ibcplicitly the

principle goes as follows:

statement is true, and if

"If it is false that a statement is false, then the

a statement is true, then it is false that it is false."

Now there are qualifications to

ment of it holds in the

principle to be tested for in
To the extent that

be added to this principle, but this simple state-

cases in which it is needed.

We do not consider this

this test, but rather assume that it is mastered.

our assumption is false, the test is also a test for this prin-

ciple.
The Three Content Components.

In a frequently-mentioned study

tent and attempted to

Wilkins (1928) specified four types of con-

see which kind made syllogisms more difficult.

selected her first, second, and

fourth types for our tests. E r four types

Concrete familiar, in which the

a.

and qualities with which

We have
are:

content mentioned is concrete articles

the subject has been associated.

However, no

statements

are made which the subject has

reason to believe to be true or false, because the

specific objects referred to

are notkaown by the subject.

For example instead

of

saying, "All cats

are black", a statement with which the subject is likely to

disagree, a statement might

likely to meet with agreement
sibility of disagreement,

be, "Mary's cats are black".

The latter is less

or disagreement, although there is still the pos-

if the subject identifies the Mary of the statement

with some Mary he knows.

b.

Symbolic, it which symbols like 'x',

places, instead of terms

that refer to particular

tence using symbolic

c.

objects.

are used in key

An example of a sen-

content is the following: 'All A's are B's.'

Unfamiliar, in which the

subject is expected to

'A', and

basic terms are scientific terms with which the

be unfamiliar,

or are "nonsense words invented to sound

like scientific terms" (p. 13).

a sentence with unfamiliar content.

For example, 'All hentras are globiculous', is

IV-9

d.

Suggestive, in which the content is familiar, but the truth status of this

content is known by the subject.

Furthermore, "the truth or falsity of these state-

ments gonclusiongwas at variance with their validity." (p. 13).

In other words

the truth status of the conclusion was different from the validity status of the
argument.

A person might be tempted to makk an argument valid because he agrees

with the conclusion, or invalid because he disagrees with it.

In our tests the three of Wilkins' four content components that we used are the
ones we feel a person is most likely to encounter in realistic reasoning situations
in which his reasoning ability will make a difference. Our three content components
are the concrete familiar, the symbolic, and the suggestive. We do not think that
what she calls 'unfamiliar' content is likely to be encountered in situations that

make a difference.

If a person meets arguments with such content, he will

generally
not know what to do with the conclusion anyway.
This is not to say that arguments with such content are useless in teaching,
for they can be very helpful in isolating questions of form, and they can be useful
in arousing interest, if handled properly.

Furthermore it might well be that they

are good indicators of grasp of form. But ultimately our interest is a practical
one, so we, given the testing time limits forced upon us by the situation, did not
include this type of content in our tests.
The six items in each item group were assigned to the three content components

in the following way.

Fbur of the items had concrete familiar content, a fifth had

symbolic content, and a sixth had suggestive content. There were some formal

variations among these, incidentally, which ve shall describe next.

A basic logical symbolic form was assigned to each item group.
Column 3 in Table IV-1

and Table

IV-2.

It appears in

This basic form was used for two of the

concrete familiar items, the symbolic item, and the suggestive item. The third

and fourth concrete

familiar items embodied alight formal changes, in

order to pro-vide

greater variety in the representatives of each principle or combination.

IV-10
Because class and conditional

reasoning differ in their basic structure, these

variations differed.
In the conditional reasoning
the conclusion, unless

there were good reasons not to do so. That is, if the

conclusion called for by the

is negative; and if the

basic symbolic form is positive, the one provided

conclusion called for is negative,

vided is positive; unelss,

1. If the basic form is

then the one pro-

as we said, there are good reasons not to do so.

What are the reasons?

temptation to judge it

test the third concrete familiar item negated

There are two:

invalid, then denying the conclusion reduces the

valid, making the item considerably easier. So for those

four item groups in which

the basic form is invalid,

the third concrete familiar

item conforms to the basic form.

If the conclusion is

becomes awkward to state.

he will walk to school"

school", which is

Note that the denial of "If Sam misses the bus, then

is not "If Sam misses the bus, he will not walk to

not awkward to state.

case that if Sam misses the bus,

Hence the third concrete
ti

itself a conditional statement, a denial of it

Instead the denial is, "It is not the

he will walk to school",

which is awkward.

familiar items conformed to the basic logical form in

additional cases (Groups 5 and 6).

The fourth concrete

the basic form by

having one of the basic sentences negated throughout, and by

having premises in

different order, when there are two or more premises.

11/-2 presents symbolic

In class reasoning
class membership
exclusion.

familiar item in conditional reasoning varied from

Table

representation of this variation.

the third concrete familiar item contained at least one

statement in the premises instead of all class inclusion and

A class metbership

statement asserts the membership or nonmedberahip

of one thing in a class, whereas a class

about the relation between classes.

inclusion or exclusion statement speaks

The fourth concrete familiar

item in the class reasoning test contained at

least one statement of partial inclusion.

The composition of each

item group is presented in Tables IV-1 and IV-2.

Item Format.

We sought an item format

Provide a way for

does not follow,

that would meet the following criteria:

a subject to show that he know that a proposed conclusion

even though it does not contradict the premises.

Avoid technical language,

yet ask:whether an argument is valid.

Provide more than two alternatives

Provide fairly equal numbers of each type of answer.

Allow for separation of

in multiple choice form.

different items of the same logical form.

Be understandable to fourth

graders without insulting twelfth graders.

Avoid confusing truth end validity.

Not require elaborate directions.

Not require an elaborate scoring procedure.

Minimize irrelevant errors.

After much experimentation

the subject to

and discussion we settled upon a format which asks

suppose (a) certain premise(s), and to decide whether, on the basis of

the supposition,

a fUrther thing would be true.

the 'proposed statement'

and the original supposition the 'supposed statement(s).

There are three possible

answers, 'Yes', 'No', and 'Maybe', each of which is ex-

plained in the directions

explanation of each

has gone before.

by means of six sample problems.

PUrthermore a brief

response appears at the top of every page.

The response ' Yes'

follows necessarily.

We shall call this hirther thing

indicates that,the subject thinks the proposed statement

'Bo' signifies that the proposed statement contradicts what

'Maybe' means that the proposed statement neither follows neces-

sarily nor contradicts

-- that itu truth is not necessarily determined by the estab-

lishment of the truth of the premises.

IV-12

WO introduced the 'No'

response partly in order to make possible three choices

instead of the two suggested by the distinction between validity and invalidity.
We vented three choices in order to lessen the attractiveness of wild guessing and to

lessen the chances of its success.

We were able to introduce this 'No'

response as

a correct answer by taking conclusions to valid arguments and then denying them.
Fbr the 'No' response, the proposed statement is then a denial of a conclusion that

follows necessarily from the supposed statement(s). A second reason for having

'No' response is that doing

so enabled us to avoid a preponderance of 'Yes' answers

over 'Maybe' answers.

In introducing this 'No'

risk.

response we were well aware that we were running a

There is an additional principle involved in each of these items: That the

denial of a statement which follows

necessarily from supposed statements can not be

accepted, if the supposition is accepted. This principle is elementary, but it must

be invoked.

Failure to invoke it, as contrasted with failure to know the basic

principle(s), could explain errors that are made.

possibility of error of measurement.

Thus there is an additional

We feel that the advantages gained are

worth
the risk.

The reader is referred to the copies

of the tests in the Appendix to see the

sample problems and initial directions.

the-page interpretation of the

Let it suffice here to present the top-of-

answers and an item from the class reasoning test:

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.

MAYBE It maybe true or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

7.

Suppose you know that

All the cars in the

garage are Mr. Smith's.

All Mr. Smith's cars are Fbrds.

Then would this be true?
All of the cars in the garage are Fords.

A.

YES

B.

P3

C.

MAYBE

IV-13
In order to avoid possible confusion separate answer sheets were not used. Subjects
circled an answer on the right.
Although there are many strong arguments on both sides we directed students not

to guess wildly and used a scoring formula. We did this because, in addition to
the standard arguments, we feel that a critical thinking test should not encourage
wild guessing, end that part of being a critical thinker is knowing when

you do not

know.

The following direction

appears on the front of both tests:

DO NOT GUESS WILDLY. There is a scoring penalty for guessing wrong. If

you think you have the answer, but are not sure, mark that answer. But
if you have no idea, then skip the question.
The scoring formula which we used is: Score = R - W/2 + 27.

The 27 was added in

order to make negative scores unlikely.

Item Arrangement.
The first six items are sample items in each test. They are drawn from other

types of logic than class and conditional reasoning. Items 7 through 78 in each test
are divided in two parts, although the subject is not informed, of this fact. The
first part contains the item groups for the six easiest forms, as determined in our

tryouts.

The second part contains the item groups for the more difficult forms.

Thus it might be possible to cut the test in half for testing younger children or
for other purposes.
Since our item groups are generally numbered in recommended teaching order, and
not necessarily in order of increasing test difficulty,

some of the groups in the

second half of the conditional test represent principles in the first

six on the

teaching order list.

Items were arranged so that no two items from the same item group appeared on the
same page, and so that the distribution of answers would appear fairly random

to the subjects.

The specific assignment of items to test item nuibers is Mown in

Tables IV-1 and IV-2.

11c14

OPERATIONAL DERUEETIONS OF ')(5PI:PS:MUM'

Our basic interest is

in determining Whether or not a given principle of

logic has been mastered.

a determination, we developed
correct only to the

our own, realizing of course that our results are

extent that our criteria are satisfactory and that the ques-

tion of the satisfactoriness

It is partly also

Since there are no'established criteria for making such

of our criteria is not strictly an empirical matter.

a matter of the meaning of the. term 'mastery' and of the meaning

of the principle involved.

Since there is

no outside criterion, one principal empirical question is

whether the criterion

measure). Another

operates reliably (or Whether it is based upon a reliable

empirical

ments are in agreement

ments must be based

about meaning require a great deal of attention

(We did seek

not go outside our own staff for such ex-

outside opinion about

the validity of the

total

present that later.)

What a person interested in

in very specific terms and decide

our research must do is to examine our criteria

for himself whether they reasonably conform to

the meaning of the phrase, 'mastery of

aspects of our criteria

judgment rather than

the form of operational definitions. The operational

fiable observations

as follows

(abstracted from/Innis,

a specification of an operation performable

this specification

comes a statement of

In short the non-empirical

empirical investigation.

definition structure that we used is

First there is

2:principle'.

are not arbitrary, but they are matters of intelligent

Our ciiteria took

investigator;

tothe

particular criterion and there is a limit to the amount of ex-

pert judgments.

1964c):

judg-

With the judgments of experts, but of course their judg-

pertise that can to exercised, we did

tests and vill

the criterion-based

upon the non-empirical meaning features mentioned above.

Since these judgments

operation of the

question might be whether

appears in a conditional clause.

by an

Next

some sort of relationship between the making of speci-

and the application of the term being defined.

effect the operational

In

definition gives an incomplete empirical interpretation

of the meaning of an abstract term, because the relationship is claimed to hold

when the operation has been performed, but no claim is made about the status of
the relationship at other times, or when other operations have been performed.
Furthermore the operational definition should have some explicit qualifiers
like 'probably' and 'under standard conditions' if there is a reasonable chance
that something might go wrong in the measuring process.

Since there ordinarily

is such a chance in dealing with human beings, it is best to include such qualifiers in our operational definitions.
First we shall give an informal presentation of one of our operational defi-

nitions; then we shall give the same definition formally; finally we shall discuss some dangers in our approach.
Put informally the operational definitions of mastery of one of our principles go as follows: Given that we have administered. the "Cornell Conditional
Reasoning Test, Fbrm X" under standard conditions and have scored it according
to the key given in Table IV -l.

Then if a subject gets right at least five of

items 8, 16, 22, 29, 35, and 39, we can be fairly sure that he has mastered
Principle & (Given an if4hen sentence, the denial of the then-part implies the
denial of the if-Tart).

FUrthermore if he gets ftwer than four of those items

right, we can be fairly sure that he has not mastered Principle #4.
One might choose to regard the foregoing as two operational definitions or
as one.

It does not matter.

definitions.

Fbr convenience we regard it as two operational

The distinction between them appears in the following pair of for-

mal operational definitions of the same principle.

The first gives a sufficient

condition of mastery, given the operation; the second gives a necessary condition

of mastery, given the operation:

If
Y is given
"The Cornell
Reasoning
standard
conditions;
then if Conditional
Y answers correctly
at Test,
least /bra
five of under
items
8, 16, 22, 29, 35, and 39, Y has probably mastered Principle IR.

IV-16

X" under

If Y is given "The Cornell Conditional Reasoning Test,

Fbrm
Principle
conditions; then it is probable that Y has mastered

if Y answers correctly at least four of these items:

on
29, 35, and 39.
These two operational definitions of 'mastery of
for estimating the per cent of students at each
mastered Principle #4 of conditional reasoning.

standard

8, 16, 22,

Principle #41 served as our criteria

grade level who had and had not

We used similar operational
corresponding

definitions for each of the other principles for which we had a

item group.
It is important to note that the above pair of operational
not specify a decision for a subject who correctly answers

definitions does

exactly four of the items

If a subject correctly answers at least five, then he is

judged

There is a reason for,, framing the operational definitions this way.

we think are borderline cases, so we have accommodated

to this fact.

Such people

the operational definitions

We do not desire to be committed, even with the qualifications

exist in the definition, when a person answers correctly

exactly four of an item

group. We do not want to say that he has mastered the principle, and we do
to say that he has not mastered the principle.

that

not want

We want to leave the question open.

One of the dangers in this operational specification of the meaning of

'mastery

of v principle' is that we might be taken to have exhausted the meaning of the

phrase with our necessary and sufficient conditions; that is, we might be taken to
have equated the meaning of this phrase with a finite set of

"behaviors", as people

SD often say. A careful examination of the logical form of our definitions should
make it clear that we are not doing this.

committed to saying that no other test of mastery of these principles can attain
validity simply by virtue of the meaning of 'mastery of X principle'.

IV-17

Furthermore our qualifying terms, 'probably' and 'under standard conditions'

protect us from being unalterably committed to the view that a subject who by wild
guessing answers correctly five of the six items in a group knows the principle.
A behavioristic interpretation would commit one to such a view. Our judging
procedure will classify such a person as one who has mastered the principle, but
the looseness of our definition enables us to avoid such a commitment.
In summary, even though operationism is often associated with behaviorism,
we want it to be clear thpt our operational definitions should not be so associated.
A second danger in our approach arises from the number of items in an item
group: six.
struck.

The use of six items instead of more represents a compromise that we

Pressures that operated to keep down the number of items were the need to

avoid taking too much school time of our subjects, and the desirability of including
a variety of principles and combinations thereof. Given these pressures, together
with the fact that our items are so directly and clearly related to their principles,
a study of our tryout test reliabilities made us willing to utilize part scores based

upon six items.

This use of these part scores was fairly successful, from the point

of view of reliability.

For a discussion of the reliability of parts and wholes at

each grade level, see the next section.

A third danger in our use of these operational definitions is the possible

con-

fusion with respect to those item groups which represent combinations of princ4''.es
and with respect to Item Groups #3 and

#4

of the class reasoning test, both of which

embody Principle #3.

We shall deal with the latter prrblem first.

Since operational definitions do

not present complete interpretations of the meaning of terms, the double pair of
operational definitions of 'mastery of Principle #3' (class reasoning) supplement
each other.

Does this imply that we must have complete agreement between these two

item groups?

No, it does not.

Systematic and random variations, errors, and mis-

understandings are bound to slip in. One does not even expect (nor achieve) complete

I V 18

agreement among thermometers of different types, although they are supposedly

measures of the same thing, temperature, and are mentioned in operational definitions of 'temperature'.
A glance of the difficulty index development on class reasoning Item Groups
#3 and #4 (Table

p. 31) shows that these two item groups are not vastly

different from each other, although Item Group

f4

appears more difficult,

probably because of its greater complexity.

The other problem, that of confusion resulting from the combination of

principles for some item

groups, maybe stated in two parts:

The mastery of 'which principle is being defined?

When alternative combinations of

judgment for a given item

principles can be used to justify a cor-

group, 'which of the alternatives is being defined?

The first part of the problem

can be answered by stating that it is

'mastery of the combination of Principles X

The second part of the problem is

satisfaction.

rect

and V that is being defined.

not me we have been able to solve to our

It is probably the result of the redundance, mentioned in Chapter

II, that is built into the list

of principles.

This redundance is the cost of

intelligibility and utility for teaching.

One way out is to

say that the necessary condition operational definition

can be used to show, if the

met, that probably

necessary condition of at least four correct is not

no one of the combinations has been mastered.

the sufficient condition operational

ficient condition definition of at

definition can be used to show if the suf-

least five right has been met, then at least

one of the combinations has probably been

a verbal interpretation, but the test is

dition instances,

Furthermore

mastered.

Thus the test results have

not serving, in these sufficient con-

as a vehicle of interpretation of particular principles or

combinations thereof.
These operational interpretations
empirical dats and the
other

represent a new approach to the relation

concept of mastery. We would like to see it tried

areas, and are using it because we feel it has promise.

between

and tested in

IV-19

RELIABILITY

Because of the principle and content components of the tests) it did not

seem appropriate to compute a split-half reliability estimate.* There just

was no way to divide the tests into two equal halves. So we are reporting a
test-retest reliability estimate based upon the groups that did not receive

logic instruction, the LNDT-1's and the LNDT-2's. They are lumped together for

these purposes.

The administrations of the test were about ten weeks apart in

the spring of 1964.

These estimates are reported in Tables IV-3 and IV -4 by separate grades

for total score, component

scores, and item group scores.

Accompanying these

estimates are the means and standard deviations on the pre-test for each score
on which a test-retest reliability estimate is presented.

We had expected to find that the

when the mean scores were near their extremes (e.g., an.item group score near 0 or

near 6). That is why we report-the

means and standArd deviations in:these

tables.
However, a visual inspection of the reliability estimates suggests that

they are relatively independent of the

difficulty of the item groups.

This

impression is supported by the tables of standard errors of the item group

scores (Tables IV-5 and IV-6).

The standard errors hover around one, even though

the difficulties of the item groups vary considerably.

In order to make a rough check

on the relative size of the standard errors, we

divided them into two groups and computed the mean standard error for each

group.

One group consisted of those for which the mean score on the item group

was within 1.5 points of 4, that is, between 5.5

sisted, of the rest) that is those

and 2.5.

The other group con-

for Which the mean scores were distant from

our mid-cutting point.

Fbr conditional reasoning these mean standard errors

were

We prefer to speak of

a reliability estimate, rather than a reliability,

because we are securing only estimates of the actual reliabilities.

TABLE IV-3. Test - Retest Reliability Estimates on the ConditionalReasoning Test

Grade 05 11=76 Grade 07 Nic75
Grade 09 N-64
Grade 11
All Grades

Type of Scoring
Score

Total

Formula

(R- W/2) +27

Content

CF

R
R
R

SY
su
Item

Group

Poss.

Score Mean

99

It

R
R
R
R

6
6
6

5
6
7

R
R

8
9

R
R

lo

11

12

6
6

6
6
6

Using Fisher'sZ.

SD

42.7 11.4

.76

Mean

SD

51.2 12.4

r Mean

.65

51.8

SD

Mean SD

11.3 .78 53.5 14.0

Mean*
r

75

48 23.6 4.9 .6o 27.1 5.5 .63 27.5 5.1 .62 27.8 6.7 .74 .65
12
5.8 1.8 .48 6.6 1.9 .5o 6.6 2.0 .63 7.0 1.8 .48 53
12
4.7 2.2 .61 5.9 1.9 .4o 5.8 2.0 .62 6.o 1.9 .56 .55

1
3

Sax.

4.3 1.6 .59 4.5 1.3 .56 4.8

1.3 1.3 .50 1.8 1.6 .48 1.5
1.1 1.2 .51 1.7 1.4 .34 1.6
3.5 1.6 .54 4.o 1.5 .27 3.9
3.3 1.5 .55 3.9 1.6 .58 4.o
3.5 1.7 .63 3.8 1.6 .4o 3.7
.9 1.3 .54 1.6 1.7 .38 1.7
4.1 1.6 .56 4.9 1.2 .57 5.o

1.3 .63 4.4 1.3 .37 .54

1.6 .61 2.3 1.7 .60 .55
1.4 .57 1.9 1.4 .45 .47

1.4 .51 3.7

1.6 .61 4.1
1.9 .5o 3.4
1.9 .69 2.3
1.3 .56 4.9
4.2 1.4 .5o 4.6 1.2 .38 4.8 .9 .43 5.o
3.4 1.7 .67 4.2 1.7 .54 4.4 1.6 .7o 4.1
3.3 1.6 .62 3.9 1.4 .39 4.1 1.5 .57 3.7
1..6 1.5 .68 1.1 1.4 .55 1.o 1.1 .28 1.4

1.4 .31 .41

1.5 .57 .58
1.5 .44 .5o
1.8 .61 .57
1.4 .64 .58
.9 .4o .43
1.7 .44 .6o
1.3 .39 .5o
1.2 .33 .48

TABLE IV-4.

Test-Retest Reliability Estimates on the ClassReasoning Test

Grade 04 N=71 Grade 06 N=78 Grade 08 N=76 Grade 10 N=54 Grade 12 N-50 All
Grades
Max.

Type of Scoring Poss.

Score

Total
Content

Formula

(R-W/2)+27

Score

12
12

4
5

7
8

10
11

12

R
R
R
R

6
6
6
6
6
6

R
F

6
6

6
6

.63

38.2

.80

30.4 7.1 .85 30.8 8.3

4.9 1.7 .44 5.5 1.7 .36 5.3 1.9
4.5 2.4 .52 5.5 2.5 .70 6.2 2.8

25.5 7.6

SY

R
R

.66

.83

48

Group
1

17.9

r Mean SD r
71.4 12.0 .84 70.0 16.0 .86

43.7 15.6

R
R

SD

99

CF

su
Item

SD

Mean

Mean SD
53.3 14.9

Mean
.88 54.8

4.3 1.8 .60 5.3 1.1 .66 5.4

3.6 1.4 .43 4.2 1.3 .69
1.7 1.7 .64 2.5 1.8 .71

2.5 1.6 .45 2.6 1.4 .39

3.5 1.5 .48 4.2 1.1 .3C
3.6 1.6 .53 4.4 1.4 .44
1.9 1.7 .51 2.3 1.8 .59
2.3 1.8 .57 2.6 1.7 .64
2.9 1.7 .50 3.4 1.5 .49
2.4 1.5 .32 2.6 1.4 .26
2.7 1.6 .44 3.4 1.5 .40
3.6 1.4 .32 3.7 1.3 .44

* Using Fisher'sZ.

.27
.53

Mean SD

5.1

.79 37.9 7-4 .85

6.9 1.9 .70 7.0 1.7 .62
8.3 2.2 .61 7.7 2.7 .73

5.7 0.7 .44 5.7

4.2 1.4 .54 5.2 0.9 .41 5.1
2.9 1.6 .62 4.0 1.7 .68 4.4
2.8 1.4 .61 3.5 1.3 .43 3.9
4.1 1.2 .33 4.7 0.9 .38 4.5
4.4 1.4 .49 5.2 1.0 .72 5.1
2.2 1.7 .48 3.7 1.7 .60 3.8
3.0 1.8 .56 4.2 1.4 .37 4.3
3.2 1.6 .38 3.9 1.3 .48 3.8
2.8 1.6 .30 3.5 1.4 .41 3.1
3.4 1.7 .40 4.9 1.1 .52 4.4
4.0 1.4 .33 4.8 1.2 .36 4.4
1.1 .47

0.8

1.2
1.7
1.3
1.2
1.0
1.7

.77

r*

.83
.79

.50
.63

.60

,55
.69
.52
.41 .38
.38 .52
.65
.76
.68

.46 .53

.45
.48
1.7 .5o
1.5 .51
1.4 .58
1.7
1.6

Mean

.52
.47

.36
.46

.41

IV-22

TABLE /V.5.

Standard Errors of Item Group Scores

Item
Group

N = 76
Grade 05

1.1112

.92

N =75

Grade 07

Ak.

- Conditional Reasoning

64
Grade 09

51
Grade 11

i9

Jai

1.15

1.12

1.08

.84

1.14

.92

1.04

La

1.28

.98

1.16

1.01

1.04

1.00

.98

1.03

1.24

1.34

1.12

1.05

_1.12

,,

.88

1.34

1.06

.79

.86

.84

.99

.94

.74

.70

10

.98

1.15

.88

1.27

11

.99

1.25

.98

1.02

12

.85

.94

.93

.98

Note: If the standard

error is underlined, the mean score on this item group

is within 1.5 points of 4; that is, between 2.5 and 5.5

TABLE IV-6.

Standard Errors of Item Group Scores - Class Reasoning

N = 71
Grade 04
1.25

N = 78
Grade 06
1.37

1.14

.64

1.o6

.72

4
5

1.19
1.08

N = 54
Grade 10

1.92

1.37

1.4o

.80

.52

.38

.95

.69

.71

.87

.98

.74

.98

.71

.92

.53

.79

Grade 12

1.09
.92

1.10

1.05

1.19

1.15

1.18

1.02

1.20

1.07

10

1.24

1.20

11

1.20

1.16

1.15

.97

1.00

1.081.25

1.111.26

1.26

12

N= 50

N = 76
Grade 08

.94

1.16

1.081.20

1.32

.76

1.05

1.15

.96

.91

Note: If the standard error is underlined, the mean score on this item group is
within 1.5 points; that is, between 2.5 and 5.5.

practically identical, 1.01 and 1.03 respectively.

For class reasoning the

mean standard errors were 1.05 and 1.32 respectively, giving a difference in

the expected direction.

But this is still insufficient to warrant the claim

that there is an appreciable decrease in error around the crucial cutting

points.
Thus we were unable to find grounds to support our expectation that the

item groups would be more sensitive around the mid-range than at the extremes.
As is expected, the reliability estimates for the total scores are the
largest, running from a low of .65 on the conditional reasoning test for seventh

grade to a high of .88 on the class reasoning test for sixth graders. Correlations on the class reasoning test tend to run slightly higher than on the
conditional reasoning test, the mean total score correlation on the conditional

reasoning test being .75 and on the class reasoning test .83.
The component reliability estimates are not so high as those for the total
scores, presumably because of the fewer number of items. The concrete familiar

components each consist of 48 items altogether, the reliability estimates running

from .60 on the conditional reasoning test for fifth graders to .85

on the class reasoning test for sixth and twelfth graders

.65 for conditional reasoning and

-- with a mean of

.79 for class reasoning.

The symbolic components, each consisting of twelve items, had correlations

running from .27 for grade eight

ten on the same test.

on the class reasoning test to .70 for grade

She mean correlations are .53 for conditional reasoning

and .50 for class reasoning.

On the suggestive component estimates of reliabilities

ranged from AO

on the conditional reasoning test for seventh graders to .73 on the class

reasoning test for twelfth graders.

The means are .55 on conditional

reasoning

and .63 on class reasoning.

The item group test-retest correlations ranged from .26 for Item Group #10
of the class reasoning test at grade level six to .77 for Item 0roup #1 for grade

IV-25

twelve on the sane test.

The mean item group correlation for the conditional

reasoning test is .52 and .50 for the class reasoning test.
Since these item group reliability estimates are based upon only six items
apiece and are for snores to be used with groups rather than individuals, they

are encouraging even though they average around .5. The total score reliabilities
are of course more satisfactory, but have the disadvantage of representing a
composite score, rather than a fairly pure score.

VALIDITY

Several approaches to the validity of the tests suggest themselves:

Examining the items and the procedures used to develop them to see if these
items do represent the subject matter that they are supposed to represent;
asking outside experts to judge whether in their opinion the test does test for

what it is supposed to test for; determining the test's correlations with

familiar measures and seeing if these relationships make sense; examining the
internal features that empirical investigation shove the test to possess; seeing

hcmrnudh sense one can make out of the investigations which depend upon the

test; and seeing the extent to which the

valid measure of whatever it is that

section we shall report

nation of validity.
validity, and sometimes

test correlates with some established

the test is supposed to measure.

In this

on all but the last of these approaches to the determiThe last approach, which sometimes appears as concurrent

as predictive validity, we shall not use, because there

is no established outside criterion for measuring that we vent to measure.

Content
Part of our argument for content validity rest, upon the procedures which ve

used to develop the test.

We made a careful study of the field of logic, and

examined different kinds of specific

examples of logical reasoning in everyday

life: newspaper editorials, I.S. Supreme Court opinions in support of decisions,

an auto vedhanics handbook.

and

We found no calms of class or conditional

reasoning which were not covered by

our principles,

The principles, as was

IV-26
pointed out in Chapter II, are basic and comprehensive.
Itirtheriore we engaged in continuous consultation with members of the
Cornell Philosophy Department who are interested in logic. Professor Keith
Donnellan, who has the major responsibility in the Philosophy Department for
knowledge of logic among graduate students, has examined and taken the tests,
and judged them to be valid.
Lastly the content of the items together with the instructions is evidence
of the test's content validity.

Each item calls for an answer to the question

which is the central question in logic:

Does this statement follow necessarily?

And the answers depend upon whether one has the right answer to that question in

each case.
principles.

Furthermore the correct answers to the items are justified by the

There is an obvious relation between the items and the principles,

as can be seen by an inspection.of the items.

The relationship goes both ways.

Not only do to principles bray the right answers to the items, but generally
speaking, getting five or six of an item group right loosely implies a knowledge

of the principles.

Admittedly this last claim is not onethat can be

provenOt
is one that we rest upon the intelligent judgment of informed interested people.

Construct Validity.
In looking at the construct validity of these tests, we are looking at

the degree to which they make sense in their internal and external relationships..

This makes our concern very broad, broad enough to include concern with content

validity as veil.

But since that has already been discussed, we shall limit

the present discussion to the other elements of construct validity.

a. Correlations with Familiar Measures.

intellectual trait, we expected to find

test scores and IQ scores.

Because knowledge of logic is an

a substantial correlation between logic

Furthermore it is partly a verbal intellectual trait

so we expected to find at least some relationship between logic test

scores and

socio-economic status (SW. Presumably knowledge of logic is something that

IV-27

increases with age (Burt, 1919; Piaget, 1950, 1958, 1959; Hill, 1961) - until

- so we expected a substantial correlation with chronological

age for our subjects when the various grades are grouped together.
expect to find any sizeable relationship between me and le%ic knowledge when

figured at each grade level separately.

Although we expected sex to be unrelated to logical ability, we were not

sure.

Burt (1919), Miller (1955), and Hill (1961) found no relationship, but

Sweeney (1953) found that men did significantly better on logic tests, though

verbal intelligence was controlled.

We obtained correlations with these familiar variables that were basically
in agreement with our expectations.
In Table IV-7, which is based on pretests for LDT's and LNET-1's

combined,

one can see, substantial correlations between IQ, grade held constant, and both
tests.

This is also the case for chronological age when grades are lumped

together.

The correlations with age by separate grade levels

are small and

generally negative, two of the nine being significant at the .05 level. This
is insufficient proof of a small negative relationship, but should there be one,

it would not be surprising.

In fact it would be rather interesting, for it

would

suggest that in a given grade the

more able logicians are the younger ones.

This

in turn suggests the explanation that inherent mental ability is a more signifi-cant
factor in logical ability than it is in grade placement. And this in turn fits in with
the substantial relationships that were obtained with IQ. These

speculations bear fUrther investigation.

There are generally small positive correlations between SES and logical
knowledge, six of the nine correlations being significant at the .05 level.

The correlations between the tests and sex are small and on either side a

zero.

There does not seem to be any relationship.

IV-28

TABLE IV-7.

Correlations Between Total Scores on the Cornell Deduction Tests

and Certain Familiar Variables: Chronological Age, IA., SocioEconomic Status, and Sex.

Grade

(Separate
Grades)

(Grades
Combined)

CA

CA

IQ

SES

SEX

.24

.00

.28

.17

.13

.17

.12

.09

.20

.10

z.g

.00

lat

!Ei

-.10

-.05

.26

-.09

Conditional Reasoning
..09

53

05

07

47

.61

..09

.64
..t

09

4o
48

-.24
-.05

188

-di

.70

122

11
5,7,9, & 11

...g

Class Reasoning
VT

=as

58

-.07

49

-.12

1E2

40

-.26

.26

41

.23

LE

.40

-.22

-.12

la

.26

-.00

o4
06

112.

./
.68

08
10
12

235
4,6,8,10, & 12

Notes:

These are based upon the pre-test scores of the LDT's and the UWI.l's.

A correlation that is significant at the .05 level is underlined.

Par SES, signs have been reversed for ease of interpretation. A
positive correlation on this chart indicates that those with high
SES's did better than those with low SES's.
Numerical assignment for sex gave boys a one and girls a two. Hence
a positive correlation suggests that girls werebettAarin that group.
Details about instruments used can be found in Chapter III.
The correlations for CA for grades combined are based only upon those
asibers of the LL 's who were used for computing the multiple regression
equations discussed in Chapter VI. For conditional reasoning, N = 64; for
class reasoning, N = 82.

Correlations were averaged via Fisher's Z.

IV-29

Item Analysis.

at each grade level

Difficulty and discrimination indices for each item

were computed.

percent of students taking

The difficulty index in this case is the

the test who answered the item correctly. A low

index signifies a difficult item.

The discrimination index

shows how successfully the item discriminates,

using in this case the total

secured
by separating out the top

score on the test as a criterion.

and bottom 27% of a group using the above criterion.

The percent of subjects in

the bottom 27% who correctly answer the item is -sub-

tracted from the corresponding

the result, the

It is

percent in the top 27%. In general the larger

more discriminating is the item, unless the difficulty indices

are near the extremes.

Tables A-1 and A-4 in the Appendix

indices for each item at each
in addition, for the LDT's

grade level for all subjects on the pre-test and

on the pre-test and the post-test. These indices

arranged according to item

presented as well, giving

the

natural

knowledge will

cultural

are

groups and the mean for each item group is

an idea of the comparative difficulty and discriminating

power of the item groups and perhaps

of

list the difficulty and discrimination

the principles. The discussion in Chapters

development

of

logic

knowledge

and

the

capacity

V and VI

for

logic

draw heavily upon these tables. .Here we'dhall make

only the most obvials comments

Tables IV -8 end IV-9 in this

detailed difficulty and

and present a few interesting means.

chapter provide something of a summary of this

discrimination ihformation, by giving the means on the

pre-tests for each item

group, component. and total at each grade level. This

summary, when it is examined from the

standing of logic, is quite

point of view of the development of under-

revealing, but that is the topic of the next chapter.

Let it be noted, however, that

ficulty levels of the various

there is considerable variation in the dif-

item groups, with a range of 14.1 - 87.6 on the

conditional reasoning test and

30.2 -

95.4 on the class reasoning test.

These

IV-30
TABLE IV-8.

Mean Difficulty and Discrimination Indices Jr b The Conditional Reasoning

Test) Based, Upon llre-Test of DDT's, LNDT-11s, and LNDT-2's Coabined.

Grade
N =

Mean Difficulty Indices

5
7
9
11
99
78
80
102

Mean Discrimination Indices

5
11
7
9
78
102
8o
99

Item
Group

71.4

74.7

77.3

78.8

1
2

22.7

27.4

24.8

35.3

17.7

26.8

31.5

35.7

55.7

68.7

59.6

65.o

56.7

67.9

66.3

76.1

56.5

66.2

57.7

65.3

14.1

24.1

28.6

43.2

68.6

80.6

77.9

85.7

70.8

78.3

78.6

87.6

10

54.7

71.2

68.8

77.6

11

54.6

66.o

64.2

66.2

12

26.3

17.8

19.4

21.6

CF

48.9

55.4

55.2

61.8

Sr

48.1

55.8

53.4

59.5

26.319.432.0

au

41.3

53.4

53.3

59.9

29.o

26.5

21.2

19.1

34.3

26.5

22.3

31.8

28.1

27.0

19.8

29.1

Mean over
all items

47.5

55.8

54.6

61.5

32.118.934.9

-17.9

16.7

5.3

27.8

-13.0

6.2

3.8

29.4

33.924.225.4

43.224.235.7

30.813.614.3

9.322.753.2

35.231.830.9

33.321.923.0

46.339.446.0

30.228.823.0

6.23.o5.6

Mean Discrimination Index forll Grades on Total Test: 26.0

Total N = 359

IV-31
TABLE IV-9. Mean Difficulty and Discrimination Indices Fbr The Class Reasoning Test,
Based Upon Pre-Test of LtT's, LEM-1's, and LNDT -2's CaMbined.,

4
94

Grade
N =

Mean Difficulty Indices

6
10
8
100
75
103

12
72

Mean Discrimination Indices

8
10
4
6
103
100
75
94

12
72

Item.

Group
1
2

76.1
55.7

90.6
64.9

91.5
67.7

93.4
82.4

95.4
82.7

53.3
32.7

22.6
27.4

21.0
37.0

5.0
15.o

10.0
20.0

30.2

42.9

49.2

66.5

75.0

24.0

40.5

39.5

39.2

45.0

43.4

42.6

47.7

58.0

66.7

20.7

23.8

31.5

25.8

34.2

64.7

70.7

66.3

80.0

79.9

30.0

14.3

19.7

10.8

22.5

61.7

76.9

79.0

86.5

87.1

40.0

26.8

38.3

15.8

25.8

33.9

38.o

36.5

58.7

63.o

44.7

37.7

31.7

46.7

37.1

42.4

51.7

67.6

71.6

28.0

41.7

43.2

38.3

45.8

51.1

56.3

58.5

66.7

58.1

41.5

20.8

32.7

17.5

26.7

10

43.6

48.2

52.5

59.3

55.3

30.0

31.6

39.5

20.8

20.8

11

52.1

62.3

66.3

79.3

78.3

38.7

27.4

45.7

19.2

38.3

12

41.7

43.9

53.2

61.1

59.7

23.3

26.8

42.6

27.5

41.7

CF

54.5

64.8

67.2

78.5

77.3

32.3

28.9

32.5

19.2

29.o

SY

43.1

44.8

46.4

55.0

61.7

22.3

20.5

25.o

24.6

28.3

SU

40.1

47.2

56.2

67.3

67.0

33.1

40.0

48.2

28.3

44.6

56.6

60.0

71.6

72.7

30.7

29.o

35.7

22.2

31.5

6.7

Mean for
all
items

Mean

49.3

Discrimination Index for All Grades on Total Test: 29.8.

Total N = 444.

IV-32
ranges extend across grades, but one can see that certain item groups are more
difficult than others.

This is in conformity with the view that certain prin-

ciples of logic are more difficult than others and are learned later. It is in
conflict with the view that one kind of logic is learned at one stage in life and
another kind learned at a later stage.

It appears that parts of one kind are

learned before parts of the other and that parts of the other are learned before
parts of the one.
The level of the mean difficulty indices (ranges: 47.5

61.5 for conditional

;9-7

reasoning and lef*N.404 for class reasoning) suggests that neither test,
When administered for purposes of securing a total score, is too difficult for
grades 4-12.
Our purpose in developing this test, however, was not primarily that of
securing a total score, but rather to enable us to determine Whether a certain
principle of logic has been mastered at a given level. Hence we do not approach
the test with a preconceived idea of what the mean difficulty levels of the
item, groups should be. Instead we in a way seek to find out What the mean dif-

ficulty levels are in order to find out What is mastered and What is not.
The discrimination index pattern, in conjunction with the difficulty index
pattern, is quite interesting from the point of view of development, but as we
saidl that is the topic of the next chapter. The following observations are
relevant here:
1.

The only item group that consistently fails to discriminate at all

levels of administration is conditional Item Group #12. This may be because of its great difficulty (mean always below 30%). With instruction in
logic incidentally, as will be seen in Chapter VI, the item group becomes
considerably easier in grade eleven with a mean difficulty of 62%.

Other cases of law mean discrimination indices are found at either

the early grades or the advanced grades, the former apparently because the item

group is difficult at that level, and the latter because the item group is quite
easy at that level.

Cases of the former are conditional numbers 2, 3, and class

number 7; cases of the latter are class numbers 1 and 2.

All low discrimination indices are thus accounted for except the fairly

low ones in grades nine (13.6) and eleven (14.3) for Item Group #6 in conditional
reasoning.

We do not have an explanation for these fairly low indices, nor for

the fact that they are so much lower than the index for this item group in grade
five (50.6).

However, they are not embarrassingly low; they are just fairly low.

With this one exception, then, all low discrimination indices are accounted
for by the purpose of the tests, Which was to ascertain whether members of a set
list of basic principles were mastered at various levels. Unsurprisingly there
are some which at a given level are either easy enough or difficult enough to
result in low discrimination indices.
Given the purpose of these tests, the discrimination and difficulty index
patterns are satisfactory.

Making Sense in the Context of the Study.

A third type of argument for

construct validity tries to show that the results of the use of a particular
instrument make sense.

Thus in a way the argument for validity of these instru-

ments depends in part on the intelligibility of the experimental findings. Hence

the next two chapters, Which present the experimental findings in the areas of
natural-cultural development and readiness development, are implicitly discussions

of the construct validity of the tests. Roughly speaking the more intelligible
the results, the more valid the teat is shown to be.

SUGGESTED PURTM3B DEVELOPMENT

Several ideas for further development of the Cornell Deduction Test Series
(or of some other deduction test series) have occurred to us before and while

1ppraising these two tests:

IV-34

Tests of other kinds of deduction should be developed.

Ordinal reasoning

and other kinds of sentence reasoning (alternation, disjunction, conjunction,

and combinations of these with each other and conditional reasoning) should be
developed first, because the principles for these types are fairly well developed.
Tests for deontic, alethic, epistemic and other forms of logic must await the
preparation of a set of principles for each of these types. Hence the second
recommendation.

Principles

of

types

of

logic

other

than

class,

sentence,

and

ordinal should be formulated. This is a very difficult task.

Satisfactory methods should be developed for using the two existing

tests with lover elementary students (grades 1-3). One possible adjustment is

to shorten the tests.

Each test is designed so that the second half (Items 43-78)

can be dropped off without destroying the item group structure. Instead six

of the more difficult item groups are simply omitted.

Tables IV-1 and IV-2

show which groups would thus be omitted.

During the pilot year of the project, we developed what seemed to be promising

procedures for administering early editions of deduction tests to lower elementary students.

For second and third graders items were simply read aloud while they

followed along and marked answers directly on the test booklets. Testing

time was broken up into separate periods of twenty or thirty minutes apiece.
Two people handled classes of twenty to twenty -five students in this manner.

Par first graders five students at

a time were tested for periods of twenty

minutes, the first of these periods being devoted simply to instructions and

practice problems.

A separate answer sheet was used (see Appendix) and the

questions were read to the children.

very busy.

One person handled the testing but was

IV-35
On the basis of this experience we feel that group logic testing with
this sort of test is possible at the lower elementary level. Incidentally Hill
(1961) tested lower elementary students in logic, but did not face the problem,
which, as we indicated early in this capteryt should be faced, of getting subjects

to judge whether an argument is invalid or not, when the proposed conclusion does
not contradict the premises.
4.

Scores on these tests should be correlated with other measures of logic

knowledge, such as grades in an elementary logic course and scores on other

logic tests.

It would be particularly interesting to see the relationship between

these tests and an open-ended logic test in which subjects are asked to supply
the conclusion themselves, the intent being to see the extent to which a multiple
Choice test like these can be used as evidence of ability to deduce conclusions,
given only the premises.

Thus one could see the extent to which these tests

fail to test for deductive creativity (if they fail at all).

These tests could be expanded so as to include more items in each item

group, thus achieving a more reliable measure of each principle, but of course
sacrificing either comprehensiveness or compactness. Such a sacrifice will be
necessary if one wants to use the test to measure and predict individuals'
scores, as can be seen from the item group reliability estimates.
6.

the tests could also be expanded to include a negation component.

Hill

found that negation was an important source of difficulty in logic for lower
elementary students (1961, p. 66). In a way the conditional reasoning test
has a negation component built in by means of the C104 items.

These had mean

difficulties Which were lover at each grade level than the totel mean difficulties and the mean difficulties for each of the components. CF4 mean difficulties at grades 5, 7, 9, and 11 are respectively 38.4, 45.8, 43.3, and 50.0
as compared with total mean difficulty indices of 47.5, 55.8, and 54.6, and 61.5.
Hence negation appears to be an important source of difficulty.

IV-36

Much correlational work

could be done in an attempt to see relation-

ships between various personality

One migOtwonder vhether

and other variables and the logic component's.

an authoritarian personality, for example,

is'relatively

poorer on the suggestive component than is some contrasting personality type.

This chapter commenced with

by presenting the three
suggestive); showing

a description of the basic structure of the

the distribution of these components within each group of

of logic which were presented

developed.

between the item groups and the principles

in Chapter II; and explaining the item arrangement

Next operational definitions of 'mastery of X principle' were

Roughly speaking these operational definitions specified getting

correct at least five out of

an item group as a sufficient condition for mastery

and getting at least four

as a necessary condition for mastery.

Next came the presentation

test-retest correlations.

of reliability estimates, Which consisted of

Split-half correlations were not used, because it was

not possible to split the tests

reported for total

Cornell Class Reasoning Test"

Reasoning Test".

into equivalent halves. these correlations were

score, component scores, and item group scores at each grade

level that took each of the

two tests.

The mean reliability estimate on "The

is .83, and is .75 on "The Cornell Conditional

The mean correlations for the components and item groups are

lower, but sufficient in

our judgment for use of the tests with groups.

Discussions of content and

sisting of three parts:

construct validity came next, with the latter con-

correlations with well known variables, item analyses,

and making sense out of

the problem of the next

the experimental results.

In effect, this last topic is

two chapters on the experimental results.

Finally a number of suggestions

Made.

tests

components (concrete familiar, symbolic, and

six items; showing the relationship

and format.

CHAPTER SUMMARY

for fUrther development of logic tests were

Chapter V.

The Natural- Cultural Development

of Knowledge of Logic

The development of knowledge

topic of this chapter.

or

logic under contempory conditions is the

We shell not attempt to make any estimation of the

degree to whidh development is attributable to nature as opposed to nurture.

We shall simply try to describe

what we have found in this group of 803

upper New York State students whose ages range from about 9 to 18, and Whose

mean IQ is roughly 114.

We do not doubt for a moment that environment has played

a significant role in the development of their logical capacities.

One cannot, for example, be

sure that there has been no previous con-

scious effort to teach them logic.

most students have received at least

parents, their teachers,

On the contrary it is presumable that

some informal instruction from their

or their peers. But we do know that our staff did

not teach logic to these students prior to the test administration which

this chapter considers.

In the next chapter we will consider What later

happened to those students to Whom we did try to teach logic.

THE LITERATURE

Jean Piaget is the best known and most prolific contributor to the
literature on the development of knowledge of logic: First we shall examine his
characterization of the formal operational stage, which he holds runs

from ages 11-12 on.

Then we shall pose questions which his views suggest

and on whidh the pre-test administration of the Cornell Deduction Tests

throws some light.

Other studies will then be considered for their

bearing
on these questions.

Piaget's Formal Operations.

A standard claim made by Piaget

is that the concrete operational period

of thought runs from ages 7-8

toll-32p at Which time it is !aimed by

the
formal operational period, the apex in the development of thought (e.g.,

V-2

Inhelder and Piaget, 1958,

emphasize formal operational

p. 1*). In the approaching discussion we shall

thought, but shell draw upon his notion of

concrete operational thought for purposes of contrast.

The Real Versus the Possible.

According to John Revell

"The most important

in The Developmental Psychology of Jean rivet,

general property of formal operational thought, the one

from which Piaget derives

all others...,

sible." (1963, p. 204).

sible, rather than just

Apparently it is the ability to deal with the pos-

the real, that supposedly is uniquely found in the

formal operational period.

different basic features

concerns the real versus the E227

We have found in Piaget three, perhaps four,

of this cognitive ability to deal with the possible,

instead of just the real:

1) The first is the ability
validity as opposed to

to judge an argument on the basis of its

judging it on the basis of one's belief about the

truth of the conclusion.

istic'. It is manifested in
Which are known to be

We shell call this the 'truth-validity character-

several ways: by reasoning from assumptions

false, by reasoning from assumptions whose truth

status is not known, and

by reasoning and concluding without regard to the

truth status of the conclusion.

According to Piaget:
Childish reasoning beteveen
fore
the years of 7-8 and 11-12 Will therepresent a
very definite feature...: reasoning that is connected
with actual belief,
vation, will
or in other words that is grounded on direct obserFor formal reasoning
be logical.
But formal reasoning will not yet be possible.
connects assumptions -- propositions, that is,
order
in which one does not
necessarily believe, but which one admits in

to see whet

consequences they will lead to (1928, pp. 250-51).

*Henceforth in this chapter,

citing the year, 1958.
in the Preface (p. mciv),

According to the distribution of authorship reported

the parts that

actually vritten

as in Chapter II, we shall refer to this work by

we are quoting and referring to are

V-3

More specifically Piaget

says the following things:

The formal

operatignal

thinker is able to

reason "on the basis of assumptions which have no neces-

sary relation to reality or to the subject's beliefs... He relies on the

necessary validity of an inference

sions with experience" (1950,

premises as simply given, without
not...; belief in the conclusion

Or even if he

p. 148).

"To reason formally is to take one's

enquiring whether they are well founded or

will be motivated solely by the form of the

deduction"* (1928, p. 251).

other hand, "the child

as opposed to agreement of the conclu-

Luring the concrete operational period, on the

cannot reason from premises without believing in them.

reasons implicitly from assumptions which he makes on his own,

he cannot do so from those

which are proposed to him" (1928,

Revell says, "In general,

Piaget finds that contrary-to-fact

suppositions...tend to be foreign
(1963, p. 208).

p. 252).
'what if'

to the thought of middle childhood"

Flavell is here emphasizing a striking part of the truth-

validity characteristic,

that of working from assumptions which one actually

disbelieves.
The suggestive component

of the Cornell Deduction Tests is aimed at this

truth-validity characteristic.
status of the argument is

In each item in this component the validity

different from the truth status of the conclusion,

and often the premises

that are offered are obviously false.

To

identify

these items in the tests,

and IV -2 in Chapter
.2)

which maybe found in the Appendix,

IV for the item numbers assigned to this component.

A second feature

ability to operate in

given situation.

see Tables IV -1

of the ability to deal with the possible is the

the framework of all the possible combinations in a

Since this characteristic was discussed at some length in

Chapter II, one quotation

should suffice here:

Presumably here he means 'belief ghat

not
t conclusion follows necessarily';
conclusion is true'.' The latter is suggested by his choice of
words, "belief in the conclusion".

'belief that the

If we accept the task

operate
of describing the structures which actually
in the subject's
minds, we have to use the criteria furnished

by the combinatorial
system in distinguishing between concrete ... and
formal
... operations (1958, p. 280).

Chapter II attempts to

clarify Piaget's notion of a combinatorial system.

Presumably the Cornell

tic of the ability to

reasoning, which implies
working within a

Deduction Tests do not test for this characteris-

deal with, the possible. If, however, doing conditional

doing propositional reasoning, is alleged to imply

combinatorial system (and,

there is some suggestion

ing test is a test

that Piaget would want

as we pointed out in Chapter II,

to this effect), then the entire conditional reason-

of this combinatorial characteristic.

to allow this,

maintain the presumption

so, until told otherwise,

We do not think

we shall

that the Cornell Deduction Tests do not test for

this characteristic.
A third feature,

3)

the ability to control

as Piaget views the real-possible distinction, is

variables in (presumably) an empirical investigation:

Two discoveries found

at the beginning of the formal level are (1)

can be separated out by

exclusion and (2) that
purpose of analyzing its

a factor can be eliminated not only for the

ovn role but, even more important, with

the variations of
the Cornell Deduction
that this

associated factors.(1158',P4M)

Tests do not test for this characteristic, unless it is

alleged

cally

characteristic and conditional reasoning ability are logi-

interdependent.

Because Piaget seems to relate all of the features of

stage to each other in

there is some ground
interdependence.

a manner that is never clearly specified,

for thinking that Piaget might allege the above logical

But again, until

would not consider

the formal

told otherwise we shall assume that Piaget

the Cornell Deduction Tests to be measures of ability to

control variables.
4)

Whether the fourth

possible distinction

feature should be considered part of the real-

or a separate feature of the concrete formal distinction,

V-5

is problematic.

This fourth feature is the ability to do propositional logic,

According to Flavell, as quoted above, the real-

possible distinction is "the one from which Piaget derives all others",

so

presumably the class-propositional distinction should be derived from it as

well.

Furthermore, as we indicated in Chapter II, there is reason to believe

that Piaget wants to connect ability to do propositional

reasoning with the

ability to deal with the possible instead of just the real.

But, as we also indicated in Chapter II, to hold this view about
propositional logic is to hold a different notion of propositional logic
than the standard one-and than the

one which is operation in this project. The

most charitable interpretation of Piaget's views is that propositional

reasoning is definitionally independent of ability to

We shall adopt this interpretation.*

In any case Piaget maintains that propositional reasoning (and thus

sentence and conditional reasoning) is not possible until age 11-12.

"The child at the concrete level (stage II: from 7-8 to 11-12 years) cannot

yet handle... propositional logic. ..." (1958, p. 1).

The conditional reasoning

on "The Cornell Conditional Reasoning Test"

corresponds to much but not all of Piaget's propositional logic.

because of time and space

junction.

Omitted,

pressures, are alternation, disjunction, and con-

Included are conditionals, which, roughly speaking, are statement&

containing the word 'if' and its

synonyms.

That Piaget finds conditionals

to be a central part of his propositional logic is seen in the following

quote which, in referring to

implications between propositions, is referring

to conditionals:

412hren though we list this

others.

propositional-class distinction along with the

Pbrmal operations, therefore, consist essentially of 'implications'

(in the narrow sense of the word) and 'contradictions' established
between propositions which themselves express classifications,
serrations, etc. (1950, p. 149).

Kinds of Fbrmal Thought.

Recently Piaget has introduced two stages in formal thought which he

calls IIIA and IIIB, the former running from 11-12 years to 14-15 years and
the latter from 14-15 years onward (1958, p. 1). Be holds that it is not
until IIIB that the child thinks in terns of necessity of relations (presumably empirical relations) instead of constancy (1958, p. 11); and that
the full grasp of the concept, all other things being equal, is attained
(1958, p. 43).

Neither of these features is a feature of formal logic and

neither is tested for by the Cornell Deduction Tests.

Interestingly Piaget does not suggest, so far as we can find, that any
principles of conditional reasoning are attained later than others, nor that
any is more difficult than another. Similarly he does not distinguish among
principles of class reasoning, nor among principles of alternation, disjunction, and conjunction.

Instead he lumps all of sentence reasoning together

(under the title "propositional reasoning") and all of class reasoning

together, saying that the latter can be done in the concrete operational
stage but that the former cannot be done until the formal stage. There
seems in his work to be no differential treatment of either type of logic,
in so far as development is concerned. He does of course recognize different
principles of reasoning, but for study of development and mastery, he lumps
them together.

Stages.

Piaget talks as if there are stages in the intellectual (including

logical) development of youngsters.

The claims about the concrete opera-

tional period running nun 7-8 to 11-12, and the formal operational period

running from 11-12 onward sound as if he does think in terms of stages.

Naturally the distinction between the meaning of 'stages of growth' and
'continuous growth' is a difficult one to draw,

question in a precise manner.

so it is hard to put the

And Piaget does not help very much, perhaps

because his views are changing.

In Judgment and Reasoning in the Child

he speaks fairly explicitly of distinct stages:

The evolution of intelligence is therefore not...continuous, but
rhythmical; it seems at times to go back upon itself, it is subject
to waves, to interferences, and to 'periods of variable lengths'
(1928, p. 215).

On the other hand his introduction in The Growth of Logical Thinking

from Childhood to Adolescence of two subdivisions of the formal operational
period, stages IIIA and IIIB, suggests on his part a belief in more continuity

of deVelopment than is implied by the above quotation.

The data of the present study has bearing only on the age range 10-18,

and thus should not be considered

concrete operational stage.

as evidence on the claim that there is a

It is some evidence on the question of the

existence of stages in knowledge of logic in adolescence. But whether Piaget

ever actually meant to raise this question specifically is unclear,

because

the distinction between ILIA and IIIB is not a distinction between types of

logical knowledge.
Unfortunately the question of the existence of stages is basically an
unclear question because there is lacking a criterion of how much of a
leveling off is necessary for there to be a stage. The question is essentially a pragmatic question, since the

We interpret the question, "Are

answer depends on one's purposes.

there Stages?" as meaning, in effect, "Is there

enough of a regular leveling off for a long enough time for the

leveling to be significant for one's purposes?" Thus it is quite important

to specify the purposes.

Out of some purposive context, the question is

v -8
at least partly indeterminate.

Perhaps that is why there is so much dispute

about Piaget's alleged stages.

The concept, stage of development, is a tricky one. We do not feel
that justice has been done it in the preceding paragraphs. Awaiting a
good analysis of this concept we shall limp along with the above intuitive

notion of purposiveness.

Perhaps this will do, given the current lack of

thorough and detailed information.

Questions Posed by Piaget's Work.

The age range covered by the present study (roughly 10 to 18) corresponds
approximately to that of Piaget's formal operational period (11-12 onward).
Because his work on this period is so extensive and well-known, we presented

his theoretical description in detail in order to show the taking-off point

for the natural-cultural part of our study. Of the four basic features of
the formal operational period (possession of the truth-validity characteris-

tic, ability to operate within the framework of a combinatorial system,

ability to control variables, and ability to do propositional logic), our

findings are relevant to the first and fourth.

These two features are the ones treated in the current study because
they are the ones that are clearly related to ability to do deductive logic.

other two are either peripheral

The

or unrelated, depending on the interpre-

tation given them.

What we have done is not to be regarded primarily as a test of Piaget's

claims, but as an extension and refinement of them. By and large we find

his claims to be a bit too

vague and/or indeterminate for careful testing.

concepts and concerns did, however,

tions Which are listed below.

His

provide us with ideas for the ques-

Although, as the reader will see, our answers to

these questions still leave much that is indeterminate, they are more

V-9

definite then Piaget's claims and hence are an extension and refinement of

them.

Of course there is a degree of testing of his claims as well.

For the genesis of the following six questions we are intellectually

in debt to Rapt. Mese questions were suggested by his work and strike
us as theoretically interesting and practically important, from the point
of view of someone making decisions about the development and placement of
materials in a curriculum.

Is there actually a development of logical ability as children grow

older?

Does this development (if there is any) come in stages?

Is conditional logic mastered by age 11-12?

Is class logic mastered by age 11-12?

Is the truth-validity characteristic achieved by age 11-12?

Within each type of logic, is there

a development of one sort of

thing before another and are some things more difficult than others at a
given level? If so, What is the nature of this differential development? Of
the above questions, we feel that the last is the most interesting

and the one toward which this study makes an original contribution. The first
question which is about whether there is development, has an unsur-prisingly
affirmative answer from the present study. The second question,
once you are pest the problem of specifying the criteria for a stage, is

though still not easy to

stage to be.

answer, very such dependent on What one regards a

linsvmrisingly the evidence of this study points toward an

answer of "Partly" for Questions 3, 4, and 5. But Question 6 is the one

that brings forth specific refinements in gross statements about class logic,

sentence logic and the truth-validity characteristic.

V-10

Other Studies.

This discussion is organized around the six questions just asked.

Is there a development of logical ability as children grow older?

In addition to Piaget the following people, as a result of their research,

contend that there is such a development: Bonser (1910),

(though his tests are

hardly logic tests), Burt (1919), Winch (1921), Moore (1929) and Hill (1961).
Miner (1955), on the other hand, did not find development in ability to
recognize fallacies in grades 10, 11, and 12. However, many of Miller's
fallacies are not fallacies of deductive logic and he worked over a span of
only two years; 80 we are inclined to feel that the weight of evidence from
the literature is in support of the initially plausible view, that there is
development of logical ability as children grow older.

Does the development come in stages?

Out of some purposive context, this question is partly indeterminate.

Hence it is difficult to report others' results on the question. We can
only indicate whether that they found struck them as being regular developmeat.
Moore (1929) believes that the development in knowledge of deductive
logic that he found in children of ages 6-12 was regular (and hence not in
stages).

Burt (1919, p. 126) and Winch (1921, pp. 138, 209, 284) found

regular improvement with grade. Hill (1961, p. 51), working with students
of ages 6-8, found regular development in their knowledge of logic. Since
Miller (1955) fbund no development in ability to recognize fallacies in
grades 10-12, his results are not inconsistent with the stage hypothesis.
On the thole, we feel that the reports of research do not support any stage
hypothesis, particularly in view of the fact that we are unable to see
support for it even in the experimental findings of Piaget himself. But because of the context dependence of the question, one must not on this basis
make a definite Judgment about the existence of stages.

Is conditional logic mastered by age 11-12?

Considerable research has been done Which suggests (in conflict with
Piaget's claims) that children can do at least some conditional reasoning
before 11-12; Bonser (1910), Burt (1919), Winch (1921), Woodcock (1941, p.

136), Hill (1961,

p. 51).

But the question of whether and to that degree

it
has been mastered by age 11-12 has not been investigated. NowIlaget does
not state that it is fully mastered by the age 11-12, but he does not indicate the extent to which he thinks it is mastered after this age is attained.
The two earlier-mentioned characteristics of the distinction between stage
ILIA and IIIB do not help, because they are concerned with empirical rather
than logical matters.

To our knowledge no pure conditional logic tests have been administered

to adolescents and analyzed for the degree and kind of knowledge shown.

Is class logic mastered by age 11-12?

The references cited under the previous question also support the contention that at least some class logic is mastered before 11-12. And Piaget

would agree with this, though

one cannot be sure whether he thinks that all

of the basic principles of class logic are mastered by the end. of the con-

crete operational period.

One might think so, since class logic is a charac-

teristic of this period, according to Piaget. However, all he says is that only
class logic is used in this period (1958, p. 1). Be never specifically

states the extent to which he thinks it is mastered, OD far as we can determine.

Is the truth-validity characteristic achieved b7 age 11-12?

So far as we can determine, Piaget is the only person Who has made
claims about the attainment of the truth-validity characteristic before age
11-12.

A nuMber of studies of this component of logical ability are mentioned

in Chapter IV, but they are all

on older people.

Since these studies gener-

V-12

ally show that suggestive arguments with content are handled less ably than
arguments with concrete familiar content, the minimum answer suggested by
these studies to the question is that the truth-validity characteristic is
not fully achieved by age 11-12; So far as we can tell, Piaget does not
commit himself to a degree of attainment of the characteristic, just as he
does not commit himself to a degree of attainment of propositional and class
logic.

Within each

J- of 1

is

is there a develnt of one sort of

thing before another and are some things more difficult than others at a
given level?

If so what is the nature of this differential

There is not much evidence on this matter. Except for his attention to
the truth-validity characteristic, Piaget seems to have ignored it.

Hill

(1961) explicitly raised the question with respect to knowledge of different

principles and found in ages 6 through 8 no pattern that she could discern

(p. 57).

However, her test, it should be remembered, did not test for any

fallacies.

In class reasoning Wilkins (1928, p. TT) found the components to

be in the following order of increasing difficulty for college students:

concrete familiar, suggestive, symbolic, and unfamiliar.
Burt (1919), as a result of his experience developing and administering

his "Graded Reasoning Tests" held that the basic mechanisms of formal reasoning are all there by the mental age of 7, and that differences are results of complexity.
All the elementary mechanisms essential to. formal reasoning are
present before the child leaves the intents' department, i.e., by the
mental age of seven, if not soneWhat before. Development consists
primarily in en increase in the extent and variety of the subjectmatter to Which those mechanisms can be applied, and in an
increase in the precision and elaboration with Which those
mechanisms can operate. The difficulty of a test depends upon

its complexity, that is in the mein upon four points: how

many connections have to be made between one idea

V-13

and another -- only threes as in the ordinary syllogism, or four,

or more? How many of these connections has he to supply himself? How
closely are these connections to be knit together -- in
parallel, in series, or in a more or less intricate system? How
far do they fall into the same category of time) space, nuMber;
etc., or differ one from another? Other points, whet is the pre-cise
nature of these connections -- temporal, spatial, numerical) causal,
etc. -- and of their interconnections -- hypothetical, disjunctive,
etc. -- are of little importance.

Burt's four aspects of complexity mentioned in the above paragraph (number of

connections, number of connections to be supplied by the subject, intricacy
of total, and extent of being in the same category) are potentially workable
categories of research. Of course they need to be defined and clarified, but
one might classify arguments according to these categories and see if subjects
at various levels can handle them.
Our difficulty with these four categories of Burt's is that they do not
seem immediately useful in making decisions about curriculum and grade placement of materials.

The reason for this is that we now think in logic in terms

of valid and invalid arguments, class and sentence reasoning, and various
rules of inference. It would be possible, but inconvenient, to think instead in
terms of Burt's four aspects of complexity, suitably reformulated. However,

before turning to Burt's categories we should try to work with the established
ones and see if differential development exists in terms of the established

categories.

Hence the present study is framed in terms of the established

categories of logic.

It would be worthwhile in the future not only to repli-

cate this study, but to compare the fruitfulness of the two conceptual
schemes.
Summary.
Plaget's discussion of four basic features of formal operational thought
(possession of the truth-validity characteristic,

ability to operate within

a combinatorial framework, ability to control variables, and ability to do

propositional logic) provided a conceptual springboard for the framing of the

above six questions abort the neturalacUltural development of knowledge of

logic.

Although there will be a testing of his claims to some extent, the

current study is to a greater extent concerned with the extension and refinement of his views.
The evidence of others supports the contention that there it development of logical ability; does not support the view that this development comes

in stages; suggests that the basic principles of both conditional and class
logic are not mastered by age 11-12; supports the contention that the truth
validity characteristic is not fully developed by age 11-12; and has little
to say about the differential development of principles and components of
logic.

The major contribution of the current study to knowledge about the

natural cultural development of logic is its exploration, conceptualization,

and tentative results on the differential development of knowledge of logic.

THE RESULTS

Like the previous section this part of Chapter V is organized in accord

with the six listed questions.
1.

Is there a development of logical ability as children grow older?

That our results indicate a positive answer to this question can be

seen in a number of ways. In Chapter IV Tables IV-8 and IV-9, which give the
mean pre-test difficulty iLdices for item groups, components, and total test,
dhows development from the lowest grade to the highest on all item groups
except "12 conditional, on all components, and on both tests as a whole.
Apparently Item Group "2 conditional is just too difficult for any development to dhow within the range with which we worked.
Inspection of total scores on the tests for LDT's, LNDT-1's, and LNET-2's

separately also reveals this development. These scores appear in Chapter

III in Table III-3. Table V-1 in this chapter presents the mean total pre-

V-15
test scores for each grade level, all subjects combined. Again development
with advancing age and grade can be seen. )br interpretation purposes this
table also presents the mean chronological age, IQ, and estimated mental age.
Even though the IQ scores are on different tests, resulting from the combination of the LNDT-2's with the others, these scores help to explain the
larger jumps found from the 5th to the 7th grade on the conditional reasoning
test and from the 8th to the 10th grade on the class reasoning test.
Table V-1 also presents the mean pre-test scores for all subjects com-bined
on the components and item groups. Again development is in evidence

in every case except for #12 conditional. If one looks more closely at these
figures than simply inspecting the figures for the youngest and oldest, the
significance of IQ as well as age and grade is again suggested.
There are altogether 112 different steps from one grade to the next tak.

ing into consideration total, component, and item group scores.Of these 112

steps, 16 are down, 2 are at the

same level, and 94 are up.

Let us

examine

the 16 steps down.

Only one of them occurs *ere there is an increase in

mean IQ from one grade to the next, and that is again for Item Group #12,
conditional.

Of the others, 6 occur in the shift from grade T to 9, condi-

tional; 4 occur in the shift iron grade 9 to grade 11, conditional; and the
other 5 occur in the shift from grade 10 to grade 12. In the first two of

these three shifts there is an actual decrease in the mean I.Q. In the
third shift the mean IQ (120) stays the

to 7, 4 to 6, 6 to 8, and 8 to 10)

same. The other 4 shifts (grades 5

are accompanied by increases in IQ.

However, there are still some steps that do not quite fit the explanation

that uses only a combination of the

chronological age, grade, and IQ factors.

Perhaps there is also a leveling off in grades 9 through 12 (CA roughly 15

through 18 years; MA roughly 200 through 260 months). We shall return to

this possibility when considering the stage

question.

TABLE V-1. Mean Chronological Age; IQ; Estimated Mental Age; and Total,
Component, and Itim Group Conditional and Class Reasoning PreTest Scores; by Grade for All Subjects Grouped Together.

Grade
N =
Chronological

Age (nos.)
IQ

Conditional Reasoning
11
05
07
09
78
102
99
8o

04
94

129
108

203
109

117
109

127

153
117

184
110

Class Reasoning
08
10
06
103
loo
75

166
113

190
120

214
120

187

44.3

159
53.4

57.8

228
71.2

256
73.4

142

112

Estimated Mental
Age (mos.) (CA x
IQ/100 before
rounding)
139
Total Score*
42.4

179
51.7

201
55.3

220
56.6

Component**

CF
SY

23.3
5.8

27.1
6.7

29.0
6.5

29.5
7.2

25.6
5.1

30.2
5.3

32.1
5.9

38.2
6.9

39.0
7.5

su

4.6

6.o

6.4

6.4

4.7

5.6

6.7

8.3

8.2

4.3

4.5

4.9

4.7

4.3

5.4

5.5

5.7

5.8

1.4

1.7

2.1

2.2

3.7

4.3

4.5

5.2

5.2

1.2

1.6

2.1

2.0

1.8

2.5

2.9

4.o

4.6

3.3

4.1

3.8

3.9

2.6

2.6

2.9

3.5

4.1

3.4

4.0

4.2

4.5

3.6

4.2

4.2

4.7

4.8

3.3

4.o

3.9

3.6

3.5

4.5

4.7

5.3

5.2

0.9

1.5

2.0

2.6

2.0

2.2

2.3

3.7

3.9

4.1

4.9

4.6

5.2

2.3

2.5

3.o

4.2

4.5

4.3

4.7

4.6

5.2

2.9

3.3

3.5

3.9

4.1

3.3

4.3

4.2

4.5

2.5

2.7

3.1

3.5

3.3

11

3.2

4.o

4.1

3.9

2.8

3.4

3.9

4.9

4.7

12

1.6

1.0

1.8

1.3

3.5

3.6

4.2

4.8

4.6

9
lo

*Tbtal score was calculated using scoring formula: R=W/2 + 27.

**Component and item group scores are number of right answers.

V-17

Development of logical ability with grade

in Tables V-2 and V-3, which present

(and thus age) can also be

the percentages of subjects at

each grade vbo met the sufficient condition for mastery

(at least five correct)

each principle or combination and vho failed to meet the necessary

mastery (at least four correct).

seen

condi-

of

tion for

A reminder: Those who failed to

meet the necessary condition are those who marked correctly three or fever
of the items in an item group. Hence those subjects vho marked correctly
exactly four of an item group neither met the sufficient condition, nor

to meet the necessary condition.

failed

Pbr this reason the percentages do not

add

up to 100%, the difference between 100% and the sum of the percentages given
for a given grade being the percentage who marked correctly exactly four
items in an item group.
An inspection of these two tables again shows development in logical
ability.

Differences among principles, which were apparent in the diffi-

culty index tables (IV-8 and IV-9) and in the absolute score table (Via)
are again apparent in these necessary and sufficient condition tables.
These will be discussed under the topic of differential development.
In sumary there is quite clearly development of logical ability in
subjects like those we tested. Even though this was not a longitudinal
study, it wild be very difficult to deny the development hypothesis and
still offer a satisfactory explanation of the data.

2. ...21gAt:...eZEEELTELIELOAPAY
As indicated earlier, the answer to this question depends upon *et degree and length of leveling off shall count as a stage -- and this depends
in part upon the general context in which the question arises. Hence for
this question there are not only problems of errors and gaps in measurement
there are also problems of interpretation of the question. We are thus in

position only to make the most tentative suggestions.

Table V-2.

Percent Meeting the Sufficient Condition and the Percent Failing to Meet the Necessary

Condition Pbr Mastery of Each Principle at Each Grade Level on the Conditional

Reasoning Test.
Grades
N =

Item

Percent Meeting the Sufficient Condition

05

102
53.

Group
1

co

5
6
I7

30
25
34

10

53

46

11

26

12

23

07

99

09
8o

11

78

56 66 62
6

41

45

40

5
63
63
52
40
4

.4

35
4o
35
11
70
69

53
46
1

12
3
35

58
33
19
79
ft

58

4o
0

Percent Failing to Meet the Necesnary Condition

05

102

07

99

09

80

11

30 26 21 22
92
94
54
48
51
94
31

26

54
51
86

80

92
36

38
36

84
20

17
30

37
91

90
89
41

35
45
80

13
23

34
33
93

73
85
4o

22
47
68
9

5
23
36

95

Table 37-3.

Percent Meeting the Sufficient Condition tnd

the Percent Failing to Meet theNecessary Condition Fbr

Mastery of Each Principle at Each Grade Levelon the Class Reasoning Test.

Grades
N =

Item

Group
1

2
3

Percent Meeting the Sufficient Condition

04
06
08
10
12
94

103

loo

75

72

56 79

91 92

97

34

55

80

21

44

75

44

Percent Failing to Meet the Necessary

04
94

Condition

06
103

33 10

10

12

loo

75

72

08

5 1 3

6
7
8

lo
11
12

15

10

lo

68

41

83

28
66

25
63

t5

6
22

15

24

43

71

72

68

148

32

31

37

46

63

67

4o

24

24

17

3o

56

63

83

85

15

13

35

46
83

22
79

19

17

24

37

11

47

73

4o

42 r4
1

22
lo
15
30

V-20

Graph V-1 depicts the total score vs. chronological age coordinates
for each test, making use of the means that are presented in Table V-1.
The lower line connects the points for the conditional reasoning test, and
the upper line those for the class reasoning test. Since the points are
approximately two years apart in each case, and since there are only four and
five sets of coordinates respectively, the graphs are difficult to interpret.

they show stages or not?

1k

We find ourselves unable to give a firm answer

to the question.
However, there is some indirect evidence against the existence of stages

or at least in favor of their being less severe, if they exist. This is

seen in a comparison of Graphs V-1 and V-2. Graph V..2 plots the same total
score means against estimated mental age instead of chronological age. The
points in the mental age graph

come closer to fitting the speculative dotted

straight line than do the points in the chronological age graph. In other
lords the degree of abruptness that is found in the chronological age graph can be
at least partially explained by variations in IQ from one grade level

to another.
Naturally this method using estimated mental age is not a precise one,
since it is based upon different tests,

IQ score and mental

and since the relationship betwen

age is a matter of some controversy. But the procedure

does seem to have merit,

so we used it.

One must make some tentative

assumptions.
Even with the mental age adjustment, there is still not exactly a

straight line for either kind of

by experimental error
sort.

reasoning.

The data might be explained

or measurement, or the existence of stages of some

Conceivably there are stages for each of the principles of logic.

they would be masked by this

that the data of this

total score treatment. Regrettably we

study are definitive enough to do more

If so,

do not feel

90 ...Total
Score

GRAPH V-1. Conditional and Class

Reasoning Mean Total Score vs. MeanChronological Age

80

70
60
50

4o
30

12

10

Class Reasoning

El
Conditional Reasoning
11

Legend: The lover set ofpoints

conditional reasoning test.

class reasoning test.

and lines is for the

The upper set is for the

A set of coordinates
point is
is plotted for eachgrade. Each
labeled by the number
of its grade level. Source
is Table V-1.

The plotted points

straight lineof

200

are connected by solid lines. A

approximate best fit is dotted.

220 210 260

Chronological

Age (mos.)

V-22

than speculate about the existence of stages in the learning of principles.

In summary the stage question is a very difficult one and no definitive
answer is warranted on the basis of these data. Mere is 'vidence against
the existence of stages (within the age range studied) as extreme as Piaget
posits in his earlier writing, but the question about the existence of more
refined stages is largely unanswered, and until the question is refinedl
largely unanswerable.

Is conditional 'tic mastered by age 11-12?

There are several different ways of approaching this question. One is

by examining Table V-3, which give the percentages at each grade level vho
met the sufficient condition and who failed to meet the necessary condition

for mastery of each principle.

Another way is to look at the total scores

at the various grade levels, as presented in Table V-10 and judge Whether any
or all of such mean scores indicate mastery. A third way is to look at

the mean difficulty indices given in Chapter IV in Table IV-8. Each of these
ways supports a negative answer to this question, whether interpreted in
terms of chronological age or mental age.
Age 11-12 years is equivalent to 132-144 months. In chronological age
this corresponds to our 5th and 6th graders, in mental age to our 4th
and 5th graders.

Hence if the answer to the question is positive, then

conditional reasoning should be mastered by our subjects in grade 5, and

certainly 14r grade 7.

Consider Table V-2, the

conditional reasoning.

necessary and sufficient condition table for

Of course one of the most striking things about it

is the difference among the item groups. But let us leave that fact aside

for the moment.

At grade 7 four of the principles (or combinations) are

definitely not mastered; the

and 12.

ones corresponding to Item Groups 2, 3, 7,

The other principles (or combinations) are probably mastered by

9C--Total

GRAMMV-2. Conditional and Class Reasoning Mean Total Score vs. Mean MentalAge

Score

Class
Reasoning

70

60

Conditional Reasoning
11

50

Legend: The lower set of points and linesis for

conditional
reasoning test. Theupper set is

the

for the class

reasoning test.
A set of coordinates

is plotted for eachgrade.

Each point is labeled

level.

by the number of its grade

Source is Table V-1.

The plotted points

are connected by solid lines.

A straight line of
I

120

140

160

180

200

220

approximate best fit is dotted.

240

Estimated Mental Age (mos.)

260

280

V-24
roughly half of the subjects and probably not mastered by about a third
(with about a sixth falling in the borderline area). The situation in grade
5 dhows a lesser degree of mastery but on the same order as in grade T.
Hence the answer to the question that is here suggested is that there is mastery
of some of the principles by some of the people of age 11-12, but rarely (if
ever) mastery of all of the basic principles and frequent non-

mastery of most of them.

When one looks at the question through the mean total scores one sees

roughly the same answer, though it is not as refined.

The mean total scores

for all of our subjects in grades 5 and 7 are 42 and 52 respectively out of

a possible 99.

On the face of it these scores suggest that the basic prin-

ciples embodied in the conditional reasoning test are not mastered at this
level.

Higher scores on the test are possible, as demonstrated by the mean

score of 80.7 reported in Table 111-3 in Chapter III for llth grade students

who have been taught conditional logic by one of our staff members.
The mean difficulty indices given in

Table IM in Chapter IV also sug-

gest that conditional reasoning is not mastered by age 11-12, the mean total

test difficulty for grades 5 and 7 being

around 50%.

It should be

remembered

that this test was not designed with

culty index of around 50%.

Instead it vas designed with the intent of seeing

whether given groups of students

intent requires that decisions

had mastered certain principles.

Such an

about item inclusion and exclusion be based

upon Whether the item is deemed

mastery of the principle.

the intent of securing a mean diffi-

on the face of it to be an indicator of

To select an item on the basis of its difficulty

would be to some extent to build

in an answer to the question of

ithether a

given principle is mastered. We have not done this.

V-25
In summary, it seems clear frmn any and all of the three ways of presenting the data that the basic principles of conditional reasoning are not
mastered by age 11-12, given the contemporary cultural background of these
subjects.

Incidentally it should be noted that, although there was considerable

improvement from grade 5 to grade 11, conditional reasoning is not mastered
by the older group either.
The mean chronological age of our grade 11 subjects is slightly under
17 years and the mean estimated mental age is 220 months (slightly over 18
years).

The same four principles and combinations are not mastered by

these students.

73%, 85 %, 68 %, and 95% of the 11th graders failed to meet

the necessary condition for mastery of the principles and combination

corresponding to Item, Groups 2, 3, 7, and 12 respectively.

And the mean

score on the total tent was only 57 out of the possible 99.
This is not to say that the 11th graders are incapable of mastering the basic
principles of conditional logic. It is simply that under contemporary

conditions they do not do so.

That they are capable of much more is

shown
by the performance of the 11th gradera to were taught logic by a member of

our staff.

This is one of the striking findings to be presented in

the

next chapter.

Is class logic mastered by age 11-12?

This question will be approached from the same three vantage points.

If the answer to the question is affirmative,

then, considering the age

equivalents mentioned under the conditional reasoning discussion, our sub-

jects in grade 6 should have mastered class reasoning.

Consider Table V-3, the

necessary and sufficient condition percentage

table for class reasoning.

notable,

Again differences among item groups are

V-26

though they are generally not so extreme as in conditional reasoning. And

again some of the principles are not mastered by the majority, since over 50%

fail to satisfy the necessary condition for the principles and combinations
corresponding to item groups 3, 4, 7, 8, 9, and 10. Only for two principles
is the sufficient condition satisfied for more than 50% of the grade 6 sub
jects.

Thus the necessary and sufficient condition tables indicate that the

basic principles of class reasoning are not yet mastered at age 11-12,
whether put in terms of mental age or chronological age.
A similar conclusion can be drawn from the mean total score, which is
given in Table V-1 as 53 out of a possible 99 for the 6th graders. And the
mean difficulty index (58%) at the 6th grade level, which is given in Table
IV-9 in Chapter IV, supports the view that the basic principles of class
reasoning are not mastered by age 11-12, though of course there are some
students who have mastered some of them.
Even the oldest subjects, those In grade 12 (mean chronological age of
slightly under 18 years and mean estimated mental age of 256 mos., or
slightly over 21 years*) have not fully mastered the basics* of class reason-

ing, although they seem to come fairly close. For only two principles have
over 80% of them attained the sufficient condition, and for five principles

(or combinations)

tion.

over 25% of them have failed to meet the necessary condi-

Their mean score on the test is 73 out of the 99 possible, and their

mean difficulty index is about 73%.

With a group at this state of advancement, the question, "Have they

mastered class reasoning?" needs a

more refined answer than the question

*Whether a mental age of 21

is an
years makes sense, given contemporary IQ theory,
interesting question, but
ivhen we are at this
one which we do not have to resolve here.
the product of IQ
level,
we can simply arbitrarily define 'mental age' as
and CA divided
pect is a better indicator (for
by 100, and use the result as whet we susour purposes) of
mental development than

either Igor CA would be alone.

V-27

explicitly requests.

The answer would probably go something like this:

Probably some of them have mastered all of the basic principles, some have
mastered most and some have mastered only a few.

The above statement is

qualified with the word 'probably' because our analysis is of groups, not
of individuals.

The above answer would explain the data.

Thus we do not find full mastery of class reasoning even at age 17-18.
Hence Piaget's characterization of the concrete operations *period as one in
which class logic can be done deserves qualification, if our data are taken
at face value.

Class logic is not, under the cultural conditions of our sub-

jects, fully mastered during the concrete operations period. Not until
some years later is mastery of the basic principles of class logic approached.

However, Piaget's contention that class logic is easier than sentence

logic receives some support from the above analyses, though the stage form
of his conclusion is not supported.

That is, right down the line class

logic scores are generally higher than conditional logic scores (this assumes
that conditional logic scores are representative of scores that would. be
Obtained on a complete sentence logic test). But it is not a matter of one
being mastered at one stage, and then the other being mastered at another
stage.

Instead neither is fully mastered by age 17-18, and both appear to

be developing fairly regularly up to that age.

Mainly because the conditional and class reasoning tests were given at
different grade levels, we did not perform tests of statistical significance for
the differences between performance on them, feeling that such compari-

sons could be made at some other time on subjects who take both tests. But
Graphs V-1 and V-2 do suggest differences in overall performance on class
reasoning and conditional reasoning. It is the sort of difference suggested

by these graphs that supports Piaget's implicit contention that class logic

v-28
is easier than sentence logic. Naturally the support must be qualified by
the lack of tests of statistical significance, which we feel should wait
until more directly comparable scores are available.

Is the truth-validity charactrelniclelemAR5211:127

The data suggest an answer of "Partly" to this question. Tables IV-8
and IV-9 of Chapter IV present the mean difficulty indices for the suggestive

items, of Which there were twelve in each test. These suggestive items, it
au.
will be remembered, usitsuch that the truth status of the conclusion is different from the validity status of the argument, so a person who cannot

judge the validity without being swayed by his beliefs will do poorly on
these items.

The mean difficulty indices for the conditional reasoning test suggestive items are 41.3 and 53.4 respectively for grades 5 and 7; hence the
average suggestive item was marked correctly about 41% of the time by 5th
graders and 53% by 7th graders.

On the class reasoning test the correspond-

ing figures are 40% and 47% for 4th and 6th graders. These percentages
indicate that there is a degree of achievement of this characteristic, but
that the achievement is not complete.
The percentages for the 11th and 12th graders on the conditional and
class tests respectively are 60% and 67%. These show greater achievement of
the truth-validity characteristic, but still indicate that the achievement

is yet incomplete.

This result is of course to be expected from the

findings

reviewed in Chapter IV under the topic of Wilkins' suggestive logic content.

In sum the truth-validity characteristic is partly achieved by age
11-12 and to a greater extent by age 18, though it is still incompletely
achieved on the average by students of the latter age.

There are incidentally

interesting differences between conditional and class reasoning on this truth-

validity characteristic, differences which will be discussed under the next

question.

V-29

6.

Within each type of logic, is there a development of one sort of thing

before another.

If so, what is the nature of this differential development?

Since there are virtually no suggested answers from the literature on

this question, and although some of the differences that will be suggested
in this section seem fairly clear-cut, this research is largely exploratory.
Theoretical background and/or replication are needed.

Differential Development of Knowledge of Principles of Logic.

1)

The Fallacies.

The most striking difference among the raw scores,

difficulty indices, and necessary and sufficient condition percentages for

the principles is the difference between the principles which express the
basic fallacies and those which express the basic validities. The principles
expressing the fallacies are those which specify certain conditions which
fail to entitle one to draw a conclusion of a certain type. Conditional
reasoning Principles 2, 3, 7, 11, and 12 listed in Table II-1 and class
reasoning Principles 3, 6, and 7 listed in Table 11-2 in Chapter II are the
fallacy principles.

As shown by Tables IV-1 and 1V-2 in Chapter IV, these

principles (except for conditional 12) are embodied in Conditional Item Groups

2, 3, 7, and 12 and Class Item Groups 3, 4: 7, and 8. Their symbolic manifestation is also presented in the tables in Chapter IV. For quick identification in Tables IV-1 and IV-2 one can use the answer 'Maybe' as a sign of
a fallacy item group.

All other principles are classified as 'validity principles'. They

describe or indicate a logical move that one is entitled to make. There are
two kinds of item groups embodying the validity principles, those in which
the proposed statement follows necessarily (thus giving a valid argument),
and those in which the proposed statement is the denial of a statement which

follows necessarily.

Items fitting the distinction between these two types of

item groups are keyed 'Yes' and 'No' respectively. Hence items embodying

validity principles are those keyed either 'Yes' or

An inspection of Tables 1V-8 and IV -9 in Chapter IV shows that at the

lowest grade levels (4 for class and 5 for conditional) the 'Maybe' item
groups have the four lowest mean difficulty indices for each of the tests.
These 'Maybt' item groups are the most difficult in each test at that level.
Roughly the same situation holds in the necessary and sufficient condition
percentages given in Tables V-2 and V-3. Conditional Item Groups 2, 3, 7,
and 12 and Class Item Groups 3, 4, 7, and 8, which are the 'Maybe' item
groups, are the most difficult in each test, with one minor exception.* At
this ale level (CA of 10-11; estimated MA of 11-12) students seem

bet6er able to tell that something which follows, does follow; than that
something which does not follow, does not follow.
FUrthermol, there is generally considerable improvement among the
'Maybe' items.

The actual improvement for each item group is shown in

Table V-4, which is derived from Tables 1V-8 and 1V-9.

Table V-4 shows that the largest amount of item-group improvement for
each type of reasoning is registered for a fallacy item group (conditional 7

and class 3).

In class reasoning the next two largest improvements are also

for fallacy groups (8 and 7), and the fourth fallacy group (4) is among the

top ones in improvement.

(2 and 3) are high

In conditional reasoning two other fallacy groups

ones in improvement; the fourth however (12) is the

lowest in improvement, presumably because it is

FUrthermore it is

not a pure fallacy group; it is the only one Which embodies a combination

of principles.
It appears then that although the fallacy principles are the most difficult at ages 10-12, there is great improvement in knowledge of these princi-

ples as students grow older.

*In the sufficient condition table for class reasoning, Item Group 10 intrudes by
two percentage points.

V-31
TABLE V-4.

Item
Group
1

Difference Between Lowest Grade and Highest Grade Mean Item Group
Difficulty Indices; Item Group Answers
Conditional
Class
Difference
7.4

Answer
Yes

Difference
17.3

Answer
No

12.6

Maybe

26.7

Yes

18.0

Maybe

44.2

Maybe

9.3

No

23.3

Maybe

19.4

Yes

15.3

Yes

8.8

Yes

24.8

No

29.1

Maybe

29.1

Maybe

17.1

No

34.5

Maybe

16.8

Yes

15.6

Yes

10

22.9

No

15.7

No

11

12.6

No

26.2

No

12

-5.3

Maybe

4.9

Yes

Note: Item group answers are given because they indicate the type of item
group: 'Maybe' for fallacy item groups; 'Yes' for item groups in which
the conclusion to a valid argument is offered; and 'No' for item groups
in Which the denial of a conclusion to a valid argument is offered.
The question inevitably arises as to whether, using these imp.ovement
figures, the actual improvement in amount of knowledge is necessarily being
compared.

After all there is inevitably little improvement on a test where

there is little room for improvement, it might be argued.

In a way there is point to this view, but it neglects the unique
feature that is built into these tests. They are intended to be tests of
mastery.

If someone has mastered a particular principle or skill, then there

is not much room for further development of knowledge of the principle or

skill, although there might well be room for other kinds of development.
Given our assumption that these are tests of mastery, then ',the comparisons

V-32

of amount of improvement are legitimate. But it ishould10 bourse be remem.

bered that there might well be improvement in other things Which is neglected

in the comparisons that are being made. Nbre specifically, given our assumptions, there is on the whole more improvement in the fallacy principles than
in the other principles for which we tested. This does not preclude there
having been an even greater improvement in some of the principles, skills,
and/or combinations thereof for which we did not test.
This defense of principle improvement comparisons does not extend to
component and total score comparisons, for those are not master: scores.
In those comparisons differences in improvement scores might not reflect
differences in amount of improvement in the thing being measured. Artificial
test ceilings might be operative.
2)

The converse.

Roughly speaking a converse of a statement is the

statement with the parts reversed. One basic fallacy rule is that a statement does not imply its converse.

In other words:

That all A's are B's does not imply that all B's are A's.
'If p, then q' does not imply 'If q, then p'.
These two rules are among the most basic in the practical application of
logic.

They are put more precisely as conditional Principle 7 and class

Principle 3, which correspond to the item groups of the same number.

A striking thing is that in each type of reasoning it is the converse
principle which is the most difficult at the lowest level, and in which there

is the greatest improvement over the years, Conditional Item Group #7

registering a mean difficulty index difference of 29.1 and Class Item Group
#3 a difference of 44.2 (see Tables IV -8, IV -9, and V-4). Similar results

appear in the raw score table, Table V-1. The two elementary converse item
groups start with the lowest raw scores and register the greatest absolute
gain over the range that we tested.

V-33
The results are not quite so striking (though almost so) then the necessary and sufficient condition tables, Tables V-2 and V-3, are examined. In
each case the converse principles tie for the position of greatest difficulty.

Furthermore the simple class reasoning converse item group (p3) does register
the greatest gain in percentage meeting the sufficient condition and the
greatest reduction in percentage failing to meet the necessary condition. And
the conditional reasoning converse item group registers the greatest reduction

in percentage failing to meet the necessary condition. But it does not register the greatest increase in percentage satisfying the sufficient condition,
presumably because not enough subjects even at grade 11 have mastered the
converse principle.
Another interesting fact about the converse principles is that at the
top grade levels in our range they had (as shown in Tables IV-8 and IV-9)
high mean discrimination indices, the highest for conditional reasoning
(53.2%) and among the highest for class reasoning (45.0%0 which is 1.7 percentage points from the highest).

Hence an understanding of the fallacy of

conversion is, of the subtest factors with which we worked, among the most
closely related to total performance on the logic tests.
In sum the greatest improvement within the range of levels measured in
this study occurred for the converse principles of both conditional and

class reasoning.

These principles for ages of roughly 10-12 were the most

difficult and for those around 17-18 were among the most discriminating. This
is in some contrast to the contraposition principles, which we consider next.

The Contrapositive.

Roughly speaking the contrapositive of a statement is the statement with

the parts reversed and negated.

contrapositive.
In other words:

A statement does imply its

V-34
That all A's are B's does imply that all non -B's are non -A's.

'If p, then q' does imply 'If not q, then not p'.
These two rules are put more elegantly as Conditional Principle 6 and Class
Principle 8 in Tables II-1 and 11-2 in Chapter II. They are directly tested
for by Conditional Item Group 6 and Class Item Group 9, as shown in Tables
IV-1 and IV-2 in Chapter IV.
These item groups are of medium mean difficulty at the outset (56.5% and

51.1%) and there is

a very modest amount of improvement in each case (8.8 and

15.6) compared to the other item groups in a given type of reasoning test.
Each of these improvement figures is the third smallest for its type of
reasoning.

Furthermore at the outset each contraposition group has a high

discrimination index (50.6% and 41.5%), which drops considerably by the end of
the period with which we worked (14.3% and 26.7%). This is just the re-

verse of the trend for the conversion discrimination indices.

One wonders how to explain these differences between the conversion and

contraposition principles.

Perhaps the conversion principles are more the

sort of thing that people can learn than the contraposition principles,

ability at which is essentially native, rather than acquired, and develops

early in life, if at all.

That would explain the difference in the amount of

improvement between the two kinds of principles. And it could explain

initially poorer performance

the

on handling the fallacy of conversion, which might

not yet be learned at the early levels. nrthermore, it could explain

the greater early discriminating

power of the contraposition item groups, as

contrasted with the greater discriminating power later on of the conversion

groups, on the assumption that learning plays a larger role in test perfor-

mance as people grow older.

This explanation is in accord with our experience in teaching logic.
We found that the conversion principle, when explained, was fairly easily

V-35
understood by all students, but that the contraposition principle vas either
understood right away or not at all. Contraposition did not seem as teachable
as conversion.

A second possible explanation is that these subjects over the years

receive deliberate instruction in the conversion principles, and not in the
contraposition principles. The fallacy of conversion is regarded as a great
evil by social studies teachers. Contraposition, on the other hand, is in
the experience of the writers regarded as an important tool by only a few
mathematics teachers.
A third possible explanation runs as follows:

The high-scoring students

in the early grades answered the conversion and contraposition items on the
basis of whether the conclusion feels or sounds like the premise. Given that
nobody at these early ages is any good at logic, the ones who worked this war
could get the high total scores and would get the conversion items wrong and
the contraposition items right. In the early grades that would make the
former items more difficult and the latter more discriminating. Later on
students begin to reason logically and the high scorers are those who do so.
They get the conversion items right because they know better, but they miss
the contraposition items because a little knowledge is a dangerous thing.
Hence the conversion items become easier and more discriminating) while the
contraposition items become little easier and less discriminating.*
Research is needed on the reasons for this difference between the conver-

sion and contraposition principles. It could be quite fruitful in helping our

understanding of intellectual ability and development.

Transitivity.

A transitive relationship is one which, so to speak, passes with order

*This explanation was suggested by Prof. Jason Millman.

14,-36

preserved through an intermediary. Implication and class inclusion are

transitive relationships. In other words:
Given that p implies q, and that q implies r, we can conclude
that p implies r.
Given that the class of A's is included in the class of B's and
that the class of B's is included in the class of C's, we can
conclude that the class of A's is included in the class of C's.

The transitivity principles are #5 conditional and #2 class. They correspond

to the item groups of the same number. In the class reasoning test, Item
Group 11 calls for the double application of the transitivity principle.
At the outset in our range the transitivity principle item groups are
of medium ease, with mean difficulty indices of 56.7%, 55.7% and 52.1%. Although
they start out at this level of ease there is still considerable

improvement registered for each in the range covered: 19.4, 26.7, and 26.2

percentage points respectively.

On the whole their discriminating

power as
given in the discrimination indices is somewhat better than that of most of
the item groups.

Thus the transitivity principles, although they start out fairly easy,
are ones in which there is considerable improvement given the range within

which these tests were given.

The Comparability of the Two Types of Logic.

One rather striking feature of the above discussions of fallacies, conversion, contraposition, and

transitivity that it appears that there are

definite similarities between the two types of logic being studied. The

that suggests itself is that

psychologically, although conditional

reasoning is more difficult, there

types of reasoning) and that

thesis

are basic similarities between the two

grasp of the three basic rules of conversion,

contraposition, and transitivity follows similar developmental patterns in

each type.

This is a thesis which can only be suggested. on the basis of

this study.

It must be checked by further investigation.

V-37

b.

Differentia) Development of the Three Components: Concrete Familiar,

Symbolic, and Suggestive.

As shown in Table V-5 there is development in all three of the components on both tests.

In each case the greatest development is on the

suggestive component, next the concrete familiar component, then the systolic
component.

Table V-5 also shows a very striking difference between conditional and
class reasoning in the comparison of components at each grade level. In the
conditional reasoning test the differences among the components are small,
varied, and not statistically significant.
In contrast there are regular differences among the same components on
the class reasoning test, the concrete familiar being easier. At three
(6, 8, and 10) of the five grade levels, the differences between the concrete familiar and the symbolic components are statistically significant
(5% level) and in the other two are nearly so.

The differences between the

concrete familiar and suggestive components are less marked but an inspection
of the chart shows that they are regular. At grade 6 the difference is
statistically significant; at grade 4 it comes within two tenths of a percentage point of being so; and at the other three grades the differences
are about two-thirds of what is needed for statistical significance.
These tests of significance are the relatively conservative Tukey test.*
When the less conservative t-test was performed on the above differences at

each grade level, all except those at grade 4 turned out to be statistically
significant.

Furthermore, using the t-test, none of the conditional reason-

ing component differences are significant. Because of a conservative leaning,

4

we report the Tukey test results in Table V-5, but do find the t-test results
rather striking, so we mention them too.

*See: Ryan, Thomas, A.

"Nhltiple Comparisons in Psychological Research ".

Psychological Bulletin, Vol. 56, No. 1, Jan. 1954.

TABLE Y-5. Comparisons of Mean Difficulty Indices of Three Components: Concrete Familiar, Symbolic,

and Suggestive.

Dirt.**
Needed

3 means 22.9 27.5

26.6 26.1

Needed
2 means

24.6 24.1

Dirt.***

21.2

25.4

19.2 19.2

17.1

13.5 17.1 17.7 17.7

15.8

14.6 18.5

*Underlined differences are significant at the 5% level.

**Tukey test of significance at 5% level.

The difference given is the differenceneeded for

significance in comparing the top and bottom of threemeans.

***TUkey test of significance at 5% level.

If, and only if, the difference betweentop and

bottom mean is significant is this test

needed for
applicable. The difference listed is the difference
significance in
making either of the other twocomparisons.

V-39
Except at the lowest level (grade 4) the suggestive component is not
so difficult as the symbolic component. The differences are not statistically
significant using the Tukey test, as can be seen from an inspection of
Table V-5.

One wonders why there are these regular superiorities in the Blass
reasoning test of the concrete familiar component over the symbolic and
suggestive components, but not in the conditional reasoning test. Two ideas
have occurred to us, one for the CF-SU comparison and the other for the
CF-SY comparison.

Each will have to be checked by further research.

It might be that something analogous to the suggestive component is

already a part of all of the conditional items because of their "iffy"
What the subject is asked to suppose is the conditional statements

nature.

on the conditional reasoning test is not simply that something

is the case,

but rather that, on another supposition (the if-clause) something would be

the case.

In the class reasoning concrete familiar items, if the subject

accepts the initial invitation to suppose something, then he is working with

what he believes to be true -- for the purposes of the test. But with the
concrete familiar conditional statements what he is working with, though he
might well believe it to be true, is the implication of another supposition.
Thus he is forced to think in terms of what is implied, rather than what
is true.

Perhaps the above hypothesis is unsatisfactory -- either because it is

false, or because it is vague.
standable.

But it does appear reasonable and under-

It would explain the difference between the two tests on the

CF-SU comparison.

According to this explanation, the conditional CF items

are also SU items and thUs have the SU difficulty built into them. Thus we
would expect no difference between the CF and SU components on the conditional
test.

The class reasoning CF items on the other hand do not have this

v-ito

component built in, so we can expect a difference, on the assumption which is

supported by the literature cited earlier that SU items are generally more

The research that has been done has used syllogisms,

which are categorized as class reasoning.

This hypothesis might also explain the apparent greater difficulty of

One might well expect this, if something

analogous to the suggestive component is built into all the items on the
conditional reasoning test.
Turning to the CF-SY comparisons, the greater difficulty of the symbolic
items on the class test is not surprising, since working with variables
presumably more difficult than working with familiar categories.

is

On the

other hand there is no comparable difference in the conditional reasoning

test. This is surprising. One possible explanation is that the symbolic

component in the conditional test does not use variables, but instead talks about
the existence or non-existence of letters.

That is, the items that

are used are of this type:

If there is an X, then there is a Y.
There is not a Y.
Therefore there is not an X.
They are not of this type, in which the letters are clearly variables.

If p, then q.
Not q.

Therefore not p.
Perhaps it is the use of symbols as variables, rather tlisn simply the use of

symbols that makes for difficulty. If so, then the difference between the
class and conditional reasoning tests on the ccmparison between the concrete
familiar and symbolic components is understandable.

v-'11
It appears then that the order of difficulty of these three components,
if one accepts the auxilliary hypotheses mentioned above, is as
our subjects:

concrete familiar, suggestive, and symbolic

follows for

(listed in order

of increasing difficulty). This order of difficulty, which appears starting

roughly at ages 12-14, is the same order that Wilkins found (1928, p. 77).

CHAPTER SUMMARY

This chapter presents a review of the literature and the results of our
study of the development of knowledge of conditional and class logic of
students roughly in the chronological age range 10-18 under the environmental
conditions of a contemporary upstate New York area. These students had not
to our knowledge been deliberately exposed to instruction in deductive logic.

The Literature.

The review of the literature focused on the work of Jean Piaget. Fou
basic features of his formal operational period of thought, which he holds
runs from age 11-12 onward, were described. They are 1) possession of the
truth-validity characteristic; 2) ability to operate within a combinatorial
framework; 3) ability to control variables; and 4) ability to do propositional
(sentence) logic.

Only the first and fourth, as we interpret these features,

are basically logical.

The testing of Piaget's views, because of their vagueness, was not the
primary emphasis of this chapter. Instead it was concerned with questions
which are interesting, practically important, and were generated from a
consideration of Piaget's interests and concepts. Hence this part of the
study is to be considered an attempt at extension and refinement of Piaget's
work, rather than primarily a testing of his views.
The review of the literature (including Piaget's work) suggested the
following:

v.lte

that there is a development of logical ability as children

grow older.

That no stages in this development have definitely

been identified.

That there is no work on the

conditional logic in adolescence.

That one might infer that Piaget thinks that class logic is
mastered by age 11-12.

That the truth-validity characteristic is not achieved by

age 11-12.

That there is practically no study of the different develop-mental

patterns of different principles and components of logic.

extent

of

mastery

of

Findings.

Briefly (and with many qualifications omitted) the findings are as

follows:

In this age range there is a development of ability to do logic as

students grow older.

If there are stages in this range, they are not noticeable

at the level of refinement of our measuring techniques.
The basic principles of conditional reasoning are not generally
mastered by age 11-12, nor by age 17.

The basic principles of class reasoning are not generally

mastered by age 11-12, nor are they fully mastered by age 17.

The truth validity characteristic (the ability to consider

questions of validity regardless of belief in truth of the
parts of an argument) is not attained by age 11-12, nor by

age 17.

The patterns of development and mastery of principles of logic

vary, but there is considerable similarity between the two
types of logic studied.
The principles expressing the basic logical
fallacies are the most difficult at ages 10 -12, but are also the ones
ment over the range
in which there is generally the most improvestudied.
The most extreme example is the principle that a statement
does not imply its converse. The principle of contraposition is one
at medium difficulty and does not become
which in this range starts
students.

much
easier
this range
at for older

The transitivity principle starts in

medium difficulty, but is considerably easier for older

students.

11.43

Of the three components tested, generally the concrete

vas the easiest; next came the suggestive; and most
familiar
difficult was the symbolic.

Of the three components tested, the greatest difference between

earlier and later ages was in the suggestive component.

Class reasoning appears to be easier than conditional reasoning

at all levels.

Needed Research.
The following types of further research are called for:
2.

A more refined study of the possibility of stages of development

of logical reasoning ability.

3.

An extension of the investigation to younger age groups, using

perhaps the first half of each test.

This investigation should be repeated on different kinds of

children.

1.

Investigations siniflar to this one should be performed for other types of

logic and for loose reasoning with these two types and the others; required
first are prior investigation of these other types of logic and of loose
reasoning.

A check on the comparative difficulty of the two types of

reasoning, and the principles and components, using both tests
on the same students.

A check should be performed

variable.

on the effect of a symbol's being a

The question of the existence of basic psychological

between conditional and class
reasoning
differences
should be investigated further.

Further tests should be performed on the hypothesis that

mastery of the principle of contraposition is more related to
inherent ability than is mastery of the principle of
conversion.
Perhaps controlled amounts of time could be spent
teaching each at various age levels and the amounts of
learning analyzed.

CHAPTER VI.

The Development of Readiness to Neater Logic

The topic of this chapter is to be clearly distinguished from that of

the last, because we are here concerned will what students can come to bow

-- not with what they know already.

First of all, because readiness

is a
key concept in this chapter, we shall present an examination of the concepts
readiness and readiness to master a principle. Then, after noting the dearth of
literature about readiness to master logic, we shall describe the experimental
procedures, including the teaching that we did, and present the results of this
inevitably limited study of the readiness question. These

results will be organized around two questions:

What did our subjects learn?

Can we state that others are ready to master?

WHAT IS READIRESS?

The forthcoming discussion of readiness and readiness to master a

1.12.....2CIASI is

logic.

oriented toward the ultimately practical concern, the teaching

of

Hence some qualifications -which might have to be made, were we

considering these concepts

will not be nade.

in

other contexts, will not be crucial here, and

Fbr our purposes the ensuing discussion should suffice.

At some other time and in

some other place, a general discussion of the

concept readiness would be in order. This is not to deny that much of what

follows is general, but simply to limit the

problem.

This is not

intended

to be a definitive treatment of the concept readiness.

arAt21122111-1-AMTEI
When one inquires whether

a-person is ready to do something, one is

not asking whether he has done the something, and one is not simply asking

whether he Eill.do it, for

he is ready to do it.

someone might never do the thing, even though

One is asking in part %tether he has the capacity to

VI-2

do it. But not only this.

Someone might have the capacity to do something,

but be declared not ready to do it because he is not willing to do it. To

say that a person is ready to do something is to say that he has

the capacity and sufficient willingness to do it, which in turn implies

that he probably will do it, given suitable conditions.

Dispositional Terms.

In philosophy there is a technical name for terms that apply to traits

Which are evidenced under suitable conditions. Such terms are called
' dispositional terms'.

The standard example is 'soluble'. To say that a piece of

sugar is soluble is not to say that is has dissolved, nor that it will ever
dissolve, but simply that under suitable conditions it will dissolve.

Since readiness is a trait that is evidenced under suitable conditions, the

term 'ready' is a dispositional term.

'Mastery of a principle' is also a dispositional term. If one says of
a person that he has mastered a principle, then one implies that under
suitable conditions, the person will behave in a certain manner. For example,
if one claims that a person has mastered the principle that affirming the
antecedent of an accepted conditional commits one to the affirmation of the
consequent, then one is committed to the person's accepting as valid simple

arguments of the form used for Conditional Item Group 1.

So far that has been said about the term 'mastery'

is noncontroversial.

That is, it is not controversial that a person who has mastered the prin- ciple in
question should be able to do something of the sort, under suitable

conditions.

What might be argued is whether the person should be expected

to get correct at least four of Item Group #1 on "The Cornell Conditional

Reasoning Test, Form X".

The particular behavioral expectations that one

VI-3

has might be argued, but that

something of the sort is required hardly

seems arguable.

3.

......1"InaafL11111E11112E-

Tae combination of terms, 'ready'

phrase 'ready to master

and 'master a principle', appearing

a principle', is doubly dispositional, once

for the readiness concept and

once for the mastery concept. Roughly speaking,

to say that a person is ready to

the disposition, given suitable

master a principle is to say that he has

conditions, to develop another disposition,

which, given suitable conditions,

he will show in an appropriate way.

Thus the inference path from

his readiness to master

to statements about what he

More manageable,

the inference path from a person's behavior

has mastered.

'mastery of X principle' given in

procedure for traveling

a person's behavior to a statement about

a principle is not a simple one.

although certainly not simple, is

The operational definitions of

Chapter IV represent a rule of thumb

on such an inference path, a procedure which perhaps

imperfect in detail, but

in the

is

we think satisfactory in general approach.

Elusiveness of Readiness.

In addition to this doubly

mastery, there is another

dispositional feature of readiness for

problem in inferring from behavior to readiness for mastery.

The readiness disposition

disposition to dissolve.

is more like the disposition to explode than the

When sugar has dissolved, it has not lost its

disposition to dissolve, for if

residue will dissolve again.
it explodes, it will not

we let the water evaporate, the sugar

But if some powder is

explosive, then once

any longer have the disposition to explode, no

matter how long we wait around.

a principle -- and then masters it
Therefore we can not get direct

Similarly, if a person is ready to master

-- he is no longer ready to master it.

evidence that a person is ready to master

something (evidence of the sort

is soluble)

that tells us directly that a certain powder

-- although we can get direct evidence that he was ready.

Just as we can directly find

that some powder was explosive, so can

we find that a person waa ready.

explosive when we discover

that it did explode.

was ready to master a principle

principle unfortunately is
knowing whether it exists

before it is exercised.

Once exercised, it is

in the fact that it was present -We want to know whether a

know this before he exercises this disposition.

This

concept readiness to master a principle obviously

makes problems for this

Is there then a
principle?

Readiness to master a

one of the latter type. We are interested in

purposes, like this study.

person is ready, but we want to

characteristic of the

that are retained while exercised and

as they are exercised.

gone, and we are no longer interested

except for special

We discover that a person

by noting that he did master it.

Thus there are some dispositions

some that are lost as soon

We directly find that some powder was

sort of readiness study.

way to find out if a person is ready to master a

Ideally the flay to make this discovery would, be to find an

identical person, provide

identical person masters

the suitable conditions,

and then see if this

the principle. If so, then the subject is ready.

If not, then he is not ready.

There are obvious difficulties
unavailable; even if he
could be, it would be
time we want to see if

here: first, the identical-person is

were, he could not be identified; and even if he

inconvenient to arrange to put him through the paces

our subject is ready. Furthermore, the phrase

'suitable conditions' is vague.

A Substitute for the Identical Person.

every

A rough compromise

as a way of meeting the identical- person difficulties

VI-5
is the study of a range of people. Such study will call for the measurement
of one or more variables which are correlated with the extent, after the
suitable conditions have been provided, of mastery of the principle.
Thc_e correlations could be used in the development of multiple regression
equations, which would serve the function of predictor equations since the
values of the variables, as determined be_ fore the introduction of the

suitable conditions, would be correlated with the extent of mastery, as

determined after the introduction of the suitable conditions. The worth
of these predictor equations would depend on the combined strength of the
relationships between the predictor variables and the variable to be predicted
(extent to which the principle has been grasped).
Then one or two lines must be drawn separating that degree of achievement which is deemed to indicate non-mastery and that degree which is
deemed to indicate mastery.

These lines need not necessarily be sharp,

but for some purposes, it is convenient to make them artificially sharp.

This artificial sharpness, which is the state of the lines drawn by the
operational definitions described in Chapter IV, 'sill result in some mistakes in the treatment of borderline cases, but for practical purposes,
this likelihood of mistakes must be accepted.

Some of the mistakes can

be avoided by deliberately introducing an area of uncertainty in the present

study (exactly four is the area of uncertainty).

Thus, neglecting for the moment the difficulty inherent in the

vague-

ness of the term 'suitable conditions', a rough scheme might be developed

for determining someone's readiness to master a principle. The values of
the correlated variables for a given person can be put into the prediction
equation and a predicted degree of grasp of the principle comes out. On
the approach (the one taken here) which makes use of an area of uncertainty,

VI-6

there will be two lines, which will be called the "necessary condition
line" and the "sufficient condition line". If the prediction falls below
the necessary condition line, then we judge that the person is probably
not ready to master the principle.

If the prediction falls between the

two lines, then we withhold judgment. If it falls above the sufficient

condition line, then we judge that the person is ready to master the

principle.
On the simpler approach Which does not make use of an area of uncertainty,

the judgment would simply depend on whether or not the prediction falls
above or below the line.
Judgments about groups would not be as subject to error as judgments
about individuals. A procedure for making judgments about groups will be
described later.

Suitable Conditions.
la

The phrase 'suitable conditions' in the previous analysis of readiness

reminds us of the practical considerations that enter into our use of the
concept.

Suppose that it would take six months of full-day instruction to

teach a certain level of child that affirming the consequent is

We would be tempted to say that he is not ready to learn this fact yet,

because it would be just too much trouble.

teach

it to him in fifteen minutes, then we would say that he is ready.

Hence if we say that a person is ready to master something, we imply that

it would not require an unreasonable amount of effort to teach him.
Since people will differ on what they consider to be an unreasonable
amount of effort, an ideal piece of readiness research would enable one
to predict the results of varying amounts of effort and let each person
judge for himself whether the effort required for a given result is reasonable.

VI-7

But the difficulty of performing this task, the production of predictability for
varying degrees of effort, is great.

The current study has attempted

to investigate the results of one given amount of effort, roughly indicated

three weeks of group instruction (size about 20-30) for about

or 50 minutes per day, such instruction provided by a teacher with special

training in logic. In a later section we will say more about the

teaching that we did, but the previous statement suggests some of the
dimensions of effort and to most people will suggest some that we had to
ignore.

The phrase 'suitable conditions' refers not only to the amount of

effort, but also to the nature of the effort. The effort must be of the
As a result of this fact, a piece of readiness research like

the present one is again in a difficult position.

advance which is the right sort of effort to make?

is tried, if it fails, there is always the possibility that another would

have succeeded -- that they really were ready, if only we had tried a
different approach.

The inevitable compromise is to try an approach that seems feasible

to an experienced person who knows well the subject matter to be taught,
or, If time and funds permit, to try several of these. But it must be
remembered that, following such a course, one can much more easily declare

that a certain sort of person is ready than that he is not ready.

find that people of that sort do master the thing in question, when provided with a given sort of instruction, then we can say that a person of

On the other hand if we find that people of

that sort do not master the thing in question,then it is with much less

VI-8
confidence that we say that such a person is not ready, because there
might be another feasible and successful way of conducting the instruction.
This is one of those rare cases when a positive answer is easier to give
than a negative one.

In the present research we were able to provide at each grade level

only one set of 'suitable conditions'. They were the best we could do,
given our limits of time and funds. But the results of this readiness
study apply only to the sorts of conditions that we were able to provide.
This is an important fact about these results.

Summary.

This analysis of the concepts, readiness ard readiness to master a

principle, has produced a number of interesting features and some difficulties

that any readiness study must face.

subject's

behavior to an assertion or denial of his readiness to master a principle

of logic is a complicated one. Here is a list of the complications
discussed above:

The concept is doubly dispositional.

Readiness is an elusive disposition in that once shown, it

no longer exists. It is explosively dispositional.

Readiness is a practical notion in that the means needed to produce the mastery for which the subject is ready must be feasible.

Since there are possibly other ways of achieving said mastery,

it is difficult to deny conclusively that a subject is ready
for the mastery.

Overlooking the vagueness of 'suitable conditions', briefly the

proposed analysis is as follows: Tony that a subject is ready to master

Cisatleasttp...2ziehasthedissirincile)tioikiiven suitable

conditions to devel

the di:..sition to show

riven suitable conditions

VI-9
the sort of behavior that a person Who has mastered principle X would shor.

Admittedly the foregoing analysin is circular with respect to the concept

mastery., but the focus here has been on readiness.

Mastery was discussed

in Chapter IV and can be argued separately. Its analysis is assumed here.

Any readiness study must simplify by trying to find out just that can
be acquired by certain sorts of people When they are exposed to a certain
set (or limited number cif sets) of conditions.

After the research is done

and the results announced, there is still much cautious judging to be done.

The inference leap to another person is one that must be undertaken with
care when the relevant factors have been examined. The judgments about
the feasibility of the means selected must be made. If the subjects used
did not succeed, then one must consider the question of whether better
methods might yet be developed.'
Hence the results of a readiness study are likely to be rather modest.

THE LITERATURE

There is practically no literature on the question under consideration in this chapter. Piaget's bountiful contributions concern themselves
only with what children know, not what children can learn. The only thing
that we have been able to find that is relevant is the effort of Hills

(1961) to see the effect

on children in 1st through 3rd grades of a rather

restricted way of teaching: telling children, when giving them a logic test,

whether each answer is correct before they go on to the next problem.

But

of course Hills' main purpose was other than seeing the effects of teaching
logic.

The current study is the first study of readiness to learn logic with
which we are acquainted.

VI-10

PROCEDURES

Assignment of Grade Levels to Class and Conditional Logic; The Pre-Test

At the beginning of the spring semester, 1964, the pre-test was ad-

ministered to all subjects, who are described in Chapter III. At grades

4, 6, 8, 10, and 12, the pre-test

was

The Cornell Class Reasoning

Test,
Form X ", at grades 5, 7, 9, and 11, "The Cornell Conditional

Test, Form X".

Reasoning

These tests are described in Chapter IV. Within a week

after the administration of this Idretest, formal instruction in either class

or conditional logic, depending on the grade le-A,

was commenced with the

LDT's, of 'which there was one class (size about 20-30)

levels 4 through 12.

at each of grade

Class logic was taught to those that took the class

logic test, and conditional logic to those who took the conditional logic
test.

This pre-test was given primarily in order that the scores might be

used in the prediction equations. It also did serve as a control in the

analysis of covariance comparisons; served to alert the LDT's to the nature

of the content they were to learn; but unfortunately also presumably

served to provide some logic instruction for the control group.

Total Teaching Effort

The initial agreement with the school system called for some member

of our staff, a person trained in logic and experienced in teaching at the

grade levels to which he

was assigned, to take over each of the LDT classes for

one period per day for 15 instructional days for the purpose of

instruction in logic.

Each daily period was to last from 4o to 50 minutes.

Naturally minor modifications of this plan were required to fit specific

situations, but it was essentially followed.

VI-11

The Actual Teaching.

Zech staff teacher vas instructed to teach the logical principles roughly

in order, using whatever style of teaching seemed to him to be most appropriate

and going as far down the list as he could in the available time. No staff

member taught more than two classes per day, and when there were two classes
per day they were in the same type of logic. Thus staff teachers were provided
with time to plan and to develop written exercises. Each member of

our teaching staff vas also a graduate student at Cornell University.

The following general procedures, which are in part based on Crombach's
recommendations (1954, p 272), were followed:

Each principle to be taught was made explicit somewhere in

the course of instruction, though it might have been near the beginning,
middle, or end of the instruction aimed at that particular principle.
The language in which the principle was stated, and the person
(teacher or student) by whom it was stated varied frameless to class
and principle to principle.

The use of technical terminciogy V82 kept to a minimum,

but was not completely avoided.

c.

Frequent written exercises and many examples of varied sorts

were used.

Each staff member was free to develop his own exercises

and examples, though they used each other's ideas. Examples of the
teaching materials will be found in the Appendix.

A modified Idler circle system was used as a model in doing

class reasoning.

There was no discussion of any items on the Cornell Deduc-

tion Tests, nor was practice given in the specific mode of response
used in the tests.

VI-l2

Homework was given in grades 7-12.

Although students were provided with an evaluation of their

work, the students were told that their degree of

did not bear upon any school grades that they were to

achievement in logic
receive. They

were also

told that a record of their progress would be given to their

teachers and principals.

h. No attempt was made to counteract the positive effects of

novelty, notoriety, attention, and whatever else might go into the
There were two

Hawthorne effect.
reasons for this:

First ve were

trying to find out what was possible and took advantage of whatever
motivation was available.

Second, there were disadvantages under

which our instruction suffered as a result of its not being part of

the regular school program. Any compensating factors were welcome.
A rough estimate of the amount of time spent on teaching each principle at each grade level was attempted by our staff teachers. This is a
very difficult thing to do, since no teacher times himself in this way and
since much teaching is aimed at more than one thing at a time. Furthermore
we used an &ler circle system for teaching class reasoning, making it
difficult to allot the time explaining the system to any particular prin-

ciple. That is, the techniques of the use of circles in doing class logic
apply to all the principles; hence much of the instruction applies to all
of the principles, making class reasoning time allotment somewhat meaningless,

but not completely so.

The principle of the symmetry of exclusion, for

example, has a corresponding diagram.

We did feel obligated to give some indication of the differential
effort given different principles.

Educational research so frequently

suffers from a lack of information about teaching that is done. We hope

VI-13
TABLE VI-1 Rough psi

Estimates of the Effort kvoted to the

Teaching of Each Principle at Each Grade Level.

Conditional Reasoning
Principle

Item Group

lo

11

ec.-apination

Grade
80

9
80

11
70

150

60

60

65

125

75

75

65

70

80

100*

125

To

25

40

40

To

50

7o

710

65

5o

4o

4o

7o

40

45

45

3o

35

15

15

30

5o

loo

loo

6o

30

30

6o

10 & 12

85

2
3 & 4

75

65

70

75

4o

75

6o

70

To

40

50

60

50

55

20

25

65

45

40

120

120

120

10,11U2

120

120

120

Colibination

200
li

225
150

70
160

75
165

15
165

*Mote:

Times are given in minutes.

'100' stands for 100 minutes.

that the reader will take this estimated time allotment for that it is:

a series of very rough and sometimes not very meaningful guesses. If we

had it to do over,

we would probably work entirely with principles of the

circle method of doing class reasoning

rather than with the more elegant,

but virtually unteachable principles

estimates appear in Table VI-1.

which we used.

In any case rough

I. caa be seen that the greatest

amount
of effort at the lower grades

was devoted to the lower-numbered principles,

whereas for the upper grades the

effort was fairly evenly distributed among

all the principles.

These estimates do not add

for a number of reasons.

up to the total time spent in the classroom

To specify a few: time was spent in organization

and control of the classroom;

some time was not allotted because it vas

used in the teaching of general

notions underlying these principles; and

these allotments are imprecise guesses.

The Post-Test.

Approximately six weeks after the

test was administered to all

taught by members of

conclusion of instruction,

a post-

subjects, those to whom logic was deliberately

our staff (the UT's) as well as those to whom it was

presumed not to be taught (the LVDT-1's). The post-test vas the same test
that was administered as a pre-test, the class reasoning test to grades 4,

6, 8, 10, and 12; and the

conditional reasoning test to grades 5, 7, 9, and

A check vas made to see if logic had been taught between test

administrations to the students to whom we did not teach logic; no evidence

of such instruction could be found.

RESULTS

The results of this readiness study are divided into two parts, one dealing
wIth the nature and extent of learning of logic that vent on in the

VI-15
groups to which we deliberately taught logic, and the other dealing with
attempts to predict What other students could learn, given comparable
instruction.

Each of these parts is concerned with a readiness question.

The first part asks What the students in this study were ready to learn
before we taught them logielthe second asks what some other students are
ready to learn.

General Qualifications.
Four important qualifications should be kept in mind as one reads

this section.

First we were necessarily working with small numbers of

subjects (about 20-30 at each of nine grade levels). The reason for this
was that it was important to worm intensively, rather than extensively in
this study.

In particular it was necessary to make sure that proper logic

was taught rather than something else. And it vas desirable to give the
staff teachers adequate time to plan, prepare written exercises, and read
students' papers; they were pioneering the teaching of this subject matter
at grade levels lower than those at which it is ordinarily taught.
As a result of this small number of subjects results are more erratic
than they presumably otherwise would be. And also in our comparisons
between the LDT's and the LED-1's some actual differences have inevitably
failed to reach statistical significance.
In order to state the other three qualification we must first distinguish

among various possible causes of improvement in scores on a test of knowledge

of logic.

Although these distinctions and the ensuing discussion are

oversimplified as a result of neglect of interaction between factors, the

points that result frcet the discussion do not suffer therefran.

pursue the simplified line.

Possible causes of improvement in scores are:

Deliberate teaching of logic.

Hence we

The taking of logic tests which are intended to measure a student's

knowledge of logic. There are two features to be distinguished here:
that which results in the learning of logic.
b) that which results in the learning of how to take this
and/or other tests (test-wiseness).

Other school influences.

Influences outside of school.

Maturation that does not depend on contributions from

the environment.

Some of these factors can be introduced at. the discretion of the

school authorities and some are more or less beyond their control. Factors
1 and 2 can be introduced by the school authorities. Factors 3, 4, and 5
are increasingly out of their control.
As a practical matter, when we ask whether a student is ready to learn
something, we are concerned with the probable results of the deliberate
introduction of factors under our control. Hence we are in this study
primarily concerned with the effects of factors 1 and 2.
Since the test-wiseness part of the second factor will give spurious
results, we would like to discount these. When we simply ask what our students
learned as a result of teaching and look only at the before and after test
results, the results of this test-wiseness factor creep in. When we ask,

however, whether they have learned more than they otherwise would have,
and use a comparison with a control group to help answer the question, the effect
of this factor is presumably controlled for to some extent at least.

With the other part of the second factor, the learning of logic as a
result of taking the

test, the situation is somewhat the opposite.

We

velcame the operation of this factor when we ask what the students have

learned as a result of teaching (although its impact is undetected to the

extent that it increases the pre-test score).

When we use control groups,

VI-17
this factor is controlled for, even though we

would rather that this not

happen. After all, our interest is in the learning that occurs as a result

of what we do. So the results of the comparisons with the control groups
mask the results of actually taking the tests,

Which can be standard parts

of instructional procedures.
Another source of error is the operation of factors

classes, both LET's and LNDT-1's.

4 and 5 on entire

Ektraneous occurrences, like a loud

noise while the test was being taken, the presence of a smart

in a class, a superior regular teacher who

-aleck or two

predisposed his class to

attentiveness and eagerness, the beginning-of-ile-year assignment of members

to a class on the basis of a criterion which is motivational or related

to motivational factors, etc., are examples.

That such things probably

occurred and that we had no check on them other than alertness and care
in avoiding them is a weakness of the study. As we will explain later)
we suspect that some factor of this sort operated to the detriment of the
LET's in the 9th grade.
Tb have provided a statistical check on this sort of thing it would
have been desirable to use a number of groups with random assignment as
LET's and LNDT-l's.

Such a procedure would have been considerably more

expensive than the one we followed.

In summary the general qualifications we have specified are as follows:
a.)

There were a small number of subjects at each grade level.

b.)

Test-wiseness will result in spurious changes When the LET's

alone are being considered, but will be controlled for in
comparisons with the LNDT-1's.

c.)

Learning of logic attributable to the taking of the tests will

justifiably have an effect on the LET pre- and post-test
comparisons (though some of this effect on change in knowledge
will be hidden), but unfortunately will be blotted out in the
comparisons with the LNET-l's.

VI-18
d.) Factors affecting total class groups can without detection
(except for internal consistency evidence) give spuriously
high or low scores.

What Idd. Our Subjects Learn?

The answer to this question gives an indication of what they were

rem

to learn.

Generally speaking, they did not during the instructional

period learn much conditional reasoning until the upper secondary level, at
which time they made a vast improvement, primarily in avoiding the fallacies.
On the other hand in class reasoning moderate improvement was registered at
all'levels with which we worked, except the lowest. These general statements
need amplification and qualification.

Conditional Reasoning.

Total Scores.

The situation in conditional reasoning can be seen in various ways,

each of which emphasizes different aspects of the overall situation. First
let us look at total scores.

One can visually compare the mean IiVr pre-

and post -test total scores on the conditional reasoning test as given in
Table III -3 in Chapter III.

One cannot see much overall improvement until

Grade 11 where the difference is 17.4 points. This is a rough =sophisticated

comparison, but it has its merits.
One can also compare LET's with the LEET-1's on the post test, taking
into account differences in the ire-test and IQ. The results of such a
comparison are summarized in Table VI -2.

The figures are presented in detail

in Table A.5 in the Appendix. These comparisons show that in grades 5 and
there is no statistically significant superiority one way or the other. In
grade 9 there is a statistically significant superiority favoring the students

who were not taught logic, and in grade 11 there was a greater superiority
favoring the students who were taught logic. The superiority of the taught
11th graders is quite striking, with a difference of
means on the post-test.

15.2

points in adjusted

(The superiority is 22.2 points before the means are

VI-19
Adjusted.)
The situation in grade 9 is puzzling.

Not only did the control group

have a significantly higher adjusted mean, but the mean score (unadjusted)
of the taught 9th graders was actually lower

(2.6 points) on the post-test

than

it was on the pre-test. Possible explanations that have occurred to

US are the following: First, it might be that at this level teaching is

just beginning to take hold, but just enough to be confusing.
might be that at lower levels, people work on the basis of

That is, it

"feel" only and

that at this level the teaching that has occurred has been effective enough
to interfere with the "feel" method of judging arguments, but not effective
enough

to provide a reliable replacement.

A second possible explanation is that the particular specimen of teaching

was simply confusing. A third possibility is that the problem

lies with the students themselves. It might be that there was an important
difference between the two groups which did not show up on any of the measures
that we used.

We do know that the 9th graders whom we taught were alleged

to be of high enough ability to be taking algebra, but for some reason or

other were not doing so.

Our staff teacher reported that they did not care

whether they learned logic.

have been a problem class.

It might

We tend

to favor this the third explanation, which, if the proper one, makes suspect
the findings about conditional reasoning at the 9th grade level. It was
only at this level that such motivational problems were reported by members
of our staff.

Component Scores.

As can be seen in Table VI-2 the situation is roughly the same as viewed
through the component scores.

The llth grade group that was taught

conditional reasoning did significantly better on all three of the components

VI-20

TABLE

VI-2.

Statistically Significant Post-Test Differences Between

Students Who Were Taught Logic and Students Who Were Not
Taught Logic, Using Analysis of Covariance with Pre-Test
and I.Q. as Covariates.

Conditional Reasoning
5
11
9
7
27
24
17
26

Grade
Niatlit2

26

25

Total Score

23

Class Reasoning
10
4
6
8
25
25
27
22

22

22

34
R

24
-

18

12
23

17
R

Component

CP

SY

..

MD

MD

..

41M

MD

...

..

..

...

..

MD

MD

GM

4m

..

'.

MD

ND

..

GB

MD

al

IID

Mb

GP

..'

UV

MO

ca b

s:..cycs

SU
Item Group

9
10

MI

MI

11
12
lo

s use

..gcateas t a

R
.C1

ference favoring the group that vas taught logic. The syMbol
is used to indicate a statistically significant difference favoring
the group that was not taught logic. A dash is used to indicate lack
of statistical significance. The 5% level was used throughout.

A yard of caution:

The sixteen comparisons described in any one

column above are statistically dependent.

VI-21
built into the test.

This did not happen at any of the other grade levels,

and at the 9th grade level on the concrete familiar component there is another
reversal.

Item Group Scores.

The great superiority of the 11th graders who were taught over those
who were not taught seems to lie mainly in the fallacies, but also in
contraposition, basic and practical, and perhaps with affirming-theantecedent items.

By 'basic contraposition' we mean the valid move from

'If p, then q! to 'If not q' then not p'.

By 'practical contraposition'

we mean the valid move from 'If p, then q' and the denial of 'q' to the
denial of 'p'.

The former appears as Principle 6 and Item Group 6. The

latter, which we have also called 'denying the consequent' (Chapter II),
appears as Principle 4 and Item Group 4. The two forms are logically similar,
as can be seen by saying them over to oneself.
The four fallacy item groups (2,3,7, and 12) and the two contraposition item groups (4 and 6) stand out uniquely on the covariance
comparison.

On all of these item groups, but only these item groups, are

the 11th graders that we taught significantly better than the ones we did
not teach.

This finding is perhaps the most striking one of this study.

It also comes through in Tables L-7 and A-8 in the Appendix, which
give item group pre- and post-test measures on the groups to which we taught
logic.

A portion of those tables is presented in Table VI-3, which shows

the difference between pre- and post-test measures for each item group in
the 11th grade.

The measures are mean difficulty indices, percentages

meeting the sufficient condition, and percentages failing to meet the

necessary condition. As can readily be seen, these six item groups stand
out, and the fallacy item groups stand out most srikingly.

VI c2

TABLE VI-3.

Conditional Reasoning Pre- and Post Differences for llth

Graders to Whom Logic Was Deliberately Taught.

N=26
Improvement in

Improvement in
Percentage Meeting
the Sufficient
Condition
15

Item
Mean Difficulty
Indices
9

Group
1

Increase in
Percentage Failing
to Meet the Necesnary Condition*
- 8

2**

50

3**

53

73

-69

4***

11

27

-8

6***

19

39

-19

7**

27

42

-35

65

-61

-11

15

-13

-15

lo

11

- 1

44

50

-65

12**

*A negative number here indicates

a reduction in those failing to meet the

necessary condition and thus shows an improvement.

**The fallacy item

groups are marked by a double asterisk.

***The contrapoelon item

Note:

groups are marked by a triple asterisk.

'which
This table is taken from Tables A-7 and A-8 in the Appendix,
all grade
present the
pre- and post-test scores and differences for

levels to vhich

conditional reasoning was taught.

VI-23

On the other hand there is

among the students taught logic a slight

worsening on the 'only-if' item

groups, numbered 8 and 9.

Perhaps the

word 'only' was previously quite clear, but that what they were taught about
the word 'if' by itself confused their understanding of 'only if'.

reference

to Table VI-1 shows that about an hour was spent teaching the

'only-if' principles.

Perhaps this amount of time was insufficient.

It is interesting to note that there was no improvement on the 'if-and-

only-if' item group (numbered 11),

to this concept.

even though about an hour was devoted

Perhaps again the distinction between 'if' and 'only if'

was confused.
There is slight improvement on the transitivity item groups, numbered
5 and 10, but one does not know whether this represents a real improvement
or not. There is a greater improvement on the most basic move of all, the
affirmation of the antecedent (represented by Item Group 1), but not enough to
reach statistical significance in the compariscn with the students not

taught logic.

At all grade levels, however, there is a lack of a really

striking teaching--Isuu-d improvement on this, the most basic conditional

reasoning principle.

This is shown in Table VI-41 and is also suggested

by the lack of statistical significance of the post-test differences

between the LDT's and the LNDT-1's, which lack can be noted in Table VI-2.

TABLE VI-4.

Conditional Reasoning Pre- and Post-Test Differences for

Principle

the Affirmation-of-the-Antecedent Principle.

Improvement in
Grade

Improvement in
Percentage Meeting
the Sufficient
Condition

Increase in
Percentage Failing
to Meet the Necessary Condition*
18

27

Mean Difficulty
Indices
9

7
9

24

17

11

- 3.2

11

26

* A negaiive number here

15

- 8

indicates a reduction in those failing to meet the

necessary condition and thus shows an improvement.

VI -24

Summarr and Possible Import for Teaching.

In summary we appear to have been

dents to avoid the basic fallacies

and practical contraposition

doing, however, we might have

able to teach upper secondary stu-

and to some extent to recognize basic

as valid moves in conditional reasoning. In so

interferred with their grasp of the basic

valid moves from Tonly_ife

sentences.

We seem to have had little effect

on their grasp of the transitivity of the if-than relationship.

level, as well as at earlier

levels, the teaching might have had some

positive effect on knowledge of

affirmation of the antecedent,

statistically significant.

At this

the most basic conditional reasoning move,

but, if so, the effect was small and not

Since affirmation of the

antecedent is in a

way
a transitivity principle (the affirmation carries through from the antecedent

to the consequent), these

results fit together.

The most striking fact about the

great effectiveness

results of this teaching was its

on total score at the upper secondary level,

with its either negative effect

fUture time, the location and

as contrasted

or lack of effect at lower levels. At some

sharpness of the implied dividing line should

be investigated.
The sufficient condition

percentages suggest rather strongly that the

basic principles of conditional

the upper secondary level

but that there is not much

reasoning can be pretty well mastered by

-- at least for students like those in this study,

point in trying to teach conditional logic in

elementary and lover secondary.*

Furthermore these results suggest that

the
things that can be taught

are the fallacies, contraposition to some extent,

and perhaps the validity of

partly teachable, but also

affirmation of the antecedent, which might

seems to develop on its own without deliberate

teaching.
* Incidentally this sentence

and others that refer to upper secondary, etc., are

kind of line
deliberately somewhat vague because of the difficulty of placing agy
on the basis of
avoid placing the 9th
these results alone. In particular ve vent to

graders

on either side of a line,because of doubts about the motivation of the 9th

grade LDZ's.

VI-25

That transitivity and

the valid moves with 'only if' seemed to develop

their own and that deliberate

on

teaching did not help at the levels with

which we worked is
percentages
meeting the sufficient
levels

condition during the teaching period at all grade

, together with the percentages that have mastered them in the 11th

grade on the pre-test: 81,

8, and 9. These figures

81, 88, and 96 respectively on Item Groups 5, 10,

can be found in Table A-8

in the Appendix.*

Learning Words vs. Learning Logic.

An intriguing questionais
registered in the four

of logic, or whether

fallacy item groups represented greater understanding

it represented a grasp of the meaning of the word 'if'

as it is properly used.
is commonly taken

that of whether the great improvement

It might be argued that the use of the word 'if'

to imply what is properly meant by 'if and only if', or

in other words that

'if' is taken to introduce a necessary as well as

sufficient condition.
knowledge of logic;

we only changed their vocabulary.

* A possible objection
at the lover grade levels is that the

The suggestion then is that we did not improve their

to this claim about the teachability of transitivity

5th grade staff teacher estimated that

combinations of principles.

Item Group 10 calls for either

the double application
for the
of the affirmation of the antecedent principle, or
combination
antecedent
of the transitivity principle and the affirmation of the
principle.
Since our instructor reported spending no time on

combinations, how
teachable at the can we suggest that this particular combination is not
lower grade levels, one might ask.

In reply we would
of the
say that an effort was made to teach the affirmation
antecedent
principle and the transitivity principle, but it had no
noticeable effect. Hence we conclude
or with
that the combination of one with another
itself would not
have been effectively taught at the lover grade levels.

VI-26
This is an extremely puzzling issue. As Benjamin Lee Whorf has
pointed out, our language, our conceptual structure, and our power to

think are intimately related. But it does make sense to suggest that the
distinction that we represent by the words 'if' and 'if and only if' was already
known to our subjects (even though they do not use these words to represent it)
and that they used 'if' to mean what we mean by 'if and only

Thus even if there is an intimate relationship, there is also a

if'.

meaningful question.
If this verbal interpretation of the learning that went on is accepted,

then the learning has less significance than otherwise supposed. Under
this interpretation we were not teaching the students to think; we were

teaching them a common vocabulary.

Now this is important too -- in order

to facilitate communication in the making of crucial distinctions -- but

it is

not as important as teaching them to think more clearly.

This verbal interpretation is consistent with what happened with

transitivity, which seems to be not simply a verbal matter, but also a
conceptual one.

The hypothesis that we were teaching them the meanings of

words allows that there would not be much learning of transitivity. But
the learning of contraposition and what there was of affirming the antecedent
is contrary to the verbal hypothesis, because these things seem to be clearly

conceptual, and not simply verbal.

The counter-effect of this latter

evidence
might be explained away be the suggestion that for these things they learned

a few rules by rote and did not understand them.

As one can readily

see, the question bears fUrther investigation.

Class Reasoning.
The situation in class reasoning is rather different from that in

conditional reasoning.

The dividing line, if there is one, comes in the

upper elementary levels, somewhere around age 12; but that there is a line

is not clear.

The teaching of class reasoning did not have much effect on

the 4th graders, but generally had a positive effect on those from 6th grade
up to 12th grode. This effect is not nearly as striking as the effect
of teaching conditional reasoning to the 11th graders.

Total Scores.

The change in mean total scores for the LDT's over the interval between
pre-test and post-test is positive in grades 6, 8, 10, and 12, but not in
grade 4.

This can be seen in Table 111-3 in Chapter III.

The improvements

registered are 9.5, 3.8, 8.3, and 6.8 respectively. These are considerably
smaller than the improvement of 17.0 points registered by the 11th graders
on conditional reasoning.
In a post-test comparison with the

holding IQ and pre-test

constant, the LET's in class reasoning were significantly superior in grades

6 and 12, but not in the other grades, given the degree of refinement of our

experimental procedures. This is shown in Table V1-2.

Component Scores.

The situation is generally the same when scores are broken down into
their components. As indicated on Table VI-2, there is some statistically
significAnt superiority of the LET's at grades 6 and 12, and also at grade

This happens most frequently in the symbolic component.

Since in the

teaching frequent use vas made of such phraseology as 'All A's are B'e',*
this difference in the symbolic component is understandable. It is perhaps
noteworthy that only in the 6th grade were the LET's statistically significantly superior on two of the components. In no grade was this the case for

three components.

In grades 8 and 12, it was for one component only, the

symbolic component.

* The letters at the end of the alphabet (e.g., 'X' and 'Y') were used in
the test, not the ones at the beginning, which were used in teaching.

111 -28

Item Group Scores.

In the class reasoning item groups there is nothing that compares with
the remarkable improvement made by the 11th graders on the conditional
reasoning fallacy item groups.

The moderate general overall improvement

from grade 6 onward, when instruction is given,*can be seen in Tables A -9

and A-10 in the Appendix. As can be seen in Table VI-20 for one item
group, #12, there was a statistically significant superiority, IQ and pretest item -group score held constant, at three grade levels on the post-test.
This item group embodies the most complicated logical structure on the test.
Perhaps the instruction was fairly successful in providing an orderly way
of dealing with complex arguments.
EXceptions to this general overall improvement during the period of
teaching are Item Groups 1, 10, and 11. Let us discuss them.
Item Group 1 is a special kind of item group in that it is so basic:
it tests for the meaning of 'all' and negation. It calls upon someone to
judge that a statement of the form, 'At least some A's are not B's', is
inconcistent with the corresponding statement of the form, 'All A's are B's.'
Tb some extent this item group is a test of whether the subject understands
what is requested of him on the test, because presumdbly people at these
levels are fairly well acquainted with the meaning of 'all' and of simple
negation.

So we are not surprised by the lack of improvement under instruc-

tion on this item group.

The lack of improvement under instruction for Item Groups 10 and 11 is
somewhat puzzling.

Item Group 10 calls for the combination of the transitivity

and contraposition principles.

Simple transitivity is tested for in Item

Group 2 and basic contraposition is tested for an Item Group 9. There was
some improvement in each of these, and according to hypothesis abovefthere

VI-29

was, improvement in ability to deal with combinations. So there is

something of a puzzle.

However, the pre- and post-test differences on

Item Groups 2 and 9, the basic

transitivity and contraposition item groups,

are not large enough to make the problem serious. When the study is

replicated with more subjects,

one can see if the problem arises again.

It is similar with the lack

of improvement in Item Group 11,

calls for the double application

which

of the transitivity principle. The

single

application of the transitivity

principle is tested for by Item Group 2, and

there is general improvement

after teaching on that item group.

But

again

the improvements are fairly small,

so it is not fully clear that there is

puzzle.

Summary.

There is nothing striking about

There appears to be moderate and

age 12 onward.

the effects of teaching class reasoning.

fairly general improvement in groups from

Two things that stand out slightly are the improvement in

handling complexity represented

by Item Group 12 and improvement on the

symbolic component.

Possiblelanations of the Differences Between Conditional and Class

Reasoning.
Inevitably one wants to try to

between the score patterns

explain the differences that we found

on the conditional reasoning test and those on

the class reasoning test.

Why the striking improvement

on the conditional reasoning test and

grade 6 upward on the class

between the improvements

speculate at this time.

in the 11th grade

the only moderate improvement from

reasoning test? Why was there such a contrast

on the fallacy item groups? We can only

A possible explanation of

the general situation is that

since class

VI-30

reasoning is easier, the students

possibility of improvement.

This explanation has the weakness of not

explaining Why in cases where there

improvement, there

on the class reasoning tests had less

does seem to have been room for con-

siderable

was still not an improvement comparable to

that registered in conditional

Tables A-9 and A-10 in the
that for item Group 3, which

reasoning.

Appendix.

Many such cases can be found in

Fbr example in Table A-9 one can see

is a conversion fallacy group, although the

mean difficulty indices on the pre-test were 340 45, 51, and 71 in grades 4,

and 10

6, 8,

respectively, the improvements were only 4, 11, 21, and 13.

A second possible explanation

fallacy item groups,

is that in the conditional-reasoning

a great share of the learning was of the sufficient-

condition meaning of the word

'if', as opposed to the necessary-and-

sufficient-condition interpretation. Since there was no corresponding

learning for the class

verbal

reasoning test, the improvement was less, the

explanation holds, although the

amount of actual improvement in knowledge

of logic was about the same.

A difficulty with this

verbal learning did not

that if the learning is of

explanation is that it does not tell why this

occur before the 11th grade.

It might be urged

a new meaning for a word, one would expect 7th

and even 5th graders to be able

to do it.

This challenge to the

explanation

of course has no evidence to

could be grasped at such

offer that a new meaning for this word, 'if',

an early level. So the question is unresolved, and

more research should be done.

Can We State What Others Are Beady to Master?

The attempt to

answer this question is based on the development of

multiple regression equations

into which one can introduce values of cer-

VI-31

tain variables and produce a predicted

score.

Two sets of these

equations

were prepared: one made use only of information which is generally available

in school records; the other in addition made use of pre-test scores.

In the preparation of these equations,

logic test were grouped together.

combined, as were grades 4, 6, 8, 10,
score, the three component

that made use of pre-test

and 12. Equations were prepared for

total

since there was a set of values prepared

scores as well as one Which did not, there are 64

equations altogether ((1 +3+12)

computed for each equation.

The weights given in those

X 2 X 2). A cross-validation estimate was

The non-variable values that go into these

equations appear in the Appendix,

Tables A-111 A-12, A-13, and A-14.

tables are the factors by which one must

are introduced into the equation. These products

together with the given constant

Here is an example of

Hence grades 5, 7, 9, and 11 were

scores, and the twelve item-group scores.

Since there are two tests, and

multiply the scores that

all LDT's who took a given

are summed to give the predicted score.

en attempt to predict a score on Item Group 4 on the

conditional reasoning test:

a.

An I;
,

e of the Use of a Multi le lie

scion

=tion for

on.

Assume the following

values of the variables,

Which values are the

mean values for the 11th grade LET's:

Grade:

11

Chronological Age: 199.8 months

IQ: 116.0
Bocio-Economic Index: 3.5
Sex: 1.5 (male=1; femsle2)
Total Score on pre-test:

Pre -test score on this

63.7

item group (44 here):4.26

Combining these figures with the weights and constant given in

Table A-111 one has the following equation:

Predicted score = (-.001) (11) + (.0074) (199.8) + (.0064)

(116.o) + (-.127) (3.5)

(.314) (4.26) -

+ 1.34

1.26

This figure, 4.91,

1.26 =

+ (.423) (1.5) + (.0384) (63.7) +

+ 1.48 + .74 -

.45 + .63 + 2.44

= 4.91, predicted score.

can be rounded or not, depending on what one is

going to do with it.

Making Use of the Generated Predictions.

Naturally the utility of such

predictions depends on the multiple

correlation coefficients and the

cross validation.

These figures,

together

with the standard errors

are given in Table VI-5.

interest in predicting mastery

multiple correlations and

Given our primary

of the principles of logic, the corresponding

estimated cross validations are not high enough to

warrant predictions of individual

item group scores.

But predictions

for

groups seem possible.

Inserting the mean scores for a group enables one

to generate a predicted

mean score for the group.

But we would like to

diction of the size of the

go further.

percentage of the group who, after comparable

instruction, will have mastered

definitions set forth in Chapter
the percentage who will meet

at least five correct out

We would like to make a rough pre-

the principle.

Given the operational

IV, the interest is in roughly predicting

the sufficient condition for mastery -- getting

of the six items in the item group that corres-

ponds to the principle.

predicted mean, particularly

One cannot assume a normal distribution around the

when it is 4 or better,

as one can see frma

working out a few comparisons between

the mean difficulty indices and these

percentages as they actually occurred on the testing that was done.

VI-33
TABLE VI-5.

Multiple Correlations, Cross Validation Estimates,

Standard Errors for Prediction Equations.

Conditional
94
Without
using
Pre-Test
Pre-Test
Rm Rxv
SE
Rm Rxv
SE

N=

Total
Score

and

Class
123
Using
Pre-Test
Rm Rxv
SE

Without
Pre-Test
Rm Rxv
SE

.81

.77 12.07

.73

.69

13.85

.87

.86

.81

.78

CF

.82

.78

5.17

.71

.66

6.31

.84

.82

.77

.74

5.51

SY

.72

.65

1,74

.67

.62

1.83

.78

.75

.76

.73

1.77

SU

.72

.67

1.69

.69

.64

1.74

.84

.82

.77

.74

1.91

.59

.47

1.05

.52

.41

1.10

.75

.71

.93

.59

.53

1.12

.77

.71

1.45

.65

.58

1.69

.71

.66

.84

.67

.62

.88

.59

.48

1.60

.59

.50

1.59

.76

.72

.66

.62

1.38

.63

.53

1.42

.51

.40

1.54

.6o

.53

.54

.47

1.69

.61

.50

1.28

.46

.32

1.42

.63

.56

.89

.60

.54

.91

.56

.42

1.55

.40

.22

1.69

.73

.68

.94

.64

.59

1.05

.66

.58

1.81

.6o

.53

1.91

.65

.59

.46

.26

1.48

.40

.21

1.51

.76

.73

.54

.46

1.77

.36

1.55

.00

.31

1.56

.65

.60

.0o

.56

.47

1.49

.53

.46

.53

.45

1.45

.68

.64

1.21

1.04

.56

.49

1.11

11.71

Component

Item
Group

lo

.67

.59

1.44

.60

.53

1.53

.58

.5o

.65

.56

1.46

.49

.37

1.66

.74

.70

.56

.43

1.76

.52

.42

1.79

.64

.57

Note:

Cross validation indlces estimated by means of the Lord-Nicholson

formula. (Brogden, 1954jpp 377-400)

ibis the multiple correlation

Rxv is the cross validation estimate

SE is the

standard error.

1.39

VI-34
So we plotted each mean difficulty index against each percentage
meeting the sufficient condition on the pre- and post-tests for each of the
two types of reasoning studied. These plotted points are reproduced in the
Appendix as Graphs A-1

and A-3.

A line of best fit was derived from these

graphs and appears in Graph VI-1 in this chapter. A similar process was
carried out for the percentages failing to meet the necessary conditions.
See Graphs A-2 and A-4 in the Appendix. The line of best fit thus derived
also appears on Graph VI-1.

Now we are in a position to complete the treatment of the example.

Taking the predicted score of 4.91 we divide by 6 in order to secure the
corresponding mean difficulty index. In this case it comes out 82%.
Tracing this value up through two intersections, we obtain predicted
percentages of 13% and 72%. The former is the percentage predicted, to
fail to meet the necessary condition and the latter is the percentage
predicted to meet the sufficient condition. In other words we predict that
somewhere around 72% of the students in our hypothetical case will have
mastered Conditional Principle #4 and that somewhere around 13% will

definitely not have mastered it.

Incidentally the actual percentages

for

the 11th grade LDT's were 15% and 77% on the post-test.
Making use of the previous analysis of readiness for mastery and
accepting the limitation imposed by our being able to provide only a given
amount of a given cost of instruction, we can now exemplify the type of

statement that we would like to make about the readiness

of the members of

a group to master a given principle of logic. Let us assume that these

statements are to be made about the 11th grade Lt's before they received

any deliberate instruction.

percentages

What we want to do is to estimate the

VI -35
GRAPH V1-1.

Empirically Derived Relationships Between Mean Difficulty

Indices and the Percentages Meeting the Sufficient Condition
and Failing to Meet the Necessary Condition.

100

90

NM.

w
it

80

0
0

ti
0

P4

70

aro

60

MM.

0
v4

4---SUFFICIENT

NECESSARY

0
C.)

4,

50

40
CO

30

112

20

10

10

20

50

60

Mean Difficulty Indices

70

80

90

100

VI-36
who are ready to master

a given principle (say Principle 4) and who are

not ready.
Referring to Table A-8 in the
at least 50% of the

Appendix we see that the pre-test shows

that

students have already mastered Principle 4 and

at least 23% have not mastered

met the sufficient condition

it. This follows from the fact that 50%

and 23% failed to meet the necessary condition.

About the remaining 27% no judgment is made.

Nov we have predicted that
72% and 13%, for an increase

after instruction the percentages will be

of 22 percentage points and a reduction of 10

percentage points respectively.

So we can say the following:

Since 50% have already

mastered the principle, the question of

their readiness does not arise.

At least 22.%

At least 13%

are ready to master the principle.

are not ready to master the principle.

We do not want to be
committed about the readiness of
other 15%.
(100 -50 -22 -13 = 15)
the

Several qualifications

need to be made, but for illustrative purposes

we made the specific unqualified

statements given above.

Obviously

with

a group of the size of the 11th

imply such precision.

More rounding should occur.

'probably' and 'approximately'

This procedure for
a given principle

And such words as

should be introduced.

estimating the per cent of students who are ready

to master

is novel, and we do not knows way to estimate

a confidence interval.
that this procedure
we have.

grade 18DT's (N=26) one would not vent to

The distribution is not normal. We can at this time

gives the best estimate that we can make on the

The development of procedures for estimating

only say

basis of the data

a confidence interval is

a task to which study needs to be given.

VI -.37

It should be noted that the procedure is based upon students vho

have taken the pre -test and thus have had the benefit of whatever learning

accrues therefrom.

If the equations without pre-test values are used,

it should be kept in mind that part of the instruction given to those

students was in the form of the administration of the pre-test.

Summary and Overview.

We have developed a method for making statements about the percentages

of a group who are ready to master

a principle and who are not ready.

When

pre-test scores are not available, the sufficient condition statements must
be about percentages who either have mastered the principle or are ready to

do so.

The necessary condition statements would not be altered by the lack

of these scores.

One would still talk of the percentage who are not ready.

These equations can also be used to predict group mean total component

scores. This is a more traditional

use of such equations.

The trouble

is

that there is as yet little operational interpretation of such scores.

FURTHER RESEARCH

The entire experiment should be replicated, but with more classes at

each grade level to minimize the effect

of class-wide factors.

The

intensive

treatment used in this study leaves the results open to influence by

idiosyncracies in the treatment.

Furthermore the equations should be checked with

students in other environments and with those in grades between

used in this study.

those Which were

The dividing line, if it exists,

beyond which conditional reasoning can be effectively taught should be

located more definitely.
Research of this sort should also be extended to lover ages and an

attempt should be made to increase the amount of time devoted to teaching

logic in order to

see whether that would make much difference at the

various age levels.

The length of teaching time could also be used as a

variable in the multiple regression equations.

The tests should be

given on a pre- and/or post-test basis at various

levels, including college, in order

the meaning or

to develop a greater understanding of

predicted total and component scores.

Another set of

multiple regression equations might be developed,

making use of personality

variables.

These have been neglected in this

study.

The possible anomalies between the

groups 10 and 11 and the

item groups representing their parts should be

investigated end resolved

improvement in the

class reasoning combination item

if possible.

One wonders thy when there is

parts (and in ability to combine parts), there is not

an improvement in the thole.

Last and perhaps

most important, the possibility that verbal as opposed

to conceptual learning accounts

improvement among 11th

require sophisticated

for the marked conditional-reasoning

graders must be investigated.

indirect investigation.

CHAPTER SUMMARY.

The first part of this chapter

concept, readiness to
dispositionality of the

statements that make

its explosive -type

This problem will

was devoted to an analysis of the

master a principle. It was noted that the double

concept makes it difficult to justify empirically

use of the concept.

Another source of difficulty is

dispositionality.

The result of the

analysis goes roughly as follows:

is reedy to master Principle X is to

suitable conditions,

To say that Y

say that Y has the disposition, given

to develop the disposition to show, given suitable

conditions, the sort of

behavior that a person who has mastered Principle X

VI-39

would show.

For simplicity's sake there is a circularity in the use of

the term 'mastery' to define

'readiness for mastery', but this is resolved

for present purposes by appeal

given in Chapter IV.

is unavoidable.

to the operational definition of

'mastery'

The vagueness of the phrase 'suitable conditions'

The concept is vague in this respect.

After noting the dearth of

to learn logic, we described

literature dealing with the topic,

capacity

our experimental procedures, and said what we

could about the nature of the teaching that was done.

Discussion of the results
which was the more speculative,

readiness statements.

was broken up into two parts, the latter of

because it attempts to make present-tense

The first part was concerned

actuall did learn while under

instruction.

discussion of their prior readiness

with what our subjects

Thus it was essentially a

for mastery and in effect made

ast7

tense readiness statements.

There was a marked contrast between the learning of conditional

reasoning that was effected and the

learning of class reasoning.

There

was

practically no learning of conditional

but in grade 11 there

was a vast improvement. This improvement was con-

centrated mainly among the fallacy

presenting contraposition
the other hand there

item groups, but also to some extent in

end affirming the antecedent. For class

those

reasoning on

were small amounts of improvement fres

grade 6 onward, the largest

improvement under instruction being registered

with the most complex items.

the

reasoning in grades 5, 7, and 9;

This latter improvement perhaps indicates

efficacy of the instruction in

handling and ordering complexities. /s.

provement in handling symbolic items was also noted.

One possible explanation of the marked difference between conditional

and class reasoning is that the major learning that accounts for the Vast
improvement in conditional reasoning was the learning of the sufficientcondition meaning of the word 'if', which has no analogue in class reasoning.
One difficulty with this explanation is that it leaves one wondering why
the meaning of this word was not effectively taught in grades 5, 7, and 9. On
the basis of these results it seems that, given comparable students,
class reasoning instruction might profitably be commenced at about age 12,

and that conditional reasoning

possibly sooner.

instruction might be commenced by age 16,

It was roughly at these ages that we found readiness for

greater percentages of mastery.

The attempts to make

mastery of particular
multiple regression

present-tense statements about readiness for

principles of logic were based upon the development

equations which made use of the variables, grade,

chronological age, IQ,

socio-economic-status, and sex; and in addition the

optional use of pre-test scores.

enough for use in

of

The predicted scores are not reliable

individual cases, but a scheme was developed for making

predictions of percentages of groups.

These predictions, assuming the

previous

analysis of 'readiness
percentages who

to master Principle X', become statements about

are ready to master the principle and who are not ready to

nester the principle.

These statements about percentages assume comparable

teaching treatment and thus the

The major weakness

readiness statements generated are overprecise.

of this study lies in the small number of subjects

with 'which we worked (about 20-30

levels for each

type of reasoning).

per grade level, with only 4 or 5 grade

This smallness

the requirement that the treatment

carried out by people

be intensive:

was made necessary by

that the instruction be

who knew logic, who knew how to teach at the level

assigned and who had adequate

time for class preparation and evaluation.

Chapter VII - The Past and /Nature of the Project

Although no date can be given to mark the beginning of the Cornell

Critical Thinking Project, since it just grew Topsy-like, several events
earlyein its existence can be dated. Two general critical thinking tests
--ere prepared in 1961. A tentative theoretical grounding of the project

was pelished in 1962 and has been under constant revision since then.
set of short reviews of works related to critical thinking was prepared
and made available in 1962.

These reviews were supported in part by a grant

of the Ford Foundation to the Cornell Junior High School Project. In

addition funds from the same source were used to pay for a collection of
works related to critical thinking and a set of specimens of every critical
thinking test that we could find. These tvo collections proved very helpful
as research has progressed.
The Readiness Study: The Past.
In academic year 1961-1962 the Cooperative Research Branch of the
U.S. Office of Education gave a grant to Cornell to study critical thinking
readiness.

That grant, vhich ran from Nay, 1962, to SepteMber, 1964,

the research reported on here.

suppetted

It was felt that a readiness study mould

involve some of the basic work necessary before applied curriculum and
method studies could be performed.

This basic work included clarification

of aspects of critical thinking, development of tests for aspects of

critical thinking, and rough determination of the knowledge of, and capacity
for learning aspects of critical thinking found in students of various levels.
The initial intent under this grant was to pick two aspects of
critical thinking, probably deduction and assumption-finding, ang perform
the above-indicated basic work for each at all the levels of elementary

and secondary school.

We started with deduction, feeling that it, of

all aspects of critical thinking, was the one most fully developed, and

VII-2

would be the easiest to handle.

both a pilot study for the

We hoped to make the study of deduction

other aspects of critical thinking and a

study of deduction in its own right.

As soon as we

came to a detailed consideration of the nature and

principles of deduction,

we discovered not only that it as a field has not

and categorized, but that the parts that have

been worked on by logicians

We had to select
types.

are not by any means clear and uncontroversial.

among possible types of deduction and had to select within

The content in the field

of logic not being as well worked out as

a great deal of time to this job.

Although our original

intent had been to teach the basic principles

of logic
of deduction at each grade
level, we discovered as a result of our analysis
that this would

be too much to teach in the limited amount of time

we were able to get the school

system to commit to us, so we set about to

The plan became

only one particular type

pilot year, we picked

of logic at each grade level.

for study sentence reasoning, class reasoning, and

ordinal reasoning.

These are described in Chapter II.

Because of item and test

reasoning to the elementary

school experience, most

fairly well mastered the

tryout experience we assigned ordinal

grades only, since by the middle of the secondary

of the students whom we interviewed had already

basic principles of ordinal reasoning under the

present natural-cultural
basic principles of

one of working on

system.

There seemed not much point to ask what

ordinal reasoning secondary students are ready to

master that they have

already generally mastered these principles.

VII -3

During the first and pilot year of the

study ordinal reasoning was

studied in grades 1, 3, and 6; sentence reasoning in grades 2, 5, 7, 9,

and 11; and class reasoning in grades 1, 4, 6, 8, 10, and 12.

After

extensive
item and test tryouts, one test for each

of the three types of reasoning

was developed. It was used as a pre-test and a post-test for

three typical

classes at each grade level. Shortened versions and special techniques

were used at the lover grade levels.
One of the three classes at each grade

level was selected for intensive

teaching of the type of reasoning assigned to

teaching was done by one of our staff members,

specially trained in logic.

for 10 instructional days

the grade level.

This

who were experienced teachers

In each case it ran for one period per day

(two weeks).

As a result of this pilot experience, we

learned several important

lessons:

That the amount of time devoted to the instruction was

insufficient. We concluded this because we felt that we had

in

general

learning logic.

not come anywhere near out students' limits for

probably valid.
This was a subjective impression, but quite

student's time
That in future readiness studies, much more of a
must be
Because ve were both
devoted to the study of logic.
operating on a small
what
scale and insistent upon control over
was being taught, we had to
staff members into
introduce our own
resulted
the schools. This inevitablytime
by
we
in strong limitations
the schools on the amount of a student's
A
could have.
possible way around this problem, a way not open to us this
last

is to provide training in logic

year because of budget limitations,

diming a
the teaching of logic for a large number of teachers

summer, and then make use of their

entire school year.

and

regular classes during the

That some test revision was still necessary.

That the first ,year's study should be treated only as a

study for the reasons given above and that the second year should

pilot

deduction,

consist of an improved study of readiness for learning

first year.
more of a student's time committed than during the

with

VII-4

5.

That this improved study would require greater

concentration of our resources and a consequent reduction
in the scope of the effort.

6.

That sentence reasoning would have to be cut down to

conditional reasoning, omitting disjunction, alternation, and
conjunction reasoning. This would be in order to reduce the
number of principles to be taught,--that is, to make the
teaching task more manageable. We judged the body of basic
content in conditional reasoning to be roughly equal to that of
class reasoning.

To achieve a further reduction we chose to eliminate the lower primary

The results of the first year

indicated that, at least as we were going about teaching, not much learning
of class and conditional reasoning was going on in those years. Although
ordinal reasoning was learned in those years, ordinal reasoning is not as

Since something
had to go, we dropped the years in which the most central parts of logic,

given our limited amount of teaching, were not getting across to the students.

The concentration of resources was achieved by increasing the amount

of teaching time to 15 instructional days on logic, generally acccopenied
by an additional 5 days in which our staff member devoted himself to the
advancement of the subject matter which was being replaced by the logic
instruction.

The second year's study is that described in the body of this report.
The Tests.

The original plan called for one deductive logic test, consisting of
six items per grade level of about 50% difficulty for the given grade level

(with some extras at both ends).

The result was to have been a 90-item

test modeled after Burt's Graded Reasoning Test (1919).

VII -5

As indicated earlier one of the results of the analysis of logic

was the decision to split deductive logic into parts in order to make the

teaching more manageable.

longer suffice.

Hence one test of logic in general would no

But there was another reason for changing the original

The analysis of logic showed not only that there exists a

small set of basic principles for each of the types of logic, but that a

general deductive test would yield

We felt we could refine the

a score devoid of intrinsic meaning.

measure sufficiently to have it be in terms of

specific principles of logic.

Scores on these principles together with

the operational definitions of mastery

of the principles, which are given

in Chapter IV, yield a score which hopefully has meaning which is not as

on a jumbled array of assorted deduction items.

Hence instead of one deduction test which

of norms, we developed a test for each of the types of logic studied

which would yield scores

on the principles of logic, scores which are

interpretable in ordinary language.

The original plan called for 90 items. There are 144 items in the

class and conditional reasoning tests (72

additional
66 items are in the ordinal

this study.

items apiece).

An

reasoning test, which is no reported on in

It is being held in reserve for a future phase of the study.

In the construction of the three tests,

prepared -- and supplemented

were conducted.

an original pool of 740 items was

as test development, interviews and tryouts

Approximately 30 students were individually interviewed

in the preparation of the tests, and about 500 students were used in various

stages of tryouts.

VII -6

The Readiness Study:

The Future

This report covers only the first of a number of phases in the study

of critical thinking readiness in grades

1-12.

In the tentative

designations
of phases to follow, we have broken the process of doing a readiness study

of an aspect of critical thinking

aspect; the development of

into three parts: the analysis of the

a test (or tests) for that aspect; and the

combined process of testing, teaching,

All three parts would in

and developing readiness equations.

general take about three years to carry out for

each aspect of critical thinking. Starting one aspect per year would
result in an overall time commitment of about 10 more years.

VII -7
Table VII-1.

Academic
Year
1967-68

1968-69

1969-70

A Tentative Time Table for FUture Phases of the Critical

Thinking Readiness Study.
Testing,
Aspect
Analysis
Assumption-Finding

Test
Development
Ordinal Reasoning

Generalizing

(revision)
Class and Conditional
Reasoning
(Grades 1-3, adaptation)
Alternation and Dis-

Alternation and

Reliability of

junction Reasoning
Assumption-Finding
Generalizing

Disjunction Reasoning:
Grades 1-12
Assumption-Finding:

Observation Statements

1970..T1

1971-72

1972-73

1973-74

Reliability of
Authorities
Hypothesis Warrant

Detecting Ambiguity

Overvagueness and

Theory Warrant

Grades 1-6
Class and Conditional
Reasoning:
Grades 1-3

Grades 1-12

of
Observation Statements

Generalizing:
Grades 1-12

Reliability of

Reliability of Obser-

Authorities
Hypothesis Warrant

vation Statements:
Grades 1-12
Reliability of Autho-

Detecting Ambiguity

rities:Grades 1-12
Hypothesis Warrant:

Overvagueness and

Grades 1-12
Detecting Ambiguity:

Overspecificity
1974-75

Teaching,and
Readiness
Equations
Ordinal Reasoning:

Overspecificity

Grades 1-12

1975-76

1976-77

Instead of aspect
analysis and new test
development, this portion
)will be devoted to
repeating some
investigations, and
writing a total
report.

Overvagueness and
Overspecificity:
Grades 1-12

Theory
Warran
t:
Perhaps
Grades
1316

In this chapter we shall try to indicate the nature of the problems

that we set for ourselves, the interpretation of our key concepts, the

procedures used to pursue our objectives, and

our results.

Many

qualifications

will be omitted in order that this overview be in condensed form. The

reader is referred to the specific chapters in the body of this report for

fuller treatment of each topic.

This particular chapter is organized in

the same manner as the report as a whole and

esaarally use chapter headings

sub- headings.

Introduction:

The Statement of the Problems.

This study is concerned with two general empirical questions, which

provide the focus of the study, and
Which, although auxiliary with

several conceptual ones, the answers to

respect to the empirical questions, are important

for the clarity and understanding they provide in dealing with the

empirical questions.

The empirical questions are 1) that of readiness for

mastery of logic, and 2) that of

the natural-cultural development of mastery

of logic.

The readiness question is

concerned with the level (age, grade, etc.)

at which and the extent to which

students are ready to master logic.

natural-cultural development question

and the extent to which logic

The

is concerned with the levels at which

is mastered by students who have not been

exposed to deliberate instruction

attributable to both natural and

in logic.

Presumably this development is

cultural factors; hence the title:

"Natural-

Cultural Development of Logic".

These two questions

are closely related.

If we find out that a certain

principle of logic has already

been mastered by someone as a result of

VIII -2

natural-cultural processes, then the question of readiness for mastery of

the principle does not arise.

FUrthermore the process of estimating what a

student is ready to master is facilitated by knowing the degree of mastery

at a given moment.

In investigating the readiness

question, which was our

primary interest, we unavoidably gathred information on the natural-cultural

development question, which is the subject of considerable other

especially that of Piaget.

research,

Both questions are discussed in this

report.
In order to discuss these questions some conceptual problems had to be

dealt with first.

First of all we had to delimit the

area of logic and

select a concept of a valid argument that is in agreement with intuition and is

relatively non-controversial. Next, we had to develop operational

definitions of 'mastery of

a principle of logico; but in order to do this

ve also had to do some York on the concept of an operational definition.

Lastly the concept of readiness required analysis.

It might well be that the actual

greater in the area of conceptual
findings.

contribution made by this study is

clarification than in the actual empirical

The empirical findings are limited by the geographical and

num.erical limitations of

the subjects with whom ve worked.

But the

conceptual

suggestions that were developed

can be applied to other studies, of which

mare are needed.

In order that we can

assume the understanding of the conceptual points in

the presentation of the empirical

findings, the conceptual points will in

this chapter precede the empirical

findings which depend upon them.

The Mature of the Subject- Metter: Logic.

The subject matter, the mastery of rich is the subject of this study,

VIII -3

is logic, which itself is

thinking.

an important part of our basic interest, critical

It is that part of

conclusion follow necessarily

of it.

The centrality of this part

critical thinking Which deals with whether a

from the premises that are offered in support

of critical thinking

has been argued

for elsewhere (Finis, 10)4

Logic is not concerned with

but simply with Whether

whether the premises or conclusion are true,

the former necessarily implies the latter.

Ti o Types of Logic.

There are many different

relationship of necessary

types of statements between Which this

implication can hold.

field of logic is categcrizsd

Of the many types of

Roughly speaking, the

according to the types of statements treated.

logic which exist (no one has successfully defined

them all), we have selected

two of the most common, conditional logic and

class logic.
Conditional logic deals with

if', and/or their

words just mentioned

statements containing the words, 'if', 'only

synonyms. It is called "conditional logic" because the

are frequently used to introduce conditions upon which

the truth of the rest

of a statement containing them depends.

The part

of

a statement introduced by the

rest (except for the word
"consequent".

word 'if' is called the

"antecedent".

The

'then', when used to introduce the rest) is called

the

A simple argument in conditional logic proceeds by

presenting a conditional

statement, then affirming or denying either the

antecedent or the consequent,

affirmation or denial of the

conclusion follow

and drawing as a conclusion either the

consequent or antecedent respectively. If the

necessarily, then the argument is called valid. If the

conclusion does not necessarily

follow, then the argument is called invalid.

VIII-1t

Here is an example of

a valid argument in conditional logic, vhich has

been adapted from the

test used in this study:

Premises:

If the car in the

parking lot is Mr. Smith's, then it is blue.

The car in the parking

Conclusion:
The car in the parking

Class logic, vhich

deals with statements

is also concerned with the validity of arguments,

'which have a different

relationship to each other.

consequent appear and

In conditional logic,

independently of each

or groups.

or the other.

Statements in class logic

interchangeable unite not sentences that can stand

other, but terms and phrases which refer to individuals

For example, the

W. Smith's cars",

the antecedent and

reappear in the argument essentially unchanged, except

for the possible negation of one

have as their basic

structure and a different

statement, "All the

is a class logic

statement are the terms,

vhich refer to groups.

"the

cars in the garage

are

statement because the basic units in the

cars in the garage" and "Mr. Smith's cars",

The statement gives a relationship between these

groups.
Here, for example,

item from one of the

is a valid argument, vhich is an adaptation of an

tests that we developed:

Premises:

All the

cars in the garage are Mr. Smith's cars.

All W. Smith's

cars are Fords.

Conclusion:

All the cars in

the garage are Fords.

VIII-5

Principles of Logic.
selected mhat seemed to be

Pbr purposes of instruction and testing we

the basic elementary principles in each of the

above two types of logic.

They are listed below and can be found

exemplified

in Chapter II:

Conditional Logic:

Given an if-then sentence,

the affirmation of the if-part implies the

affirmation of the then-part.

Given an if-then sentence,

itself (as
the denial of the if-part does not by
a result
of its being an if-part) imply the denial of the then-part.

Given an if-then
not
sentence, the affirmation of the then-part does
by itself (as
a result of its being a then-part) imply the affirmation
of the if-part.

Given an if-then
of the if -part.

1
2

2.

sentence, the denial of the then-part implies the denial

The if-then relationship is transitive.

An if-then

sentence implies its contrapositive.

The if-then relation is non-symmetric.

Given an only-if

sentence, the denial of the only-if part implies the

denial of the major part.

Given an only-if

sentence, the affirmation of the major part implies

the affirmation of the only-if pert.

The denial or affirmation of

implies respectively

one

part of

an if-and-only-if

statement

the denial or affirmation of the other part.

Given an only-if
not
sentence, the affirmation of the only-if part does
by itself (as
a result of its being an only-if part) imply the
affirmation of the major part.
Given an only-if
itself
sentence, the denial of the major part does not by
(as a result
of its being the major part) imply the denial of the only-if
part.

Class Logic:

Whatever is a member
vice versa.

of a class is not a non-meOber of that class and

vni-6

Whatever is a member of
a class is also a member of a class in *WI
the first is included.
(This implies that class inclusion is
transitive.)
Whatever is a member of
necessarily

*
*

relationship)

a member of a class included in that class.

Class exclusion is symmetric.

Whatever is a member of
the first.

a class is not (as a result of that

a class is not a member of a class excluded from

Whatever is not a member

of a class is not (as a result of that
relationship) necessarily
also not a member of a class in which the first
is included.

Whatever is not
a member of a class is not (as a result of that
relationship) necessarily
a member of (nor a non-member of) another class
which is excluded from the first.

Whatever is not

a member of a class is also not a member of any class

included in the first.

For purposes of discussion

we have grouped most of these

conversion, negation of
'only-if' principles.

of development of knowledge and readiness,

principles according to the following headings:

antecedent, contraposition, transitivity, and

Although the terms used in this list of principle

groups are primarily logical terms,

principles are for psychological

system. The groupings and

a statement, which, roughly speaking,

the original statement.

are Conditional Principles 3

negating the first part

The conversion principles

and 7, and Class Principle

3. We call them

because they indicate logical fallacies.

Negation of Antecedent:
invalidlmove.

This is a rough classification

around with the beginning and end exchanged, does not

necessarily follow from

"fallacy principles",

reasons as well.

of

descriptions are inexact but useful.

Conversion: The converse of

is the statement turned

decisions about categorization

This is also an

It consists of

of a statement and on the basis of that negation,

concluding that the second

Principle

part must be negated as well.

Conditional

VIII-7

2 and Class Principles 6 and 7 are negation-of-antecedent principles, and

thus are also fallacy principles.
Contraposition:

This is a valid move.

It consists in negating the

second part of a statement and on that baAis concluding that the first

part should be negated also.

Conditional Principles 4 and 6 are contra-

position principles, as are Class Principles 5 and 8.

Transitivity:
studied.

This too is a valid move for the types of logic we

It is best explained by a symbolic example:

If X has a given

transitive relationship to Y, and Y has that sere transitive relationship

to Z, then X has that relationship to Z. Being taller than is a transitive

relationship, for example.

A transitive relationship is one, so to speak, that

carries through an intermediary when an intermediary exists. Conditional

Principle 5 and Class Principle 2 are clear cases of transitivity principles.

Conditional Principle 1, is like a transitivity principle, even though it
does not quite fit the symbolic form mentioned above. It is similar

in that the affirmation of the antecedent is carried through to the

consequent.

'Only-if' principles:

These are conditional principles which we have

grouped together because of the

arguments to which they apply.

occurrence of the phrase 'only if' in

The phrase 'only if' indicates a necessary,

but not necessarily sufficient condition for the truth of the rest of the
sentence.

arguments.

Three of them indicate valid arguments and two indicate invalid

The principles are 8, 9, 10, 11, and 12.

together because although there

We group all

are important differences, they seemed to

us to be conveniently grouped for teaching.

viii-8
An intensive study of Pieget's logic was conducted for the purpose of

comparison with the types of logic presented above. Basically we found the
logics similar although we had reservations. We concluded that Piaget's
conclusions and cur conclusions are therefore relevant to each other.
Because of the large amount to be learned in each of the two types
of logic we studied, we assigned these two types of logic to alternate grade
levels within the range studied. Conditional logic was tested for and
taught at grade levels 5, 7, 9, and 11, while class logic was tested

for and taught at grade levels 4, 6, 8, 10, and 12.

Basic Data on Subjects.
Although we had 803 subjects distributed in grades 4-12,(inclualve), not
all were used in the readiness part of the study. All vere used in the
natural-cultural development part of the study. At each grade level one

class-size group was selected for

purposes of instruction in logic. There

were 217 of these subjects, called the "LDT's" (for 'Logic Deliberately

Taught) who by design

were intended to be fairly representative of the

upper New York State school system from which they were drawn.
From the same school system comparable classes were selected, again

one at each grade level.

These subjects, called the "LNDT-1's" (for

Not Deliberately Taught, 1st group'),

'Logic

were 211 in number.

From another upper New York State school system another 375 students,
called the "LNDT-2's", were selected for use, together with the LDT's and

the LNDT-1's, in the natural-cultural development part of the study.

The following table gives

"SES" stands for socio-economic

some gross figures about these subjects.

status, and is based upon a seven-point

occupational rating scale developed by Warner (1949, pp140-141).

1 is

VIII -9

high and 7 is low.

IQ scores for LDT's and LNDT-1's are Lorge-Thorndikes;

for LINDT-2's, predominately the California Test of Mental Maturity.

Further details about the instruments and data-gathering procedures can

be found in Chapter III.

Table VIII-1

Gross Basic Data on Subjects.

LDT's
LNDT-1's

217
211

Mean IQ
116.8
116.3

LIET-2's

375

108.4

None Calculated

112.7

3.6 (LDT's and LNDT-1's only)

Total

803

SES
3.8
3.5

The Cornell Deduction Tests.

Two deduction tests

logic treated.

were developed, one for each of the types of

They are called "The Cornell Conditional-Reasoning Teat,

Fbrm X" and "The Cornell Class-Reasoning Test, Form X". These tests were
constructed with the purpose of determining whether all but one of the
principles of logic listed earlier are mastered or alternatively, the

degree to which they are mastered.

The one, Conditional Principle 12,

was omitted in order that combinations of principles could be included

without lengthening the conditional

test.

Each test contained 72 items

in 12 item groups of 6 items apiece. Each group of 6 items 'embodies a

principle or combination of principles.

The six items within any one

group

were scattered.

The tests were deliberately kept short in order that

they could be administered in a 40-minute period.

Operational definitions of mastery of a principle were tied to the item

groups.

Getting at least 5 of the 6 items correct was deemed to be a

sufficient condition for probable

mastery of the principle and getting at

VIII -10

least. 4 correct was judged to be

a probable necessary condition. (Thus

getting three or fever correct implied

the following are the operational

lack of mastery.)

Fbrmally put,

definitions of 'mastery of Conditional

Principle 14

If Y is given "The Cornell

Conditional-Reasoning Test, Form X" under

standard conditions; then if Y answers correctly at least five of items

and 39, Y has probably mastered Conditional

8, 16, 22, 29, 35,

Principle 4.

If Y is given "The Cornell

Conditional-Reasoning Test, Form X" under

standard conditions; then it

/,_Aciple 4 only
is probable that Y has mastered Conditional
if Y
answers correctly at least four of these items:
8, 16, 22, 29, 35, and 39.*

Making use of the distinctions

(1928) ve includ

among types of content made by Wilkins

"'tree content components

familiar, the symbolic, and

contains four concrete

in the tests:

the concrete

the suggestive components. Each item group

familiar items, one symbolic item, and one suggestive

item.
Concrete familiar items

no reason to believe that

knowledge, accept

are as the name suggests;

a subject will, because of background factual

or reject the conclusion of concrete familiar items.

previous sample arguments

are about familiar concrete
student will believe

to objects in the
items are such that

They

things, but there is no reason to think that a

items used symbols instead of words referring

sentences in the argument. And the suggestive component

the validity status of the argument differs from the

conclusion, which truth status is presumed to be well-

known to the subjects.

black", Which

about cars have concrete familiar content.

The

or disbelieve that all of Mr. Smith's cars are Fbrds.

The symbolic component

truth status of the

furthermore, there is

For example if the conclusion were "All cats are

presumably is known to be false, then the argument would be

* The theory and defense of

such definitions have been presented elsewhere by the

principal investigator (Ennis, 1964).

VIII -11

either valid or such that it does not

imply the denial of the conclusion,

and would have to be recognized

of class item 31, which contains

as such.

This example is the conclusion

a valid argument, thus requiring an

answer of "Yes."
The directions do not vaguely ask

for a judgment about the conclusion.

They ask whether, on the assumption

that nothing but the premises are

known, the conclusion would be true.

Stx sample problems are used to

illustrate this type of question.

Reliability.

Reliability was estimated by means of test-retest

correlations at each grade leve)

There was approximately

for the LNDT-1's and LNDT-2's combined.

a ten-week interval between test and retest.

mean of total score correlations

The

was .75 on the conditional reasoning test

and .83 on the class reasoning test.

Component score reliability

estimates are lower.

Mean concrete

neutral

test-retest correlations

class reasoning.

were .65 for conditional reasoning and .79 for

Symbolic component mean test-retest correlations were .53

and '.50 respectively.

For the suggestive component the means were .55

and .63 respectively.

As might be expected

again lower.
Validity.

the reliability estimates for the item groups were

Means were .52 for conditional and .50 for class reasoning.

Because of the lack of a dependable outside criterion, no

concurrent or predictive validity

estimates were computed.

Basically

the

type of validity with which

of the argument for

we were concerned was construct validity. Much

construct validity here depends on the content analysis

on which the best was based.

A study of reasoning in newspaper

editorials,

an auto mechanics handbook, and two

U.S. Supreme Court decision opinions,

VIII -12

together with intensive

study of logic and consultation with logicians,

provided the basis for the

A series of tryouts,

was conducted.

content analysis judgments.

concerned with both content and form of the items

The finished product was the result of eighteen months of

concentrated effort

on the nature and types of logic,

and content. An inspection of the

of the principles of logic

item format,

items reveals them to

directions,

be clear applications

presented earlier. The details of item

assignment are to be found in Chapter IV.

The rest of the argument
areas:

for construct validity is to be found in three

correlations with familiar measures, item analysis, and the role

by this test in the

current study. Correlations with familiar

measures, which are mainly based

maybe found in detail in

correlations, 188 on the

reasoning test.

Chapter IV; 423 subjects were used for these

chronological age at a given grade level were

in conditional and class reasoning respectively, which

statistically significant.* In other words, at a

given grade level, given

grades, there is if

the present system of advancement through the

anything, only a slight negative relationship between

age and mastery of logic.

then, as might be

When the various grades are grouped together,

expected, there is a clear positive relationship between

age and logic mastery, the

tests respectively.**

which are statistically

All significance
These particular

correlations being .58 and .68 for the two

Both are statistically significant.

Mean correlations with

and 235 on the class

Some of the more important ones follow.

relationships are not

upon pre-tests given to LDT's and LNDT-1's,

conditional reasoning test,

Mean correlations with

-.11 and -.12

played

(Lorge-Thorndike) IQ were .58 and .52 respectively,

significant and are about what one tends to find in

tests used the 5% level as the criterion.

correlations were based upon a sample of the LDT's;
= 64; class reasoning, N = 82.

conditional reasoning N

"III-13
correlations between subject matter tests and IQ.
Correlations with the
occupationally- based index

of socio-economic status averaged .20 and .26

on the two tests respectively,

both of which correlations again are

statistically significant.
There appears to be

no relationship between mastery of logic (among

students not yet exposed

to instruction in logic) and sex. The mean

correlations were .10 and

significant.

.00 respectively and were not statistically

If a relationship

does exist, it presumably is a very weak

one.

The :wove relationships

contribute to the argument

and fit into

between these two tests and common measures

for construct validity insofar as they make sense

some sort of theory (though it be at a low level).

That

there

be either no relationship

or a slight negative relationship between knowledge

of logic and chronological

given that promotion

age within a given grade is not surprising,

is not a completely automatic thing.

If bright

young

students can at least

ledge of logic is

occasionally move ahead,

related to brightness,

between chronological

on the assumption that know-

then a positive relationship

age and knowledge of logic at a given grade level

would be defeated.
If knowledge of logic
should be a positive

can be presumed to increase with age, then there

correlation between logic

age, when separate grades

intellectual ability, then

are grouped together.

scores on these logic tests

to IQ scores. On the same assumption,

there is greater

test scores and chronological

with the additional

And if logic is an

should be related
assumption that

intellectual development among the children of. he higher

socio-economic classes (excluding

upper-upper classes, of whom there were

VIII-14

few or none in this study) at least

some positive relationship between

socio-economic status and logic test

we know of no reason to expect one

score should be found. And lastly

sex to be better than another on a test

in logic.
The results on these tests conformed to expectations and are under-

standable and plausible.

Hence these correlations with familiar measure

support the argument for construct validity. Naturally this is only support,
not confirmation, since other interpretations could be given to

the data.
Item analysis information, which

and the Appendix, gives

is presented in detail in Chapter IV

us more information about the test.

theory to predict the results of the

We have no

item analyses that we performed, but

do not find the results exceptional.

The mean difficulty levels of

the total tests (ranges: 47.5* -

61.5

- 72.7 for class) show that the tests are not too
difficult for the grades at which

they were administered -- if one is

Since our major interest is in whether or not

a principle has been mastered, this criterion is not one that must be

satisfied in order

that the

test be uivable for our purposes.

That it is

satisfied is interesting, but not crucial.

The mean discrimination indices
29.8 on the class test.

the various item

were 26.0 on the conditional test and

When the mean discrimination indices were computed

groups at the various grade levels, some variability

which will be discussed

for

was evident,

under the next topic. There was one

group which consistently failed to discriminate,

* These figures are

given grade level.

means of the percentages of students getting each item right at a

VIII-15
which calls for the combined

hL

application of Conditional Principles land

This lack of discriminating

of the items in this

power we attribute to the great difficulty

group (index hovering around 20% for all grade levels).

Since the function of the

test is to see if certain principles and

combinations thereof are mastered,

the lack of discriminating power of this

item group is not a defect.

with similar groups for

It would be a defect if the test were used

purposes of discriminating among the members of

those groups.

The last type of evidence

regarding construct validity is the role

played by these tests in this

been used in other studies,
The rest of this chapter,

and other studies.

Since they have not yet

only the results of this one can be presented.

in its description and analysis of the results,

is thus

assumption
that greater understanding

The Natural Cultural

of test results contributes to construct validity.

Development of Knowledge of Logic.

Partly because Piaget's claims

instruction) of knowledge of logic
partly because we worked
study actually

about development (without deliberate

are vague and sometimes ambiguous, and

with students in the age range 10-18, there is in

this

no direct and unequivocal test of his views. But his

studies did suggest the following questions:

Is there actually
older?

a development of logical ability as children grow

Does this development (if

there is any) come in stages?

Is conditional logic mastered by age 11-12?

Is class logic mastered by age 11-12?

Is the truth-validity

characteristic (essentially the mastery of

the suggestive component) achieved by age 11-12?

VIII -16

Within each type of logic,

is there a development of one sort of
thing before another such that

some things are more difficult than

others at a given level?
Answers to the first five
of these questions are suggested or indicated

,are suggested by our

in the literature/and

study,

are fairly obvious and routine for anyone who has

worked with children and

thought about the questions.

The answers are as

follows:

Yes.

That depends on what

you count as a stage, but gross plateaus

extending over a period of years were not evident.

Partly.

Partly.

Partly.

The sixth question is

an extension of the 3rd, 4th, and 5th in that it is

concerned with the extent

sixth question which

to which various parts are mastered.

It is the

is the interesting one to our minds and is the one

toward which this study

has made contribution,

if it has made any contribution

in the area of the

natural-cultural development of logic.

Part of the
contribution lies in the
formulation of the question, and part lies in the

empirical findings.
the size of the

The latter part is inevitably quite modest because of

group and the fact that it is composed c: students from only

one small part of the world.

We found considerable
two types of logic studied,

to be easier all along

fallacies (conversion,
at ages 10-12, but there

similarity in the developmental patterns of the

with the exception that class logic appeared

the line. The principles expressing the basic

negating the antecedent) were the most difficult

was considerable improvement in these over the leriod

VIII -17

studied (age 10-18).

ones contrast with the
at medium difficulty

Among the validity principles, the contraposition

transitivity principles in that although both are

at age 10-12, there is little improvement over the

years in the contraposition

improvement in the

principles, whereas there is considerable

transitivity principles.

Although we found

parallel patterns among the types of principles in

the two types of reasoning studied,

between the two types

of reasoning when we examined the three components

of reasoning (concrete

into the tests.One

regular improvement

both tests.

we did not find as much parallelism

familiar, symbolic, suggestive) which were built

parallel

feature that

we did find was the

fairly

over the years studied in all three components on

But there was a marked difference.

On the conditional

reasoning test the three components were of about

equal difficulty at each grade level.

the concrete familiar

component was in general

easier than each of the other

components, particularly the symbolic

the mean difficulty index differences

and the symbolic component exceeded

component. At all five grade levels

between the concrete familiar component

ten percentage points, and in three of

the five grade levels the differences

and were statistically

But on the class reasoning test,

exceeded twenty percentage points

significant.

differences between tL
familiar components and the suggestive

And the

concrete
component also exceeded ten percentage

points at all grade levels tested.

One of these differences was statistically

significant.
Previous studies that have shown the
to be more difficult

suggestive and symbolic components

than the concrete familiar component have all used class

reasoning.

One wonders why

the difference in difficulty did not appear in

vin-18
conditional reasoning.
suggestive and symbolic

component in class

In developmental terms one wonders why the

components lag behind the concrete familiar

reasoning, but not in conditional reasoning.

Explanations

Which were proposed in

Chapter V suggest that the differences result from

the nature of the logic being investigated

uninstructed students

dealing with the

sentences.

The Development of

following rough and

to master a principle was studied and the

somewhat vague analysis was the result: TO say that

.o master Principle X

given suitable

use of symbols to represent

Readiness to Master Logic.

The concept readiness

Y is ready

and from the difficulties that

have in

is to say that Y has the disposition,

conditions, to develop another disposition, which is the

disposition to show,

given suitable conditions, the sort of behavior that

a person who has mastered

a prior specification

is in part provided

of the meaning of 'mastery of Principle X', which

by the operational definitions given earlier.

analysis is complicated

a disposition is under consideration.

use of the vague terms

this is because the concept being

person using the concept

The

by the double dispositionality of the concept; a

disposition to develop
analysis twice makes

Principle X would show. This analysis assumes

And the

'suitable conditions'.

analyzed is vague in the same way.

readiness must judge whether the amount of effort

needed to achieve the desired

And he makesieplicitreference

mastery or learning is worth the trouble.

to possible means for achieving the mastery

But
The

or learning but he need

not specify the means.

A person claiming that

Y
is ready to master Principle X implies

that the means of bringing Y to this

VIII-19
mastery exist and are feasible, but the means are not necessarily

specified in the statement claiming

like the phrase, 'suitable conditions',
In this part of the study,

readiness.

is vague.

we concentrated on two empirical questions

in addition to the question of

the analysis of readiness

What were our subjects ready to master?

What are fairly comparable

The first question is

Hence the concept,

for mastery:

students ready to master?

a question of description and the

second a question

of prediction.

In the first question the term

They were given the pre-test,

'our subjects' refers to the LDT's.

were taught logic for 15 days, one period

per day, and approximately six weeks later, were given the post-test,
which was the same test

as was used for the pre-test. Teaching

members of the project staff,

experienced teachers

who were given special training in logic

What Our Subjects Were Ready to Master.

two types of logic that we found in the

natural-cultural development of

teachability of principles

principles was not to be found in the

- nor in improvements in component and total

In class logic modest

improvements in total scores, component

scores, and principle scores generally

upward during the teaching

students in our study registered

secondary school students

component scores, and the

occurred in students from age 11-12

period.

But in conditional logic the younger

virtually no improvement,

by

and were

at the grade levels with which they worked.

The parallelism between the

scores.

was done

While the upper

registered a marked improvement in total scores,

fallacy principle scores after having been taught.

VIII -20

Class reasoning is apparently teachable to some extent from age 11-12

onwards.

Students younger than that did not benefit from the 15 days of

instruction that we were able to give

them. Perhaps under different

conditions

- or with more time - they also would have benefited.

From

11-12 onward there is apparently modest fairly even improvement as a

result of whatever natural-cultural

teaching of the sort we did

By age 17-18 there was

sources are operating, and deliberate

can contribute modestly to this improvement.

as a result of existing natural-cultural influences

on our LDT's considerable mastery of the basic principles of class logic.

Our teaching made a modest improvement. Overall, talking in terms of

readiness, we can say that from

age 11-12 onward our students were ready

for modest improvements in mastery

and that by age 17 -18, the

of the principles of class reasoning,

group as a whole was ready to make the modest

our saying that for practical purposes

they have mastered the basic principles

Conditional logic makes

a different story.Apparently, when given the

sort of instruction we provided,

improvement until upper secondary;

great strides.

our LDT's were not ready to make much

but by age 16-17 were ready to make

These great improvements in mastery were particularly

evident among the fallacy

principles (where there was much room for

improvement); but they also occurred among the contraposition principles,

the affirming-the-antecedent

principles.

principle and to a slight extent the transitivity

No improvement was registered among the 'only-if' principles,

though this might be because

the outset the 16-17

year o ],ds were fairly good at them.

It is quite possible that
our data suggest a line between the ready and

nonready that is sharper than

it should be. We have some question about the

VIII -21

motivation of the 9th grade LIIP's.

and large ready to master the

were not.

In any case our 11th graders were by

fallacy principles, and our 5th and 7th graders

This is the most striking finding in this readiness study.

That Which Fairly Comparable Students Are Ready to Master.

the relationships between

and the LDT post-test

Two sets of equations

scores on certain variables (described earlier)

scores, and the interrelationships among these

variables, we constructed
of predicting what other

Using

multiple-regression equations for the purpose

fairly comparable students are ready to master.

were prepared, one making use of pre-test scores on

the same logic test, and the

other not doing so.

sets are grade level, chronological

All grade levels on each test

11 (94 students in all)

Variables used in both

age, IQ, socio-economic status,

and sex.

were combined; that is grades 5, 7, 9, and

were combined for the conditional logic equations,

and grades, 4, 6, 8, 10, and 12 (123

class logic equations-

students in all) were combined for the

Equations for each item-group score, each component

Altogether that makes 64 equations.*

Total score cross validation

pre-test were .77 and .69

estimates with and without the use of the

respectively for conditional logic

.78 respectively for class logic.

ranged around .70 for conditional
with the ones using pre-test

Conditional logic item

Component score cross validation estimates

logic and around .75 for class logic,

score running higher than the ones without.

group cross validation estimates ranged around .50

and .40 respectively with and

without pre-tests; class logic estimates

ranged around .60 and .55 respectively.

These item-group cross validation

prediction of individuals

(12 + 3 + 1) (2) (2).

scores.

estimates are not adequate for the

Whether they are high enough for prediction

Not all item groups uniquely tested for a principle.

were twelve item
groups in each test and an equation constructed

for each one.

There

VIII-22

of central tendencies of
interest in working out

groups is an open question, but because of our

a technique for predicting mastery in a group, we

developed such a technique

and offer it as one which must be refined and

investigated, both empirically and statistically.

The technique is as follows:

of the variables for

the group in which we have an interest. &tract a

mean item group score.

difficulty index.

Convert the predicted score into an item

to meet the necessary condition for mastery.

operational definitions
that are predicted to

on p VIII-10.)

the Appendix in Tables

is to be found in

for constructing these

are ready to caster and

equations can be found in

A-11 through A-14. The graph for converting from

to necessary and sufficient condition percentages

Chapter VI (Graph VI-1). The following is a sample working

above-described

11th grade LET's for

percentage that

arc not ready to zeisttr it.

Weights and constants

mean difficulty indices

If the pre-test has been

obtain the percentage that already %eve mastered

the principle, and estimates of the

the percentage that

(See

The results are the percentages

obtain after instruction.

given, then one can readily

ix is

group mean

conversion to percentages m.zting the sufficient condition

for mastery and failing

Principle

predicted

Use this to enter a graph which provides an

empirically derived

through of the

Inszrt into the equation the mean values

process, making use of the mean value of the

purposes of illustration.

the subject of this prediction

Values going into the equation:

Grade
Chronological Age

IQ
Socio-economic Status
Sex (male, 1
female, 2)
Total Score on Pre-Test

Pre-test Score

on this Item Group

Mastery of Conditional
effort:

Measured
Value (rounded)
11
199.8 mos.
116.0
3.5
1.5
63.7

4.26

Weight
-.001
.0074
.0064
-.127
.423
.0384

.314

Constant:

-1.26

VIII -23
Predicted score = (-.001) (11) + (.0074) (199.8) + (.0064)

(116.0) + (-.127) (3.5)

(4.26)

1.26

+ (.423) (1.5) + (.0384) (63.7) + (.314)

= -.01 + 1.48 + .74 -

.45 + .63 + 2.44 + 1.34 - 1.26

= 4.91, predicted score.

This figure, 4.91, can be rounded or not, depending on what one is

going to do with it.

The predicted score of 4.91 corresponds

item group of 82% (4.91/6 = approx. .82).

predicted percentages of 13% failing

72% meeting the

to a mean difficulty index for this

Entering graph VI-1 we obtain

to meet the necessary condition and

sufficient condition for mastery.

Incidentally the actual

percentages turned out to be 15% and 77%.

04 the pre-test 23% failed to

met the sufficient condition.

meet the necessary condition and 50%

Hence the claim that results from the prediction

that probably at least 22% (72%

principle and at least 13%

is

minus 50%) are ready to master the

are not ready to master it.

mastered it; the other 15% (100--50

50% had already

-22 -13 = 15) make up the group about

Which we do not want to make any sort of commitment.

Needed FUrther Research.

This study has only begun to

explore the topic.

The following steps

are desirable:

1. The checking of the

response form (multiple choice) of the

tests against an open-ended test form.

Revision of the tests to adapt them to students in the

first three grades.

Lengthening of the tests

so that measurement of mastery of each

principle by an individual becomes more reliable.

VIII-24

Replication with more classes at each grade level in different

types of environment to minimize the effect of class-wide factors. The

in-between grade levels should also be used. This should be done for both
the natural-cultural and readiness parts of the study.
5.

&tension with adaptation to lower age levels (ages 6-10),

again

for both empirical ;arts of the study.

The use of variation in amount of teaching time as a variable

in the multiple regression equations.

More radically, the training in logic of full-time teachers, followed by

their instructing logic in their own classes, in order to secure more

time devoted to logic teaching in LDT classes, and at the same to maintain

a more realistic situation.

Again the readiness questions would be asked.

8. A close examination of the

source of improvement in conditional

logic among the LDT upper secondary students, with this specific question

in mind:

Is this learning merely the acquisition of new meanings for words

or is it an increase in grasp of concepts and principles?

Eapirical and mathematical investigation of techniques for predicting

percentages who are ready to master a principle, making use of the

operational criteria advanced here (or similar criteria).

Such investigation

should attempt to indicate the extent to which such procedures are likely to

be in error.
10. The investigation of the readiness development

and natural-cultural

development of other aspects of critical thinking, including analysis of

the aspect, test development, and testing and teaching

of the aspect. It

would probably be better in the future if existing classroom teachers were

used, but only after thorough training.

VIII -25

Overview.

The products of this study are as follows:

*

Two logic tests, one on each of these types of logic.

A theory of operational definitions.

Operational definitions of mastery of each of the principles in

terms of the two tests.

An analysis of Piaget's conception of logic.

A suggestion of the patterns of natural-cultural development of
knowledge of logic from ages 10-18 among students like those studied
here.

An analysis of the concept, readiness to master a principle.

A description of what amounts of logic our subjects were ready

to master (given 15 days of logic instruction).

The specification of a set of basic principles of two "ajor

types of logic, conditional logic and class logic.

A set of 64 multiple regression equations for predicting the

amount of mastery of logic (given comparable instruction) in
groups comparable to those with which we worked.

A procedure to estimate the percentage of students in a given

class who are ready. L.:poster and the percentage who are not
ready to master a given principle of logic.

FUrther research is needed on all these items.

CITED REFERENCES

Bonner, Frederick G. The Reasoning Ability of Children of The Fourth/

Fifth and Sixth School Grades.
New York: Teachers College,
Columbia University, 1910.
"Statistical Theory and Design," Annual Review of
Psychology, Vol 5 (1964) , pp. 377-400.

Brogden, Hilbert E.

Burt, Cyril. "The Development of Reasoning in School Children,"

Journal of
imental Peda o
Vol. 5, No. 2 and 3 (1519),
PI"

Tr, 121427.

Cohen, Morris:R. and Nagel, Ernest. An Introduction to Logic and Scientific

Method. New York: Harcourt, Brace and Co., 193 .
Cronbach, Lee J.
Educational Psychology.
and Co., 1954.

New York: Harcourt, Brace

Ennis, Robert H. "A Concept of Critical Thinking," Harvard Educational

Review, Vol 32, No. 1 (1962), pp. 81-111.

"A Definition of Critical Thinking." In Richard E. Ripple (ed.),

Readings in Learning and Hunan Abilities: Educational Psychology.
New York: Harper and Row, 1964. Pp. 286-304. Also: The Reading
Teacher/ Vol. 17, No. 8 (May, 1964), pp. 599-612.
"Operational Definitions," American Educational Research
Journal, Vol. 1, No. 3 (1964c), pp. 183-201.
.

Fairgrieve, M. MacCallum.

"A Mass Test With Mr. Cyril Burt's Intelligence

Westions," Journal of Experimental Pedagogy, Vol. 6 (1921),

pp. 27-33.

Revell, John H. The Developmental Psychology of Jean Piaget.

D. Van Nostrand Co., Inc. Princeton, New Jersey: (1963).
Hill, Shirley Ann. "A Study of Logics'. Abilities in Children."
doctoral dissertation, Stanford University, 1961.

Unpublished

Inhelder, Barbel and Piaget, Jean. The Growth of Logical Thinking from
Childhood to Adolescence: An Essay on the Construction of Formal
Operational Structures. London: Routledge and Kegan Paul, 1958.
Miller, Elmer H. "A Study of Difficulty Levels of Selected Types of
Fallacies in Reasoning and Their Relationships to the Factors
of Sex, Grade Level, Mental Age and Scholastic Standing," Journal
of Educational Research, Vol. 49, No. 2 (1955) pp. 123.429.

R-1

R-2
Moore, Dom Thomas Verner.
First
"The Reasoning Ability of Children in the
Years of School Life,"
Studies in Ps chol and Ps chiat

from the Catholic University of

Parsons, Charles.

"Inhelder

and Piagett

America, Vol. 2, No. 2 April, 1929).

s The Growth of Logical Thinking,

II.

A Logician's Viewpoint," British Journal of Psychology, Vol. 51,

No. 1 (1960), PP- 75-84.

Piaget, Jean.

Judgment and Reasoning in The Child. Paterson,

Littlefield, Adams and Co., 1928.
The Psychology of

Kegan Paul,

Intelligence.

Routledge

and

1950.

Logic and Psychology.

London:

New Jersey:

New York: Basic Books, Inc.,

Ryan, Thomas A.

Research,"

"Multiple Comparisons in Psychological

Psychological Bulletin, Vol. 56, No. 1 (1954) .

Smith, B. Othanel.

1953.

"The Improvement of Critical Thinking," Progressive

Education, Vol. 30, No. 5 (March, 1953). pp. 129-1314..

Strawson, P. F. Introduction to Logical Theory.

Ltd. 1952.

London: Methuen & Co.,

Sweeney, E. J. "Sex Difference in Problems Solving." Unpublished doctoral

dissertation, Stanford University, 1953.

Warner, W. Lloyd.
Social Class in America.

New York: Harper and Brothers,

1949.
Wilkins, Misuse. Cheves.

The !Effect of Changed Material

on Ability to

Do Forma Syllogistic Reasoning. Archives of Psychology, No. 102,

192g.

Winch, W. H.

"Children's Reasonings:

In School Children,"

Vol.

(1921),

Experimental Studies of Reasoning

The Journal of

Experimental Pedagogy,

pp. 123.-141, 199-212, 275-2g7.

Woodcock, Louise P.

Life and Ways of the Two-Year-Old. New York: E.

P. Dutton and Co., Inc., 1941.

Wright, Georg Henrik Von.

Routledge and

Kegan Paul, Ltd., 1957.

Logical Studies.

London:

GENERAL REFERENCES

Note:

The following list of references includes those in the

previous list of cited references.

Anderson, Howard C., Marcham, Frederick G., and Dunn,

Seymour B.

"An Experiment in Teaching Certain Skills of Critical Thinking,"

Journal of Educational Research: Vol. 38, No. 4 (December, 1944),

pp. 251 -251.

Arleta, S.

Interpretation of Schizophrenia. New York:

Brunner, 1955.
Arnold, Dwight L.
"Testing Ability to Use Data in the Fifth and Sixth Grades,"
Educational Research Bulletin, Vol. 17, No. 9 (December 7, 1938), pp.
255-259, 278.

Baldwin, G. Bernard.

Inductive Reasoning Test. Minneapolis, Minnesota:

Educational Test Bureau, 1946.

Black, Max.

Critical Thinking (Second Edition). New York: Prentice-Hall,

Inc. 1952.

Bonser, Frederick G.

The Reasoning Ability of Children of the Fourth, Fifth,

and Sixth School Grades. New York: Teachers College, Columbia
University, 1910.

Brogden, Hubert E.
and Design,"

"Statistical Theory

Annual Review

of Psychology

Vol. 5 (1954), PP- 377-400.

Bruner, Jerome S., Goodnow, Jacqueline J., and Austin, George A. A Study

of Thinking.

New York:

John Wiley & Sons, Inc., 1965.

Burt, Cyril. "The Development of Reasoning in School Children,"

rimental Peda
Vol. 5, No. 2 & 3 (1919),
Journal of
PP-

77, 121 -127.

Chapman, L. J., and Chapman, J. P. "Atmosphere Effect Re- examined,"

Journal of Experimental Psychology, Vol. 58 (1959) pp. 220-226.
Chase, Stuart. The Tyranny of Words. New York: Harcourt, Brace and
Company, 1936.

Cohen, Morris R. and Nagel, Ernest. An Int-_loisailljaWAliaail

Scientific Method. New York: Harcourt, Brace and Company, 1934.
Conant, James Bryant. On Understanding Science. New York: Mentor
Books, 1951.

Cooperative Study of Evaluation in General Education of the

Council on Education. A Test of Critical Thinkinj, (Ppm G).
New York: American Council on Education, 1952.

R-3

American

R-4

Crawshay -Williams, Rupert. Methods and Criteria of Reasoning.

London: Routl edge and Kagan Paul, 1957.
Cronbach, Lee J. Educational Psycholora. New York: Harcourt, Brace
and Company, 1954, p. 272.

Dennis, Wayne. l'iaget's Questions Applied to a Child of Known

Environment,"
The Journal of Genetic Psychology, Vol. 61 (1942),
pp. 307-320.

Deutsche, Jean Marquis.

The Development of Children's Concepts of Causal
Relations.
Minneapolis: The University of Minnesota Press, 1937.

Dewey, John.

How We Think.

Boston: D. C. Heath & Co., 1933,

Dickinson, A. E., and Tyler, F. T.

"An Experimental Study of the

Generalizing Ability of Grade II Children," Journal of

Educational Psychology, Vol. 35 (1944), pp. 432-441.
Dressel,

Paul

L.,

Evaluation.

and

Mayhew,

Lewis

General

Education

Ekplorations

in

Washington, D. C.: American Council on Education, 1954.

Educational Policies Commission.

Education.

B,

The Central Purpose of American

Washington, D. C.: National Educational Association

of the United States, 1961.

Educational Testing Service, Cooperative Test Division. Logical Reasoning

Test, .Form B.
Princeton, New Jersey: Educational Testing Service,
1950.

Ennis, Robert H. "Critical Thinking: More on Its Motivation,"

Progressive Education, May, 1956, pp. 75-76.
.

"An Appraisal of the Watson-Glaser Critical Thinking Appraisal,"

Journal of Educational Research, Vol. 52 (1958), pp. 155-158.
.

"Assumption Finding," In B. Othanel Smith and Robert H. Ennis

Language & Concepts in Education. Chicago: Rand McNally,

1961, pp. 161478.

and Millman, Jason.

"Cornell Critical Thinking Test,

York: 1961a (mimeographed edition).

and Millman, Jason.

Ithaca, New

"Cornell Critical Thinking Test,

Ithaca, N. Y.: 1961b (mimeographed edition).
.

Form 74."

Form Z."

"A Concept of Critical Thinking," Harvard Educational Review

Vol. 32, No. 1 (1962), pp. 81411.

Ennis, Robert H. "A Definition of Critical Thinking,"-In Richard E. Ripple

(ed.), Readings in Learning and Human Abilities: Educational
Psychology. New York:
Harper and Row, 1964. Pp. 286-304. Also: The
Reading Teacher, Vol. 17, No. 8 (May, 1964), pp. 599-612.
.

"Needed Research in Critical Thinking," A paper presented

to the New England Regional Conference on Educational Research,

Rhode Island State College, Providence, Rhode Island, November,

1964b.
.

Mimeographed.

"Operational Definitions,"

American Educational Research

Journal, vol. 1, No. 3 (1964c), pp. 183-201.

Fairgrieve, M. MacCallum.
"A Mass Test with Mr. Cyril Burt's
Intelligence Questions,"
Journal of Experimental Pedagogy,
Vol. 6 (1921.), pp. 27-33.
Flavell, John H.

The Developmental Psychology of Jean Pia et. Princeton,

New Jersey: D. Van Nostrand Co., Inc., 1963.

Glaser, Edward M.

An Experiment in the Development of Critical Thinking.

Teachers College Contributions to Education, Number 843. New York:
BLreau of Publications, Teachers College, Columbia University,

1941.

Gordon, Raymond L.
"The Effect of Attitude Toward Russia on Logical
Reasoning,"
Journal of Social Psychology, Vol. 37 (1953), pp. 103111.

Gottschalk, Louis.

Understanding History.

New

York: Alfred A. Knopf,

1951.

Guillet, Cephas.

"The Growth of a Child's Concepts," Pedagogical

Seminary, Vol. 24 (1917), pp. 81-96.

Hazlitt, V.

"Children's Thinking,"
Vol. 20 (1930), pp. 354-361. British Journal of Psychology,

Hempel, Carl G.

Fundamentals of Concept Formation in Empirical

Science. Chicago:

Hdrix, G. "Developing

University of Chicago Press, 1952.

a Logical Concept in Elementary Mathematics,"

Mathematics Teacher Vol. 48 (1955), pp. 542-547.

Henle, M.

"On Error in Deductive Reasoning," Psychological Report, Vol.

7 (1960): p. 80.

Henle, Mary. "On the Relation Between

Logic and Thinking," Psychological

Review, Vol. 4 (1962), PP- 366-378.

R-6
Henle, M., and Michael, M.

"The Influence of Attitudes on Syllogistic

Reasoning," Journal of Social Psychology, Vol. 44 (1956), pp. 115-127.

Hill, Shirley Ann.

A Study of Logical Abilities in Children.

Unpublished

doctoral thesis, Stanford University, 19E7Hood, Fred C.

Developing the Ability to Assess the Results of Thinking.

Urbana, Illinois: University of Illinois Bulletin, 1949.

Hospers, John.

"What is Explanation ?"

In Antony Flew (ed.)

Essays In Conceptual Analysis.

Macmillan, 1956.

London:

Pp. 94-119.

Hyram, George H. "An Experiment In Developing Critical Thinking In

Children,"
Journal of Experimental Education, vol. 26 (December,
1957), pp. 125-132.

Inhelder, Barbel, and Piaget, Jean.

The Growth of Logical Thinking from

Childhood tolAdolescence: An Essay on the Construction of Formal

Operational Structures.

London: Routledge & Kegan Pauf.,7W

Interpretation of Literature Tests. Princeton, New Jersey: The Cooperative

Test Division of the Educational Testing Service, 1950.

The Iowa Tests of Educational Development. Tests 5, 6, 7, Form Y-3S,

Science Research Associates, Chicago, Illinois, 1952.

Janis, I. L., and Prick, F.

"The Relationship Between Attitudes Toward

Conclusions and Errors in Judging Logical Validity of Syllogisms," Journal of

Experimental Psychology, Vol.33 (1943), pp. 73-77.

Kane, R. B.

"Some Effects of Beliefs About Conclusions of Arguments

on the Ability to Judge the Validity of the Argument," Unpublished

dissertation, University of Illinois, 1960.
Lee, Dorris May, and Bingham, Alma.

"Intellectual Processes," Review of

Educational Research, Vol. 29 (April, 1959), pp. 185-196.

Leeper, Robert.

"Cognitive Processes," In S. S. Stevens (ed.), Handbook

of Experimental Psychology. New York: John Wiley & Sons, Inc.,
1951. Pp. 730-757.

Lefford, A.

"The Influence of Emotional

Subject Matter on Logical

Reasoning," Journal of !...13sesallashalsilz,

Vol. 34 (1946), pp. 127-151.

Logical Reasoning Test, Form B. Princeton, New Jersey: The Cooperative

Test Division of the Educational Testing Service, 1950.
Luchins, A. S. "On the, Awareness and Denotation of Contradictions,"
Journal of General Psychology Vol. 71 (1964), pp. 233-246.

R-7
Macmillan, C. J. B. and Ennis, Robert H.

to Critical Thinking,"
Junior Hi

"Reviews of Works Related

Critical Thinki4gResearch Series)

School Pro ect

mimeographed

First Ed. Ithaca, New York: 1961,

out of stock).

Maier, Norman R, F.

"Reasoning in Children," The Journal of

Comparative Psychology) Vol. 21 (June, 1930, pp. 357-366.

Marcham, Frederick George.

"Teaching Critical Thinking and the Use

Evidence,"
The Quarterly Journal of Speech, Vol. 31, No. 3

of

(October, 1945), pp. 362-368.

McCarthy, Philip J.

"Sampling,"Bulletin No. 15 of the New

York
State School of Industrial and Labor Relations, Ithaca, New York:
1951.

Michigan State College Board of Examiners. A Test of Problem Solving.

Junior and Senior High School Edition, Form B.
Michigan State College, 1954a.

East Lansing:

Michigan State College Board of Examiners. A Test of Problem

High School Edition, Form A.
East Lansing:
Solving.
Michigan State College, 1954b.
Miller, Elmer H. "A Study of Difficulty Levels of Selected Types of
Fallacies in Reasoning and Their Relationships to the Factors of
Sex, Grade Level, Mental Age and Scholastic Standing,"
Journal of Educational Research, Vol. 49, No. 2 (1955),

129.

pp. 123-

Mohandessi, K.

"Development of a Test to Measure Logical Reasoning

of High School Students,"

(unpublished master's thesis)

University of Illinois, 1956.

Moore, Dan Thomas Verner.

"The Reasoning Ability of Children in The

First Years of School Life," Studies in Psychology and Psychiatry,
Edward A. Pace (ed.), Vol. 2, No. 2 (1929).

Relational

Morgan, J. J. B., and Carrington, D. H. "Graphic Instruction in

Reasoning,"
Journal of Educational Psychology, Vol. 35 (1944), pp.
536-544.
and Morton, J. T. "The Distortion of Syllogistic Reasoning
Produced by Personal Conviction," Journal of Social Psychology, Vol.
20 (1944), pp. 39-59.

Navarra, John Gabriel.

The Development of Scientific

Concepts In a
Younj Child.
New York: Bureau of Publications, Teachers College,
Columbia University, 1955.

Oakes, Mervin E.

Children's Explanations of Natural Phenomena. New York:

of Publications, Teachers

College, Columbia University, 1947.

Bureau

R-8
Oliver, Donald W., and Shaver, James P. "The Analysis of Public
Controversy: A Study in Citizenship Education," Vol. 1 and 2.
A Report from the Laboratory for Research in Instruction, Graduate
School of Education, Harvard University, 1962.

Parsons, Charles.
"Inhelder and Piaget's
Thinking," II A Logician's

The

Viewpoint.

Psychology, 1960, pp. 75-84.

Growth of Logical
British Journal of

Pearson, William A., and Hyram,

Critical
George H. Pearson -Hyram Test of
Thinking.
St. Louis, Missouri: Saint Louis University,
1958.
Peterson, George M. "An EMpirical
Study of The Ability to Generalize,"
Journal of General Psychology, Vol.
6 (1932), pp. 90-113.
Piaget, J. The Language and Thought of
the Child. New York: Harcourt, Brace
and Co., 1925.
.

Judgment and Reasoning in the

Littlefield, Adams and Company,

.

Child. Paterson, N. J.:

1959 (first printed, 1928).

The Psychology of Intelligence.

Paul limited, 1950.

Logic and Psychology.

London: Routledge and Kegan

New York: Basic Books, Inc.,

1957.
Richter, Maurice N. Jr. "The
Syllogistic Reasoning,"

pp. 341-344.

Theoretical Interpretation of errors in

Jouofras:liolt,-nal Vol. 43 (1958),

Robinson, Richard.

Definition.

Oxford:

Oxford

University Press,

1954.
Russell, David H.

Children's Thinking.

Boston:

Ginn and Co., 1956.

Ryan, Thomas A.

"Multiple Comparisons in Psychological Research,"

Psychological Bulletin, Vol. 56,
No. 1 (January 1959), PP- 26-47

Scheffler, Israel.

"Explanation, Prediction, and Abstract Entities,"

British Journal for the
Philosophy of Science, Vol. 7 (1956-7), pp. 293309.

Sell, DeWitt E.

Evaluation Aptitude Test. Chicago, Illinois:

Psychometric Affiliates, 1952.

Sells, S. B.

"The Atmosphere Effect,"

Archives of Psychology, 1936,

No. 200.

Shelley, H. P., and.Davis,

R. E.

Problem-Solving," Psychology

"Relationship of Attitude

to Logical

Report, Vol. 3 (1957), PP- 525-530.

R-9

Shenfeld, N. "Tolerant and Intolerant Attitudes and Logical Thinking,"

Dissertation Abstracts, Vol. 18 (1958), p. 2240.

Smith, B. Othanel. "The Improvement of Critical Thinking," Progressive

Education, Vol. 30, No. 5 (March, 1953), pp. 129434.
"Logic, Thinking, and Teaching," Educational Theory,
Vol. 7, No. 4 (October, 1957), PP- 225-233.
.

"Critical Thinking,"

Recent Research and Developments and Their

Implications for Teacher Education. 13th Yearbook of the American
Association of Colleges for Teacher Education, Washington, D. C. (1960).
.

Stewart, Daniel K. "The Laws of Thought:

Report, Vol. 7 (1960), pp. 229-232.

An Hypothesis," Psychology

Straw3on, P. F.

Introduction to Logical Theory. London: Methuen & Co.,

Ltd., 1952.

Suppes, Patrick and EULford,

Prederick. "Experimental Teaching of

Mathematical Logic in the Elementary School," The Arithmetic

Teacher, March, 1965, pp. 187-195.

Sweeney, E. J.

"Sex Difference in Problems Solving." Unpublished

doctoral dissertation, Stanford University, 1953.

Tabal Hilda.

Thinking in Elementary School Children. (Unpublished

summary of a Cooperative Research Study). San Francisco,
California: 1964.

Test of Aplication of Principles in Biology, Test of Application of

Principles in General Science, and Test of Application of
Principles in Physical Science.
Form A 1 Princeton, New Jersey: Cooperative
Test Division of the Educational Testing Service, 1950.

Thistlewaite, D. L.

"Attitude and Structure as Factors in the

Distortion of Reasoning," Journal of Abnormal Social Psychology, Vol. 45

(1950), pp. 442-458.

Thompson, Jane-

"The Ability of Children of Different Grade Levels to

Generalize on Sorting Tests," The Journal of Psychology, Vol. 2
(1941) , pp. 119-126.

Thorndike, E. L.

"The Effect of Changed Data Upon Reasoning," Journal

of Experimental Psychology, Vol. 5 (1922), pp. 33-33.

Thoule4s, R. H. "Effect of Prejudice on Reasoning," British

of Psychology,
Vol. 50 (1959), pp. 289-293.
Journal
Toulmin, Stephen Edelston.
The Uses of Argument. Cambridge: Cambridge
University Press, 1958.

R-10
Valentine, Charles W.

Intelligence Tests for. Children (Sixth Edition).

London: Methuen and Co., Ltd., 1958 '(First edition, 1945).

Von Domarus, E.

"The Specific Laws of Logic in Schizophrenia,"

In J. S.'Kdsmin,

(ed.), Language & Thought in Schizophrenia.

Berkeley: University of California Press, 1944.

Warner, W. Lloyd.

Social Class in America. New York: Harper and

Brothers, 1949.

Wason, P. C.

"The Effect of Self Contradiction on Fallacious Reasoning,"

Quarterly Journal of &perimental Psychology, Vol. 16, No. 1 (1964),

Pp. 30-34.

Watson, Goodwin, and 0.1c.1:cr,

Thinking Appraisal,
Book Co., 1952.

Welch, Livingstone.

Eduard Eftynard.

Watson-Glaser Critical

Forms Am and Pm. Yonkers-on-Hudson: World

"Some Aspects of the Development of Inductive

Reasoning," New York Academy

218-221.

of Science, Vol. 4 (1942), pp.

"Methods Used by Children

.
in Solving Inductive Reasoning
Problems,"
Journal of Psychology, Vol. 14 (1943) , pp. 269-275.
TWo
, and Long, Louis.
"Comparison of the Reasoning Ability of
Age Groups,"
The Journal of Genetic Psychology, Vol. 62
(1943), pp. 63-76.

Wigmore, John Henry.

A Students' Textbook of the Laws of Evidence.

Brooklyn: The FOundation Press, 1935.

Wilkins, Minna Cheves.

The Effect of Changed Material on Ability to Do Formal

Syllogistic Reasoning. Archives of Psychology, No. 102

(1928).

Williams, E. B.

"Deductive Reasoning in Schizophrenia," Journal of

Abnormal Social Psychology, Vol. 69, No. 1 (1964), pp. 47-61.

Willis, Mary, and Domoncovich, H.

and
A. Cooperative Literary Comprehension
Appreciation Test (Form T).
Princeton, New Jersey: The Cooperative Test
Division of the Educational Testing Service, 1943.
Winch, W. H.
In
"Children's Reasonings: Experimental Studies of Reasoning
School Children,"
The Journal of Experimental Pedagogy,

vol. 6 (1921), pp. 121441, 199-212, 275-287.

Woodcock, Louise P.

Life and Ways of the Two-Year-01d. New York: E.

P. Dutton & Co., Inc., 1941.

R -11

Woodworth, R. S., and Sells, P. B.

"An Atmosphere Effect in Fbrmal

Syllogistic Reasoning," Journal of Ekperimental Psychology, Vol. 18 (1935), pp.

451-460.

Wright, Georg Henrik Von.

Logical Studies. The Humanities Press Inc.,

London: Routledge & Kegan Paul Ltd., 1951.

The tables and graphs presented in this part of the appendix are

listed at the beginning of the

GraPhs".

report under the heading, "Tables and

"1
!

TABLE

Individual Item and Item Group Mean

Difficulty IndicesbyGrades on the Conditional

All Subjects on the Pre-test and Alsofor LDT Pre-and. Post-tests.

Sub-

jects LDT

LDT

Pre

Post

Pre

Grades

N=

102

05

Item

Number

p then q. p.

14

68
78

63

74

63

19

86

85

96

07
27

Mean

74
48
75

71

jects
Pre

63
59

67

68

78

70

78
78
77

61

76
95
75
72
71

75

34

13

24

23
26

05

18

34

Mean

32

27

41
37

37
00
44

19
37
44

04

33

Post

07

24

24

Pre

80

96

78

79
62

63

90
69
75
78

76

83

77

33 33

43

58
67
loo

88

71
96
83
92

Not q.

Item
09

Pre

Sub.
jects LDT

41

34

23

13
42

15

22

11

lo

23

30

25

28

29
21

04
25

13
21

5o

25

13

33
13
28

75

34
23

09
33

09

25

LDT's & LNDT-1's & LNDT

given in Chapter Iv. **4

-2'e. ** The basic

Pre

LDT

Pbst

09

17

17

SUbgloats

Pre

78

LDT

Pre
11

UR

Post

26

26

81

96

q.

Item Group 02. If p thenq. Not p.

Number

LDT LDT

99

27

27

Item Group 01.** If

eub-

immeoning Test ter

symbolic form for each item

The symbol ":" is used to introduce

53

71
94
59
88
82

74

e;
54

49

24

49
39

88

47

81

92

78

00

100
100

83.

83
82
65

18

92
100

12

4o

24
35

31
19

39
19

34

96
992

97

85

85

73

85

85

89

84

group is presented. Variations are

the proposed statement.

FAN3

I.T.T

gistimEotiA

4,11

c%a4,ag:

ogMtc;

Hi

M
Po

ZI

r4 UN el ri

Ps

LIA

ai

r9

s:411

3N

:111 al

cTo

;11

ar

0'
4)4

43

c;.

Cp

a
1

NCO

03

gAR.;*4

["8 fr'tN

81.1Li

t44

Po

f- P`RPMP

`401t117

A.tigiti&A

I cAaptatzfil

4 VI I'

"nP2R)S

xri

I
OD

I-1

ha P. 1-1

M"

41 .9 N

FA I-a 1-1

tna

t40 r%j%

tel

vi Jut-4 4" Al 4r 4
ro
pa

1.8
W

-43 Q

co

ODtaOD Y1

tal

V-St'V

ZA

1.%)28M

ON 0%%,0 f9 ON-4 0%

1 4 ataNcriA%/i

03 'dal 03w
03SMNn

ta3 -mos

,W VeiTra'

118401481

w qr.

M.
Qat

01V1

ChVI

1 1. VI ri ta VI

tr.1

%,ft

41.11)

1-

Na vi

4r

N
1.1

if

n-ic+

:1 18? iMV--11 1

4m

VI

el i4'3.M23&

-t":3

ri

%/I

V:n NN

gv

AN

Sbau 24%12? St(1,`

44 _QUA) CI\ 01D Co

IV vv -4 LA) -4 La VI

8\ 2 g v

1C)

-4 v1-4 co.a 71

WviOvIvir-4 r9 cn oN.o 74
O wr, -4 -4 r0 i-, 0

10 Q

P'1

0
19t

A Z'fil WO era 83.) jt

WI

g .1 if) g tie)

63

e+

11

la).% Rig V O'N

CI\

-4

10

0%n F MovCSi 0

,t-ra sio3 it-4

.00` V ej 1.S` F3 .Z1 at

steal:piss

e ,T,DA-wagg
Z vt .P-01-4 cr.
n; t. A ) .1vi p CNV1

A i ..n o3 co co co
MH-4v11-1

v nes e.,3m0::$v.

83 V.eggi'l IU:411

MO AS

`43 1St ii s a

A
IR 0

Et C113131)

0
1-#

P-1 I-1 f0 I-1

b)

CO t.4 CA 46' 0 14 c)

1:3 ctWat:

:esj

23 tS 23

113 1-+

IV
CO

i 4- CO us

!AV

I%)

P'w

VLa

Pot 8 & 8 Li 164,

TABU:A.-lb Individual Item and Item Group Mean .Difficulty Indicesby Grades on the Class Reasoning Test for All

Subjects on the Pre-test and Also for LDT Pre- and Post-tests.

All*

LDT

Subjects LDT
Pre Pre Post

Grades

o4

94

N =

25

25

All*

Subjects LDT
Pre Pre

103

Item Group 01.** All As are Bs.

06

LDT
Post

25

Subjects LDT
Pre

Pre

08

All*

LDT
Post

27 27
At least some As are not Bs.

25
:

All*

100

All*

Sub-

jects LDT LDT

Pre

Pre Post

10

Sub-

jects LDT WI
Pre Pre Post

22

72

96 100 100

93

75

22

12

24 24

Item
Number
08
16
22

91
94
85

29
35

100

84

88

96

Item Group 02. All As are Bs.

68

19
27

49

Mean
*

LDT's

***

loo

97

95 100
94
88

96
96

95
85

loo

92

92

93

89

95

93

96
85

100

93
89

93

91
92

91

95

loo

95

93

96
91

96

96
96 loo

93 96 96

92

35

28

84

96

64

8o

loo
88

35

6o

72

65

74

86

82 86
80 86
55 86
loo
694
86
5
75
30

68

89
96

94
95

96

99

93

89
78

72 90

71

83
55

82

LNDT- 2's
.** The basic
&L21DT1's &
symbolic form for each itemgroup is presented.
are
given
in
Chapter
IV
.

Variations

96
86

loo

92

97

100

loo

loo

96 96

100 loo

94

loo

95

99

96

100
92

loo

loo

96

96
100

98

All Bs are Cs. : All As are Cs.

96

62

25
56

4o

91

93 loo

92

31

92

85
8o

26

96
92

loo

91

Item
Number
07
76

96
92

99

91

39
Mean

96 100

96

88
100

96
100

76 loo

83

92

89 98

i guy 11,

RA

%%0'

ca

Vi

fV, R Cr:

tr

fil

pl

p.

g
o'S\

:a 8203%8

sa

6"

S Pa 803
..* in ...7 mp, CV MIA
Crl

'4'

UN CIO _Q CrICO

t'n

ATA431A

uN
uN

so

CO

t...-

Cu

UN

t....o.u.c...cncv

so vs 0 g....

V.

ON

co col u1

?4

(%8_,r41g.
a. 1"St8win

13

7 r-I

cr)

W%

UN

Al3

g. IVP8oVfl`..

8 PErtatts2
g
Mt

mnrTABLE-A-2

cont. 04

94 25 25

103

25

16

25

100

84

78

27 27 75 22

24

Item Group 06. All As are Bs. No Cs are Bs. : At least some As are Cs.

24

Item
Number
12

15
21
25
36
42
Mean

51

36

56

80
48
72

79

60
48
64
60
56

32

36

28

34

36

08

61
56
To
53

6o
92
100
80

52
56

86

91

82

85

91

91

93

97

96
7o

86
93

loo

loo

85

59
86

100
100

91

83

84

67
82
88
76
83

52

39

67

63 77 77

22
32

37

31

30

57

67

80

37

81

59

59 56

96
89

84
54 61
79 92 88
Item Group 07. All As are Ba. No Cs are As. : No Cs are Bs.
62

60
91

89

92
92

92

100

85

87

93

Item

Number
44
57
59

611.

70
77

33

22
48
34

32

20
52
40

Mean
34 36
Item Group 08.
48
53

65
68

71
75
Mean

12
48
56

34

45

35

20
50
53

20

24
56
52

48
32
32

60
56

38 41 47

39

52
48

52

36

48

27

79

73

68

88

36

75

79

83

73

63

82

6o

41

68

82

77

77
75

63
67

72

88

77

63

No As are Bs. No Cs are Ba. : At least some As are Cs.

Item

Number

52

25

38
4o

38
46

31
29

37

lio
24

36
4o

36
28

34

3o

52
49
44

41
4o

42

36

64

64
52
56
28

50

42 36 52

57
59
66
44

42

68
54
68
46

50

56 55 73

88
63

TT

67
52

73
59

89

52 54 67

64
76

68

86

77

79

loo

92

86

79

92

96

55

91

73
50

82

68

65 75 88

96

64 84

78
74
54

71
92
67

92
83
88

72 83 90

TABLE A-2 cont.

04

Grades

08

o6

100 27
94 25 25
103 25 25
Item Group 09. All As are Be. : All Non -Bs are also Non-As.
N =

27

75

52

61

7o
59
81

76
60

10

22

22

12

72 24

24

Item

Number

45

33

4252

54
52

24

52

49

44

63
49

72
52
36
64
48

56
52
64
52

56
56
83
48

40
52
52
84
40

43

28

20

48

51

37
69
34
45
49

56
56
44

44
44
32

34

32

55
66
73

47

56

48

61

68
64
71
loo
54

67

24

39 46

44

56

79
32
41
65

54

63

58

81
47

59

57

93
52

59 77
64
82

68
73

96
55

91

50

64
64

Mean
51 49 53
56 54
68 68 73
59 To 65
Item Group 10. All As are Be. All Non-Cs are Non-Bs. : At least some As are not Ca.

68

67 88

76

86

79
83

96

96

65

83

50

83
67

71

58 78

88

83
86

Item

Number
62
67
72

76

Mean

44

4o

64

47

20

44

39

Item Group 11. All As are Bs.

Item

Number

47

61

52

44

54

59

60

3o

60
68
16
44

68
52
24
4o
28

58

63
78

Mean

56

75
33
52

36

46

43

41
87
35
47
46
48

68
76
48
56
56

56

6o
92
28
60

53

59

53

61
78
43
54
86

61

63
74
37
48
63

55

45 50 55
76
87

36
43
69

77

73
18

55
82

64
73
41

50
55

59 59 56

51 50 63
72

96

79

28
42

50
54

63
67

75
64

50

79

88

75

55 63 73

All Cs are Ds. All Bs are Cs. : At least some As are not Ds.

63 76 68

7o 78 85

74

62

75

68

48

56
4o
36
59

79
36
62

80

84
8o
64

68
24

65

75

75

66

82

80

45

66

VA5
86

79

81

85

85
74
85
81

83

81 77 64
93

96

loo

8o

82

68

71
93
57

73
77
41

64
SI.

77

79 74 78

72 67 83
92

loo

88

79

88

88

79

89
58

88

96
92

75

92
83

78 88 85

TABLE
A4
cont.
04
06
Grades

25 25
100 27
103 25 25
Item Group 12. No Bs are Cs. No Ds are Cs. All As are Bs.
N =

94

Item

Number
93

63
37

56
32

61.

47

52

46

56*

1 Z

54

49

48
44

68
48

60
56

56

Mean 50 46 58

63
34

64

36

12
56

67

63

32

52

84
48

64
72

80

53 43 70

* Item 56 turned out to be a defective item.

10

08

71

55

89
68

56

78

73

Tz
TT

64

T8

27 75 22 22 72 24 24

: No Ds are As.

91

85 T9 100

To 57 59 77

69 71 100

85 85

82

78

57

55

78
85

TT

TT

79

84

86

T7

TT
91

72 72 83

59

71 92

68

88 92
96 88

78

72 81 94

Table A-3. Individual Item

and Item Group Mean Disartmlnation Indices

Ay

Grades on the Conditional Reasoning Test foiAll Subjects

on the Pre-Test.
Grades
2116

05

07

09

11

102

99

80

78

26
59

19
30
15
22

56
41

52
56

37

32

Item Group 01..

Item Number
07
14
19

3.3.

30

27

31
40
Mean

23

14
27
27

43
43
10
29
52

23

33

19

35

11

43

26

29

Item Group 02.

Item Number
09
13

-26

23

-15
0

15

-22

22

a5

18
26

34
Mean

-18
9

33
48
10

28

38

- 30

-22

- 30

5
- 9

29
38
43

-30
-18

17

-1

23

22

Item Group 03.

Item Number
11
24
30
32
37
41
Mean

4
16t

0
5

26

15

- 7

-7

4
6

-13

18
4

414

43
29

Item Group 04.

Item Number
08
16
22
29
35
39

Mean

27

24

14

67

36

-14

37

22

33

34

24

25

15
56
19
37
41

41

38

Table A-3 cont.

Grades
21 sz

05
102

07
99

22
33
37
37
30
63
37

30
59
37
30
41
63
43

Item Group 05.

Item Number
45
49
52
566 5

73
Mean

32
18
23

32
32
24

50
56

61
69
74
Mean

48
1
9
48
43
36

Item Group 06.

Item Number
46

57

41
41
37

41
30

67
70
48

22
37
26
30

51

33.

lbp

14
32

o
9
14
14

it
0

-199

29
14

Item Group 07.

Item Number
44
57
59
64
To

77
Mean

-4
0
-19
-22

-7

22

11

12

4
7
9

41
59
33
41
63
48
48

33
33
41
33
33
37
35

-22

9
9

43

27a

62
43
52
71

41
32

23

Item Group 08.

Item Ntaaber
12
15
21
25
36
42
Mean

27
23
23
36

27
55

32

118

53

33
14
5
76
29
29
31

Table A-3 cont.

Oradea
N=

05
102

07

09

99

80

30

52

27

Item Group 09.

Item Number
10
17
20
28
33
38

Mean

37
48

26
33

74

11
52
26
33

52

40

14

3.9

19

36
32

10
10

67

23
22

14
23

Item Group 10.

Item Number
413

56

51

48

48
6

33
33
26

56
3o
41

62
67
72
76

Mean

22

41

36

55

57

55

48

50
36

43
43

27

48

14

38

39

46

Item Group 11.

Item number
47
54
58
60
63
78

Mean

48

52

22
26

10

63

41
41

36
36
32
27
36

52

30

29

23

-u

19
56

44

rlt

76
29
14
5

Item Group 12.

Item Number

-15

48
53

65
68
71

75

11
. 7

Mean

15

. 7

18

18

4
0

0
5

3.5

-22

-5

19

-14
-33
- 5
0

The top and bottom groups fah each were compared were made up off'

27% of this N.

Table A.J.

Individuel Item and Item, Ommip Man Discrimination Indices by Oradea on

Class Reasoning Test ittr All tkibjects on the Pre-Test.

Grades

Ob

Pm
Item Group 01.

94

Item Number
08
16

22
29
35
39
Mean

44
40
52
36

68
8o
53

06
103

08
100

21
18
32

15

75

0
14

19

50

30

10
5
5
5
5

23

21

33
30
59

10
5
30

30
26

10
0
15
5

15
15
10

Item Group 02.

Item Number

07
14
19
27
4o
Mean

1$0

28
32
20
68
8
33

19
52

0
15

30
37

30
15*

25
45
0
30
15
20

Item Group 03.

Item Number
11
24
30
32
37
Mean

36

28

52

24
4
24

32
43
25
43
61
39

48
37
-4

60
4o

52
52
52

43.

4o

36
21
29

15
37
56

25
45
55

45

Item Group 04.

Item lusher
09
13
18
23
26
Mean

441

57

33

24

32

5
40
50
30
10

20
26

30

?g

60

20
25
34

Table AA cont.
Grades

I Is
Item Grow 05.

04
94

06

08

10

l03

100

75

Item Amber
10
17
20
28
33
38
Mean

36
52

16
32

-4

18
14
7

30

4
4

50

3.1.

48
30

44

40
4
80
4o
36
ho
40

-1

16

25
43

14

20
5

25
15

10
10

33
33
15
20

-10

3.3.

23

52
33
30

5
55

25
ho
35

35
10

To
5

Item Group 06.

Item Amber
12
15
23,

25
36
42
Mean

Item Group 07.

43

14
71

56
52
38

7
27

15

-5

15
1.0

16

35
15
26

Item Amber
44
57
59

64
70
77

-12

-16

32
20
7

52

37

-7

71

29
61

43.

44
59
38

39

45

Item Grasp 08.

Item Amber
48

16

65

432

71

4
36
16
28

53

68

75

26

24

k ,

59

U
52

42

44
43

40
60

-15

35
30
ho
32

55
55

-5

70
bo
65
47

ap

V-SPA19

ra

II

Aes`kmv._Ir

g
P1-1

C An3o1623 % 21MVAig '6 V Wis:11-1A1-) r 1:1 & Vt8 it g V.

1 ay. %-a -a4 AV

1 0o

& a 1Z-I 03 itti

g-.3 V-03 V es' P ii-1 % 0 *03 81

a &MS Vg 8: &Vt../ V V X 'urVIS g PS N LS till V 13 4 4 a a

to

a0dag-u, ins t g las .ingu. t; is s SU' 0621%. 0; 2L01 %LAso23a

jat

g IgSgota it :II id Zil tigna g 19 WtooSS 1 a8ivset dr,

TABLE A.5.

Conditional Reasoning Adjusted. Mean Post-Test Comparisons

of Students to %on Logic Was Deliberately Taught and
Students to Whom Logic Was Not Deliberately Taught.,

Grade

LDT
17

9
ENDT-1
23

LET
26

ERDT-1
22

47.3

54.3*

77.6*

62.4

11

LND-1

LDT
24

L1111T-1

11-

LET
27

Total
Score

46.4

45.6

55.4

56.2

CF

24.2

21.5

29.025.7

SY

6.2

5.9

7.4

7.5

6.3

7.0

10.0*

7.8

su

5.6

5.5

6.5

6.2

5.2

6.0

9.2*

6.5

4.6

4.4

5.2

5.1

4.7

5.2

5.7

5.2

1.4

1.8

1.8

1.3

1.4

1.2

5.1

2.0

2.1*

1.0

3.13.4

2.1

1.2

1.7

1.5

4.8

1.7

4.6

3.9

3.9

4.8

3.9

4.8

3.8

4.6

5.2

5.1

4.0

4.1

2.3

3.8

5.1

3.4

1.2

1.8

1.8

2.4

4.7

2.6

5.2

4.6

5.1

4.4

4.6

5.3

5.1

.5.0

5.0

4.7

5.7

4.4

4.7

3.9

5.1

5.5

3.7

4.6

3.1

4.2

4.2

4.4

1.0

1.1

0.6

1.1

3.8

1.2

Group

26

25

Component

28.1*38.0*

31.8

Item
Group

4.03.9

4.o3.9

1.21.3

3.94.0
4.04.5
3.03.7

4.4

4.7

5.5
4.7

2.93.2
1.81.2

An asterisk is' placed by each statistically sienificaatly superior

adjusted nem
Note: Mese adjusted mans are pert of the outcome of an analysis of
covariance in *la IQ and pre-test score were held constant.

TABLE A-6.

Grade
Group

Class Reasoning Adjusted-Mean Poet -Test Comparisons of Students

to Whom Logic Was Deliberately Taught and Students to
Whom Logic Was Not Deliberately Taught.

4
LDT

MDT -1

6
LND-1 WI'

LDT

NM

Total

44.8 47.3

63.8*

Score

10

15

LNLT-1 Wr

LNDT-1

Llif

68.1 65.4

81.4 77.3

36.7 35.1

40.7

7.5* 6.1

12
LIDT-1

87.8 82.6*

55.3

Compo-

nent
CF

26.2 26.8

34.8* 31.2

Sit

5.1

4.9

6.74

su

4.7

5.3

7.0

5.3
.6.2

7.8

41.0

43.7 42.1

9.1

7.9

10.3* 8.8

9.6

9.1

10.5

10.0

5.7

6.p

5.9

5.9

4.8

5.4

5.5

5.8

5.8

3.8

5.2

4.6

5.5

5.3

5.3

4.8

5.7

5.2

5.4

7.3

Item
Group
1

3.9

4.3

4.1

4.0

4.9

2.2

2.4

3.4

2.7

4.2

2.1

2.7

3.5

2.5

3.1 2.8

4.8

3.9

4.1

4.8

4.4

5.2 4.6

5.1

3.8

3.2

4.7

4.4

5.2 4.5

5.2

5.5

5.6

2.0

2.3

2.9

2.3

3.5 2.9

4.4

4.1

44 ce

2.1

2.2

3.1

2.8

4.1

2.8

5.2

4.4

5.5

4.9

3.2

2.9

4.0

3.4

3.7

4.0

4.5

4.o

5.2

4.8

10

2.3

2.2

3.6

2.7

3.2 3.4

3.5

4.5

4.2

4.3

11

2.6

2.6

3.8

3.4

4.7

4.6

4.9

5.4

5.1

5.2

12

4.0

3.8

4.6

3.7

4.9

4.1

5.2

5.5

5.7

5.4

An

asterisk

adjuiteed

is

moo:.

placed

0:1 et '4-'714

5.7

by

each

5.3

5.6

4.6

5.3

*5.7

statistically

3.9
5.2

significantly superior

TABLE A-7.

Pre- and Post-Test Mean Difficurvy Indices for Students to

%an Conditional Reasoning Was Deliberately Taught, by Item
Groups and Grades.

r
Item
Group
1

tirade

Test
Pre

Post

1h

27
66

Change

7
24

77
9

76
83
7

30

23.

Pre

Post
Change

Pre

Post

Change

Post

Change

24

35

35

11

11

49

- 2

56

Pre

Post
Change

Pre

Post

Change
Pre Post
Change
Pre

28
7

24

51

Pre

25

74
6
73

67

80

44

71

65
21
15

- 67

21

20

15
5

Post

71

83

Change

63

- 77

- 8

Pre

Post

Change

70

6
83

72

91

Pre
Post
Change

Post
Change
Pre Post
Change

50

76

48

78

-2

Pre

47

72

48
1

65

31.

30

-1

26

74
78
4

ea

97
9

15

34

24
9

84

29

41

31

68

11

11

17

50

94
53

59
63

71
82
11

68
65

87

58
37

67
86
19

28

53
80

-3
-21

33
5

91

27

71
79

88
80
-8

80

93
80

84

-13

68
62

-6

88
91
3

62

72

49

71

-7

-13

-1

16
13

28
10

18
62
44

-3

-18

TABLE A-8. Pre- and Post-Tests Percentages of LDT's Who Met the Sufficient
Condition for Mastery, and Percentages Who Failed to Meet the
Necessary Condition for Mastery of Principles of Conditional
Reasoning, By Item Group and Grade.

e
e
ng
Necessary Condition
5
11
9
27
26
24
17
04
17
24
08

aye
Sufficient Condition
Item
Group
1

grade

11

26
81

Test
Pre

27
48

24

17

54

53

63
15

59
5

53

Post
Change
Pre

04

Post
Change
Pre

22
18
85

00

00

00

59

77

04

00

0
06

65
08

-26
63

04

17

00

81

96

74

17
59

6
23

73
5o

33

74

-22
21

50
- 9

24

63
-11
41

25
4
21

47
- 6

Post
4

00

0
00

96
15
12

Change
Pre

19

13
-

91

12
-12
100

00
- 8
73

94
6
88

12
-61
81

82

12
-69
23

83
-

8
96

6
53

Pre

19

61

41

77
27
81

44
25
30

70
9
57

47
6
18

83
2
42

33
- 8

21

41

67

Post
Change
Pre

25

41

31

Post
Change

35
-22
00

06

81
39
35

38
13
92

76

-12
06

33
-34
96

94

12
-19
58

77
42
88

93
- 3
41

79
-13
21

82
-12
24

23
-35
00

18
6
12

15
15
00

12
0

08
8

41

08
12
4
23

Post
Change

26

Pre

o4
o
37

09
9
65

12
6

Post
Change
Pre

37
0

48
-17

65
12

Pest
Change
Pre

52

75

71

56

88
13
63

76

81

22

-15

10

Post
Change
Pre

47

81

52

59
4
50

41
- 6
29

89
8

52
0
56

17
0

47
6

11

Post
Change
Pre

29

41

50

19
07

46
4
04

18
-11

12

Post
Change
Pre

0
00

56
0
78

37
8
92

59
18
88

post

04

50

85

96

no

22

19
-

07
26.
-

00
4

06

17

37

-11
96

50

4
22

08
04

4
1.7

4
8

27
4
100
35
-65

4
In the neemarY conditon calms a minus sign In troutt of a change
Change

Note:

77

35

8
12

35

41
11
04

53

15
-

12

indicatea *Fragment, since a minus sign there shove that fever on

the post-test than on the pre-test failed

to meet the necessary condition.

TABLE A-9. Pre- and Post-Test Mean Difficult Indices

for Students to Wks

Class Reasoning Was Deliberately Taught, by

Grades.

Item
Group

25
91

10
22
96

64
-12
68

97
6
74

93
- 2
72

96

98

- 1

84

89

70
2
34

86
12
45

90
18

90

98
9
85

38

56

84.

4
49

11
41

72
21
49

13
61

92
7
77

55

51

Post
Change

37
-12
65

14
69

78

80
19
80

78
1
88

Post
Change

69
4
54

80
11
82

87
9
92

84
4
81

95
7
91

84

88
4
40

85

Post
Change

61
7
36

56

93
2
70

Post
Change

34
- 2
39

47
6
42

59

19
54

75
19
64

77
7
83

34
- 5
49

50
8
54

67
13
70

84
20
68

90
7
'78

53

69
15

65
- 5
61

73

86
8
63

55

56
- 3

Pre

Post
Change
Pre Post
Change
Pre
Post
Change
Pre
Post
Change
Pre

Pre

Pre

Pre

Pre
Post
Change
Pre,

Post
Change
Pre
Post
Change
Pre

Post

4
25
76

24

8
27
95

Test

Grade
N=

Item Groups and

4
47

56

52

41

5
59

99

74

73
10
88

83

78

85

43

77

72

-3

70
27

79

83
11

39
- 8
46

53
- 3
59

43

65
6

46

58
12

-3

Change

6
81

81

TABLE A-10. Pre- and Post-Test Percentages of Students Who Were Deliberately
Taught Class Reasoning Who Met the-Sufficient Condition Fbr
Mastery and Percentages Who Failed to Meet the Necessary Con-

ditim, by Item Group and Grade

Item
Gro
1

Test
Pre
Post
Change

Meeting The
Sufficient Condition
12
6
8
10
4
25
22
24
2
27

Grade
N=

50
-12

100
16

39
16

00

84

100
16

-4

00
O

00

31

20

04

05

35

04
-16

04
O

09
4

68

59

40

22
-37

To

82

Post
Change

58

76
32

89
19

86
4

Pre

04

16

30

Post
Change

04
0
12

44
28

56
26

77
36

84

89

-28

16

15

32

59

62

8o

63

46

08
4
46

28
12

26
11

59

84
25

65

63

88

31

44
-36
20

56

36

27
73

_7

04

18
-28
09

42

64
28
76

81
18
93

82
9
77

96
8
88

19

12

07

09

- 12

0
09

72

93
0

86
9

88
0

16

22

23

37
15

Pre

Post
Change
Pre

4
27

Post
Change

39
12

Pre

00

Post
Change

08
8
04

32

00
4
27

Post
Change

27

10

Pre

11

-4

79

46

8
08
08

04

5o

80

63

46
23

63
13

64
-16

52

26

36

71

20
12
16

44
18
41

68
32
41

79
8
5o

3o
-26
19

48
32

30

- 11

14

79
29

30

12

24

37

32

33

Post
Change
Pre

12

20

19

36

19
-18
67

36
4
46

Post
Change

15
- 4

36
0

74
7

64
18

Pre

Pbst
Change
Pre

16
08

Pre

35

20

67

Post
Change

42
7

64
44

78

u.

55

59

73
14

-u

32

14
-18

24
00
0
04
00

_4
08

04

_4
21
08
- 13
011

00

05

_4

0
36

55

27
-28

12
-25

16

56

- 13

23

13
- 8
46

11
37

-9

59
26
79

55

15

56
19

55

52

15

0
18

75
- 4

67
12

36

15

14

67

52

26

48

64
28

-36

15
4

04

-27
32

59

88
21

12-

44

Pre

Note:

96

42

12

96
0

Pre

Post
Change

89
-11

Failing to Meet The

Necessary Condition
8
10
4
6
2
2
27
22

55

21

25
-21
08

17
9
08

00

-A

00
8

In the necessary condition columns a minus sign in front of s change

indicates improvement, since a aims sign there shows that fever on
the post-test then on the pre-test failed to meet the necessary condition.

TABLE A.U.

Conditional Reasoning Weights and Constants for Multiple

Regression Equations Making Use of Pre-Test.

Grade

CA

7.826

-.4261

IQ

SES

Total Corresponding ConsPre-Test

tent
Sex Pre-Test
Score
Score

Total
Score
Compo-

:7480

18.62

.1179

-1.531

.0385

-.460

-.652

-;o584

.970

6.86

.0189

-.174

.419

.0803

-.127

2.37

o084

-.294

.503

.0614

-.120

2.77

.173

.0290

.011

1.94

-.018 -.0113

.614

1.68

-.472

.133

5.15

.495

nent
CF

2.760

SY

.845

su

.899

-.0372

.254

-.0166

.0155

-.056

1.307

-..0647

.0035

-.0914

1.379

-.0809

- .001

.0014

.0064

-.127

.423

.008

-.0039

.0190

-.082

-.275

.0367

.324

.39

.123

-.0193

0163

-.215

.747

.0761

.005

3.81

.648

-.0215

-.0142

.28o

.0137

.406

.51

.088

.0093

.0202

.082

.0282

.129

-.1424
-.olta

Item
Group

0062

.018

-.097

.0003
.0384

.314

1.26

rvy
clin

..

.271

.0173

.0224

-.097

-.453

.0192

.016

10

.177

-.0018

.0384

-.001

-.249

.0239

.305

11

- .322

.0176

.0123

.006

.638

.0757

alio

- 3.61

12

1.489

-.0954

-.0245

-.254

-.090

.0037

.029

.82

3.49

9.05

TABLE A-12.

Conditional Reasoning Weights and Constants for Nialtiple

Regression Equations Not Using Pre-Test Scores.

Grade

Total
Score

CA

.116

-.5330

.4074

-1.332

.592

10.33

CF

5.327

-.2557

.1770

.457

-.495

11.21

sr

1.216

-.0531

.0443

- .169

.462

1.55

SU

1.211

-.0501

.0280

.280

.483

2.58

.402

-.0192

.0305

-.046

.211

1.38

1.640

-.0905

.0016

-.080

-.489

5.20

1.432

-.0841

.0060

-.059

-.558

5.66

.152

.0100

.0362

-.108

.464

-3.30

.263

-.0074

.0147

-.055

-.083

- .76

.512

-.0302

.0134

-.196

.757

2.96

Lilo

-.0430

-.0132

.069

.126

1.70

.113

.0033

.0360

-.085

.085

-.113

.0123

.0347

-.092

-.465

.75

10

.361

.0020

.0601

.010

-.279

-4.88

11

.056

.0098

.0469

.028

.728

-5.07

12

1.536

-.1003

-.261

-.017

10.72

Component

Item
Group

-.0308

.96

TABLE A-13.

Class Reasoning Weights and Constants for Multiple Regression

Equations Making Use of Pre-Test.

IQ

Total dorresponding Cons-

Sex

Pre-Test
Score

.767

-1.250

.6209

.1555

.353

- .221

.0885

.415

-10.06

.0931

.0230

-.079

.237

.0289

.154

7.09

.0003

.0501

.197

.154

.0860

.104

6.06

.040 -.0009

.0212

.121

- .007

.0114

.668

- 2.21

.337 -.0257

.0254

.035

.025

.0242

.089

1.54

- .220

.0173

.407

.57

.036 -.0000

.396

3.58
1.08

Grade

CA

-.778

.2014

.3467

CF

-.049

.0450

SY

-.678

SES

Pre-Test
Score

tent

Total
Score

-39.46

Compo-

nent

SU

.160

Item
Group
1

2
3

-.112

.0229

-.0011

-.052

-.394

.0540

.0053

-.078

-.086

.0124

.0248

.002

.270

.0095

.170

-.246

.0251

.0286

.086

.275

.0144

.350

- 3.53

-.045

.0077

.0236

-.006

.181

.0304

.373

- 3.01

.057

.0135

.0104

.016

.084 -.0241

.430

.47

-.055

.0089

.0321

.092

.098

.0286

-.007

- 2.73

10

-.277

.0329

.0353

.107

.226

.0019

.297

- 5.31

11

-.092

.0124

.0302

.079

- .123

.0359

.149

- 3.71

.098 -.0051

.0128

- .175

.0294

.144

.77

12

.095

TABLE A-14.

Class Reasoning Weights and Constants for Multiple Regression

Equations Not Using Pre-Test Scores.

Grade

CA

IQ

SES

Sex

Constant

Total
Score

.741

.2405

.6369

.593

-2.221

-50.54

Compo-

nent

CF

.636

.6191

.2835

.310

-.841

-12.43

SY

-.527

.0914

.0423

-.094

-.274

- 7.57

SU

.363

.0095

.0970

.172

-.299

8.17

-.003

5.78

Item
Group
.246
.384

-.057
-.215

-.057

-.0107

.0450

.103

-.0224

.0396

.037

.0335

.0202

.0448

.0074

-.072

.0145

.0316

.169

.0162

.0483

.228

.0049

.0441

.015

.0147

.200
.0181

-.0041

-.134

-.191

- 1.97

.000

.236

.83

.023

- 3.53

2.60

1.04

-.016

-.221

3.56

.011

-.278

- 1.47

.084

-.143

3.24

.125

-.341

6.06

.074

-.134

- 5.21

.082

-.253

.15

.0453

-.013
.0427

-.062

.0127

-.355
.0106

.95

-.006
-.312

.0436
.0547
.0321

Graph A-1 Plotted Points:

Conditional Reasoning Mean

Difficulty Max vs. Percentage

Meeting the Sufficient Condition

100 II
a

90
8o

IP

0
41

10

20

60

30

9.4

I ..,

40

ammr=LLormarawriagamma
50

6o

To

Neon Difficult Index

80

90

100

Grath A-2,

Plotted Points:

Conditional Bessoxing Mean

Difficulty Index vs. Percentage
Fatting to Meet the Neceeeery
Condition

4111

or

30

1111101011.

20

10

10

201H67-4-14-44--14
Difficult( Index

616.--IVIE100

Graph A-3. Plotted Points: Class Reasoning Mean Difficulty Index

vs. Percentage Meeting the Sufficient

Condition.

100
ell

90 1--

6
80

111

I
10

20

30

40

60
Difficulty Indices
50

90

100

Plotted Points:

Graph A-4.

Class Reasoning Wean

Difficult,'
Index vs. Percentage Failing to
Meet the Necessary Condition.

100

4011111.

Wil

tt.
20

10

10

IIIIII I.
eo

3o

4o

93

6o

To

Mean Difficulty Index

11440'

80

90

100

APPENDIX B.

THE CORNELL DEDUCTION TESTS:*

"THE CORNELL CONDITIONAL REASONING TEST, FORM X"

"THE CORNELL CLASS REASONING TEST, FORM X"
TRIAL ANSWER SHEET FOR LOWER ELEMENTARY STUDENTS

The following two tests, which are delicribed in Chapter IV, are
the ones that were used in this study. The copies appearing here are
exactly the same as those that were used.
Each consists of 22 pages with 78 numbered items, the first six
of which are sample items. Answers are given in Chapter IV, Tables IV-1
and IV-2.
The trial answer sheet, which was mentioned in Chapter IV, is
reproduced here in order to show the direction of our thinking in
extending these tests to the lower elementary levels.

Cornell Critical Thinking Test Series

THE CORNETT, CONDITIONAL-REASONING

TEST, FORM X

by
Robert H. Ennis
William L. Gardiner
John Guzzetta
Richard Morrow
Dieter Paulus
Lucille Ringel
MM.

Fill in the blanks when you are asked to do so:

Print your last name only

Do not
write in
this space:

Print your first and middle names

Your age on your last birthday

years

Your date of birth:

month

day

day

year

year

Your grade
Your school
Your regular teacher at this time

Today's date:

month

General directions:

This is a test to

see how well you do a particular kind of thinking.

We

call it

"conditional reasoning".
You will see that you already do some of this kind of
thinking.
The sample questions make clear what is expected.

DO NOT GUESS WILDLY.

There is a

scoring penalty for guessing wrong.

If you

think you have the

answer, but are not sure, mark that answer. But if you have no idea,
then !lathe question.

There are 6 sample questions,

then 72 others. You should work as quickly as
can, but do not rush.
This is not a speed test. Once you do the samples,
you will be able to

you

move right along.

DO MT TURN THE PAGE

0 1964 by R. H. Ennis

UNTIL YOUR EXAMINER TELLS YOU TO DO SO.

Publidhed. by Cornell

Critical Thinking Project,

Stone Hall, Ithaca, N.Y.

Cornell Conditional-Reasoning Test, Form X

Page 2

Answering the questions:

In answering each question, use only that you are told in that question. In order to
do this, you should imagine that your mind is blank, because some of the things
you are told are obviously false. Even so, you should suppose that they are true- for that question only.

You will be given one or more sentences with which to think. You will then be given
another sentence, about which you must decide, using only what you were told.
There are three possible answers.

This is what they mean:

YESIt must be true.

NOIt can't be true.
MAYBE It maybe true or it may not be true. You weren't told
enough to be certain Whether it is "YES" or "NO".
The meaning of the possible answers is given at the top of each page to help you
remember. Each question has only one correct answer.
Mark your answers on this booklet by drawing a circle around the right answer.
Remember: If you have no idea what the answer is, plathe question and go on
to the next. Do not guess wildly, but if you think you know, then answer the
question.

Sample questions:
Read the first question and see how it is marked.

1.

Suppose you know that

Bill is next to Sam.

1.

Then mould this be true?

B.

NO

C. MBE

Sam is next to Bill.

The correct answer is A, "YES". If Bill is next to Samothen Sam must be next to
Bill. It must be true, so a circle is drawn around "rEsn.
Here is another sample.

This time Ea circle the answer.

Suppose you know that

The sparrow is over the hawk.
Then Would this be true?

2.

The hawk is over the sparrow.

le".

You should have circled B,

If the sparrow is over the hawk:
can't be over the sparrow. It can't be true.

A. YES
B. NO
C. MAYBE

then the

hawk

Page 3

Cornell Conditional-Reasoning Test, ibrm X

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.

MAYBE It mey be true or it

may not be true.

You weren't told enough

to be certain whether it is "YES" or "NV.

Circle the answer to this next sample. Be careful:

3.

Suppose you know that

Jane is standing near Betsy.

3.

A. US
B. NO

Then would this be true?

C. MAYBE

Betsy is standing near Jane.

be
The correct answer is C, "MAYBE ".
Even if Jane is standing near Betsy, Betsy might
sitting. Betsy might be standing
or
near Jane, but she might be sitting near Jane,
something else.
You were not told enough to be certain about it, so "MAYBE" is
the answer.

Circle the answer to this next sample

question. Remember that your mind is

to
be
blank
at
the
beginning
of
each
question.
supposed

Suppose you know that

California is near Nev York.
Then would this be true?
Nev York is near California.

h.

YES

A.
B. N)

C. MAYBE

The correct answer is A, "YES",

near
even though New York and California are not really
to each other.
If California were neer to New York, then New York impuld be

neer to California.

It would have to be true.

Remember: You should suppose that that

answering.

you are told is true -- for the question you are

Cornell Conditional- Reasoning Test, Fbrm X

Page 4

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.
MAYBE It maybe true or it may not be true.

You Irv-east told

enough to be certain whether it is "YES" or "BO".

So far in the sample questions you were told only one thing.
told two things. Circle your answer.

In this one you are

Suppose you know that

The pit is inside the mouth of the fox.
The cherry is inside the mouth of the fox.
A.

/ES

Then vould this be true?

NO
The pit is inside the cherry.

MAYBE
The correct answer is C, "MaBE". All
are both in the mouth of the fox.
the cherry or not=

you are told is that the pit and the cherry

There is no way to be certain whether the pit is in

Here is the last sample question.

This time the letters "X" and "Y" are used. They can stand
for anything you like. Circle your answer:

6.

Suppose you know that

6.

X is next to Y.
Then vould this be true?

B.

la

C. MAYBE

I is next to X.
The correct answer is Al "YES",

A. YES

no matter mat X and Y stand for.

If Xis next

to I, then I must be next to X.

Nov that you have done the practice

expected.

questions you probably understand *at is

If you have any questions, ask them now.

Cornell Conditional-Reasoning Test, Pbra X

Page 5

Here is a reminder of the meaning of the possible answers:

YESIt mast be true.
NDIt can't be true.
MAYBE It may be true or it may not be true.

You weren't told

enough to be certain Whether it is "YES" or "NO%

7.

Suppose you know that

If the hat on the table is blue, then it belongs to Joan.
The hat on the table is blue.
Then ig.)uld this be true?

7. A.
B.

The hat on the table belongs to Joan.

8.

YRS
ND

C. MUM

Suppose you know that

If the car in the parking lot is W. Smith's, then

it is blue.
The car in the parking lot is not blue.

Then would this be true?

The car in the parking lot is Mr. Smith's.
9.

8.

A.

YES

B.

NO

C.

MAYBE

Suppose you know that

If Italives in the white house, then his last name

is Smith.
Tan does not live in the white house.

Then would this be true?

Tba's last name is not Smith.
10.

9. A.

YES

B.

NO

C.

MAYBE

Suppose you know that

Harry is or the football team only if he has his

mother's permission.
Harry is on the football team.
Then would this be true?

Harry has his mother's permission.

10. A.

YES

B.

ND

C.

mArm

Cornell Conditional

Page 6

- Reasoning Test, Form Y

Mere is a reminder of the meaning of the possible answers:

YESIt must be true.
MOIt can't be true.

MAYBE It may be true

or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "MO".

Suppose you know that

If Mary lives in the white house, then her last name
is Brown.
Mary's laet name is Brown.

A.

Then would this be true?

YES

93

kbry lives in the white house.

MAYBE

Suppose you know that

John is in the kitchen only if there is food in
the kitchen.

There is no food in the kitchen.

A.

Then would this be true?

YES
B.

Jan is in the kitchen.

N)

MAYBE

Suppose you know that

If the automobile in the parking lot belongs to Mr.
Brown, then it is black.
The automobile in the parking lot doesn't belong

to W. Brown.

A.

Then would this be true?

YES

NO

The automobile isn't black.

MAYBE

Suppose you know that

Joe's bicycle is not working today.
If Joe's bicycle is not
walk to school.

working, then he has to

Then would this be true?

Joe has to walk to school today.

A.

YES
B.

MD

C.

MAIBB

Cornell Conditional-Reasoning

Test,

Fbrm X

Page 7

Here is a reminder of the meaning of the possible answers:

YESIt must be true.
NOIt can't be true.
MAYBE It may be true or it may not be true.

enough to be
15.

You weren't told

certain whether it is "YES" or "NO".

that

Suppose you know

There is an X only if there

is a Y.

There is not a Y.
Then would this be true?

15.

C. MAYBE

There is an X.

16.

A. YES
B. ND

Suppose you know that

Dick was not

at home yesterday afternoon.

the football game

afternoon he was at hose.

If Dick was not at

yesterday

Then would this be true?

Dick vas not at the football game

16. A.
B.
C.

yesterday

YES
NO
MAYBE

afternoon.
17. Suppose you know that

Tom may use paints only if he

has cleaned up his

clay work.
Tom may use paints.

17. A.

Then would this be true?

B.

YES
NO

C.

MAYBE

Tom has

cleaned up

his clay work.

Suppose you know that

Iked went to a movie last night.
If Fred does not go to a movie, he feels bad
the next day.

Then would this be true?

Fred does not feel bad today.

18.

A. YES

B. ND
C. MAYBE

Cornell Conditional-Reasoning Test, Form X

Page 8

Here is a reminder of the meaning of the possible answers:

YESIt must be true.
193It can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

19.

Suppose you know that

If there is an X, then there is a Y.
There is an X.

19.

Then would this be true?

There is a Y.
20.

B.

YES
NO

C.

MAYBE

A.

Suppose you know that

Mary will be in the school play only if she likes

plays. Mary will be in the school play.

A.

Then would this be true?

YES
Z

Mary does not like plays.

MAYBE

Suppose you know that

Its is playing ball only if he has a ball glove.
Ws does not have a ball glove.

A.

Then would this be true?

YES

NO

Tom is playing ball.

MAYBE

Suppose you know that

If there is an X, then there is a Y.
Mere is not a Y.
Then would this be true?
There is an X.

A.

YES
NO

MAYBE

Test, Form X

Cornell Conditional- Reasoning

Here is a reminder of the meaning of

YESIt
BO

Page 9

the possible answers:

must be true.

It can't be true.

MAYBE It may be

true or it may not be true. You weren't told

to be certain whether it is "YES" or "MO".

enough

23.

Suppose you know that

If whales are birds, then they
Whales aren't birds.
Then would this be true?

can

fly.

23. A.
B.

Whales can't fly.

24.

Suppose you know

C.

YES
BO
MAYBE

that

If Bill lives on a farm, then

he has a pet dog.

a pet dog.
Then would this be true?

24. A.

Bill has

Bill lives on a farm.

YES

B.

NO

C.

MAYBE

25. Suppose you know that

Jerry was not

Jerry is not

asked to

play ball.

home only if

Then would this be

he was asked to play ball.

true?

Jerry is not home.

26.

25. A.

YES

B.

NO

C.

MAYBE

Suppose you know that

If Mary lives in the

green house,

is Jones.

Mary doesn't live in the

Then would this be true?

green

then her last name

house.

Mary's last name is not Jones.

26. A. YES
B. ND
C. MAYBE

Cornell Conditional-Reasoning Test, Form X

Page 10

Here is a reminder of the

YES

meaning of the possible answers:

It must be true.

IX1

It can't be true.
MAYBE It mey be true

or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

Suppose you know that

If the coat in the closet

is brown, then it belongs to

Sue.

The coat in the closet is brown.

A.

Then would this be true?

YES
?X)

The coat in the closet

MAYBE

does not belong to Sue.

Suppose you know that

There are black cats
There are black cats.

only if there are pink cats.

A.

Then would this be true?

YES
NO

There are pink cats.

MAYBE

Suppose you know that

If the bicycle in the

garage is Bob's, then it is red. The

bicycle in the garage is not red.

Then would this be true?

A.

The bicycle in the

garage is not Bob's.

NO

Suppose you know that

If there is an X,

YES

MAYBE
then there is a Y.

There is a Y.

Then would this be true?

There is an X.

A.

YES
B.

N)

C.

MAYBE.

Cornell Conditional- Reasoning Test, librin X

Here is a reminder of the meaning of the possible answers:

Page 11

YESIt must be true.

BOIt can't be true.
MAYBE It maybe true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "MO".

31.

Suppose you know that

If mice have five legs, then they run faster than
horses.
Mice do have five legs.

Then would this be true?

Mice run faster than horses.
32.

B.

YES
D3

C.

MAYBE

31. A.

Suppose you know that

If Jane fell off her horse, then she hurt herself

badly.
Jane hurt herself badly.

Then would this be true?

Jane fell off her
33.

Suppose you

horse.

B.

YES
NO

C.

MAYBE

32. A.

know that

The short pencil is not Bill's favorite pencil.

The short pencil is not Bill's favorite, only if
it is dull.
Then would this be true?
The short pencil is dull.

B.

YES
ND

C.

MAYBE

33. A.

34. Suppose you know that

If there is an X, then there is a Y.
There is not an X.

Then would this be true?

There is not a Y.

34. A.

YES

B. ND
C. MAYBE

Cornell Conditional-Reasoning Test, Form X

Page 12

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.
C.

MAYBE It may be true or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

Suppose you knor, that

If John lives in the white house, then his last name

is Smith.
John's last name is not Smith.

35.

m"-- would this be true?

John does live in the white house.

B.

YES
MO

C.

MAYBE

A.

..nommiorr

Suppose you know that

Birds can fly only if they can play the piano.
Birds cannot play the piano.

A.

Then would this be true?

BO

Birds can fly.

37

YES

MAYBE

Suppose you know that

The car will start.
If the temperature is not below freezing, the
car will start.

A.

Then would this be true?

YES
NO

The temperature is not below freezing.

MAYBE

Suppose you know that

There is an X only if there is a Y.
There is an X.
Then would this be true?
There is a Y.

A.

YES

NO
MAYBE

Cornell Conditional-Reasoning Test, Form X

Page 13

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.

C.

39

MAYBE It maybe true or it may not be true. You weren't told

enough to be certain whether it is "YES" or "NO".

Suppose you know that

If dogs have four legs, then they have three

eyes. Dogs don't have three eyes.

A.

Then would this be true?

YES
NO

Dogs do have four legs.

MAYBE

Suppose you know that

If Jean goes to the park, she will see her friend
Pat. Today, Jean is going to the park.

A.

Then would this be true?

YES
NO

Today, Jean will see her friend Pat.

MAYBE

Suppose you know that

If horses are green, then they have two tails.
Horses have two tails.
Then would this be true?

41.

Horses are green.

42.

B.

YES
NO

C.

MAYBE

A.

YES

B.

NO

C.

MAYBE

A.

Suppose you know that

The red pencils belong to Sally only if they are on

the table.
The red pencils are not on the table.

42;

Then would this be true?

The red pencils do not belong to

Sally.

Cornell Conditional-Reasoning Test, Form X

Page 14

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NDIt can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

43.

Suppose you know that

If Paul rides his bike to school, he goes the long way.
Paul rode his bike to school today.
If Paul goes the long way, he gets to school late.
Then would this be true?
Paul was not late for school today.

44.

C.

YES
IL)

MAYBE

Suppose you know that

If the chair is green, then the table is black.

Then would this be true?
If the table is black, then the chair is green.
45.

43. A.
B.

44. A.

YES

B.

ND

C.

MAYBE

Suppose you know that

If there is a blue pencil in the second box, then there

is a green pencil in the first box.
If there is a green pencil in the first box, then
there is a red pencil in the third box.
Then would this be true?
If there is a blue pencil in the second box, then

46.

45. A.
B.

YES
N3

C.

MAYBE

there is a red pencil in the third box.

Suppose you know that

If Mrs. Smith entered the flower show, then she entered

her roses.
Then would this be true?
If Mrs. Smith didn't enter her roses, then she
didn't enter the flower show.

46. A.

B.

YES
NO

C.

MAYBE

Cornell Conditional-Reasoning Test, / lrm X

Page 15

Here is a reminder of the meaning of the possible answers:

YESIt must be true.
NOIt can't be true.
MAYBE It may be true or it may not be true.
You weren't told
enough to be certain whether it is "YES" or "NO".
Suppose you know that
Bill will see Audrey, if and only if he goes to Montreal.
Bill will not see Audrey this year.

47.

Then would this be true?

A.

YES

NO

Bill is going to Montreal this year.

MAYBE

Suppose you know that

If Gary sees Sharon, he goes to
Canada. This winter Gary saw Sharon.
Gary goes skating only if he goes to Canada.

48.

Then would this be true?

This winter Gary went skating.
49.

A.

YES

B.
C.

NO
MAYBE

A.
B.

YES
NO

C.

MAYBE

A.

YES

B.

NO

C.

MAYBE

Suppose you know that

If there is an A, then there is a B.

If there is a B, then there is a C.
Then would this be true?

49.

If there is an Al then there is a C.

50.

Suppose you know that

If birds can fly, then they have six legs.

Then would this be true?
If birds don't have six legs, then they can't fly.

50.

Cornell ConditionalReasoning Test, Form X

Page 16

Here is a reminder of the meaning of the possible answers:

YESIt must be true.
NOIt can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

51.

Suppose you know that

If the bus goes to town, then it passes the old stone
Church.
The bus goes to town.
If it passes the old stone church, then it goes over
the new bridge.
Then would this be true?

51.

The bus doesn't go over the new bridge.

52.

A.
B.

YES
NO

C.

MAYBE

A.
B.

YES
NO

C.

MAYBE

Suppose you know that

If the school team loses this game, Brighton High will

win the league pennant.
If Joe does not hit a homer on this pitch, the school
team will lose this game.
Then would this be true?
If Joe does not hit a homer on this pitch, Brighton

53.

52.

High will win the league pennant.

Suppose you know that

If Jean goes shopping, she goes to Chicago.

Last Saturday Jean went shopping.
Jean visits her aunt only if she goes to Chicago.
Then would this be true?
Last Saturday Jean visited her aunt.

54.

53. A.

YES

B.

NO

C.

MAYBE

Suppose you know that

Tbm will go skating, if and only if he can borrow

Fisank's jacket.

Tom is not going skating.

Then would this be true?
Tom can borrow Ftank's jacket.

54. A.

YES

B.

N)

C.

MAYBE

Cornell Conditional-Reasoning Test, Form X

Page 17

Here is a reminder of the meaning of the possible answers:

YESIt must be true.
NOIt can't be true.
MAYBE It may be true or it may not be true.
You weren't told
enough to be certain whether it is "YES" or "NO".

55.

Suppose you know that

If Sam misses the bus, he will walk to school.
If Sam walks to school, he will cross the bridge.
Then would this be true?
If Sam misses the bus, he will cross the bridge.

56.

B.

YES
NO

C.

MAYBE

55. A.

Suppose you know that

If Bob did not buy a new baseball glove, then

he played basketball today.
Then would this be true?
If Bob did not play basketball today, then he

57.

B.

YES
NO

C.

MAYBE

56. A.

did buy a new baseball glove.

Suppose you know that

If Bill has an apple in his lunchbox, then Sally has a

cracker in her lunchbox.
Then would this be true?

If Sally has a cracker in her lunchbox, then Bill

58.

57. A.

B.

YES
NO

C.

MAYBE

has an apple in his lunchbox.

Suppose you know that

Betty is going to the movies.

Betty is not going to the movies, if and only
if Ann is going to the movies.
Then would this be true?
Ann is going to the movies.

B.

YES
N)

C.

MAYBE

58. A.

Cornell ConditionalPeasoning Test, ibrm X

Page 18

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.
MAYBE It may be true or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

Suppose you know that

If there is an X, then there is a Y.
Then would this be true?

59.

If there is a Y, then there is an X.

60.

B.

YES
NO

C.

MAYBE

A.

Suppose you know that

Elephants are pink, if and only if they are large.

Elephants are not pink.
Then would this be true?

60. A.

Elephants are large.

61.

YES

B.

11)

C.

MAYBE

Suppose you know that

If there is an X, then there is a Y.

A.

Then would this be true?

YES
NO

If there is not a Y, then there is not an X.

MAYBE

Suppose you know that

If John has the red chalk, then he is making a poster
for the play.
John has the red chalk.
If John is making a poster for the play, then he is
in the library.

Then would this be true?

A.

YES
NO

John is in the library.

MAYBE

Cornell Conditional-Reasoning Test, it= X

Page 19

Here is a reminder of the meaning of the possible answers:

YRSIt moist be true.
W)It can't be true.
C.

63.

MAYBE It may be true or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

Suppose you know that

that bicycle belongs to John, if and only if it is red.
That bicycle does not belong to John.
Then would this be

true?

That bicycle is not red.

64.

B.

YES
BO

C.

MAYBE

63. A.

Suppose you know that

If a dog can stand on its front legs, then it is a

PuPPY-

Then would this be true?

If a dog is a puppy, then it can stand on its
front legs.

65.

Suppose you

B.

YES
BO

C.

MAYBE

64. A.

know that

If there is an X, then there is a Y.

There is an X.
There is a Z only if there is a Y.

A.

Then would this be true?

YES

BO

There is a Z.

MAYBE

Suppose you know that

If Kate is in Mrs. Jones' class, then she is out on the
playground.
If Kate is out on the playground, then she is jumping
rope.

Then would this be true?

If Kite is in Mrs. Jones' class, then she is

A.

YlS
B.
C.

jumping rope.

NO
MAYBE

Cornell Conditional- Beaconing Test, Jbrm X

Page 20

Here is a reminder of the possible answers:

YESIt must be true.

NOIt can't be true.
MAYBE It may be true or it may not be true.
You weren't told
enough to be certain whether it is "YES" or "NO".

67.

Suppose you know that

If there is an X, then there is a Y.
There is an X.
If there is a Y, then there is a Z.
Men would this be true?
There is not a Z.

68.

B.

YES
13

C.

MAYBE

A.
B.

YES
NO

C.

MAYBE

A.

YES

B.

NO

C.

MAYBE

A.

YES

B.

NO

C.

MAYBE

67. A.

Suppose you know that

If Jane did not go to the movies yesterday, then

she savrher friend Pat.
Jane vent to the park yesterday only if she saw
her friend Pat.
Jane did not go to the movies yesterday.
Then would this be true?

68.

Jane vent to the park yesterday.

69.

Suppose you know that

If Nancy bought a new dress, then she vent to

the shop on Main Street.

69.

Then would this be true?

If Nancy didn't go to the shop on Main Street, then

70.

she didn't buy a new dress.

Suppose you know that

If John is not in schoe,then he has a cold.

Then would this be true?
If John has a cold, then he is not in school.

70.

Cornell Conditional-Reasoning Test, Form X

Page 21

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NOIt can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

71.

Suppose you know that

If Sally is writing a report at home, then the library is
closed.
Sally is writing a report at home.
Dick is using the classroom dictionary only if the
library is closed.
Then would this be true?
Dick is using the classroom dictionary.

72.

B.

YES
NO

C.

MAYBE

71. A.

Suppose you know that

If there are no blue pencils in the first box, then
there is a green pencil in the second. box.
If there is a green pencil in the second box, then
there is a red pencil in the third box.
There are no blue pencils in the first box.

72. A.

Then would this be true?

There are no red pencils in the third box.

73.

YES

B.

1134

C.

MAYBE

Suppose you know that

If an animal is a turtle, then it can fly.

If an animal can fly, then it has feathers.

A.

Then would this be true?

If an animal is a turtle, then it has feathers.

YES
B.

NO

C.

MAYBE

Suppose you know that

If there is a yellow marble in the first box,
then there is a blue marble in the second box.

Then would this be true?

If there is not a blue marble in the second box,
then there is not Lyellow meible in the first

box.

A.

YES
NO
MAYBE

Cornell Conditional-Reasoning Test, Form X

Page 22

Here is a reminder of the meaning of the possible answers:

YES It must be true. B. NO

It can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NV.

75.

Suppose you know that

If people have fins, then they live in water.
People have fins.
People can swim only if they live in water.
Then would this be true?

75.

People can swim.

76.

A.
B.

YES
NO

C.

MAYBE

Suppose you know that

If this animal is a dog, then it can fly.

This animal is a dog.
If an animal can fly, then it has feathers.

76. A.

Then would this be true?

This animal does not have feathers.

YES

B.

NO

C.

MAYBE

4111.1111111110fillMIIDI

Suppose you know that

If John is on the volleyball team then he is good at
volleyball.
A.

YES

Then would this be true?

NO
If John is good at volleyball, then he is on the
volleyball team.

MAYBE

Suppose you know that

There is a Y, if and only if there is an X.
There is not a Y.
Then would this be true?
There is an X.

A.

YES
NO

MAYBE

END OF TEST.

GO BACK AND CHECK YOUR ANSWERS.

Cornell Critical Thinking Test Series

THE CORNELL CLASS-REASONING TEST, FORM X

by
Robert H. Ennis
William L. Gardiner
Richard Morrow
Dieter Paulus
Lucille Ringel

Fill in the blanks when you are asked to do so:

Print your last name only

DO not

write in
this space:

Print your first and middle names

Your age on your last birthday

years
.1111111111.

Your date of birth:

month

day

day

year

year

Your grade
Your school
Your regular teacher at this time
month

Today's date:

General directions:

This is a test to see how well

you do a particular kind of thinking.

call

it "class reasoning".
thinking.

You will see that you already do some of this kind of

The sample questions make clear what is expected.

DO NOT GUESS WILDLY.

There is a vcoring penalty for guessing wrong.

If you think you have the

answer, but are not sure, mark that answer. But if you have
no idea, then skip the question.

There are 6 sample questions,

as
then 72 others. You should work as quickly
you can, but do not rush.
This is not a speed test. Once you do the samples, you
will be able to move right along.

DO NOT TURN THE PAGE UNTIL YOUR EXAMINER TELLS YOU

TO DO SO

1964 by R. H. Ennis

We

Published by Cornell Critical

Thinking Project, Stone Hall,

Ithaca, N.Y.

The Cornell Class Reasoning Test, Ram X

Page 2

Answering the questions:

In answering each question, use only what you are told in that question. In
order to do this, you should imagine that your mind is blank, because some of the
things you are told are obviously false. Even so, you should suppose that they are
true--for that question only.
You will be given one or more sentences with which to think. You will then be
given another sentence, about which you must decide, using only what you were
told.

There are three possible answers. This is what they mean:

YESIt must be true.

NOIt can't be true.
MAYBE It maybe true or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

The meaning of the possible answers is given at the top of each page to
you remember.
Each question has only one correct answer.
help
Mark your answers on this booklet by drawing a circle around the right answer.
Remember: If you have no idea what the
to
answer is, skip the question and go on
the next.
Do not guess wildly, but if you think you know, then answer the

question.

Sample questions:

Read the first question and see how it is marked.

1.

Suppose you know that

Bill is next to Sam.

A.

Then would this be true?

NO

Sam is next to Bill.

MAYBE

'01

The correct answer is A, "YES".

If Bill is next to Sam, then Sam must be
to Bill.
It must be true, so a circle is drawn around "YES ".
next
Here is another sample.
This time you circle the answer.
2.

Suppose you know that

.0.1.11=111.

The sparrow is over the hawk.

Then would this be true?
The hawk is over the sparrow.

2.

A.

YES

B.

NO

C.

MAYBE

You should have circled B, "NO".

If the sparrow is over the hawk,
hawk can't be over the
sparrow. It can't be true.
then the

The Cornell Class-Reasoning Test, Porm X

Page

Here is a reminder of the meaning of the possible answers:

YESIt must be true.

NO It can't be true.
MAYBE It may be true or it may not be true.
You weren't told
enough to be certain whether it is "YES" or "NO".

Circle the answer to this next sample. Be careful:

V./gam/ma'am

Suppose you know that

Jane is standing near Betsy.
Then would this be true?

3.

A. YES
B. NO
C. MAYBE

Betsy is standing near Jane.

The correct answer is C, "MAYBE". Even if :lane is standing near Betsy,

Betsy might be sitting.

Betsy might be standing near Jane, but she might be sitting

near Jane, or something else. You were not told enough to be certain about it, so
"MAYBE" is the answer.

Circle the answer to this next sample question. Remember that your mind is
supposed to be blank at the beginning of each question.

4.

Suppose you know that

California is near New York.
Then would this be true?

4.

New York is near California.

A. YES
B. NO
C. MAYBE

The correct answer is A, "YES",

really near to each
even though New York and California are not
other.
If California were near to New York, then New York would be near to California. It would
have to be true.

Remember:

You should suppose that that you are told is true -- for the
question you are answering.

The Cornell Class-Reasoning Test, Form X

Here is a reminder of the meaning of the possible

Page 4

answers:

YESIt must be true.

NOIt can't be true.
MAYBE It may be true or it may not be true.
You weren't told
enough to be certain whether it is "YES" or "NW.

So far in the sample questions you were only told one thing.

are told two things.

In this one you

Circle your answer.

Suppose you know that

The pit is inside the mouth of the fox.
The cherry is inside the mouth of the fox.

A.

Then would this be true?

YES
B.

The pit is inside the cherry.

C.

BO
MAYBE

The correct answer is C, "MAYBE ".

All you are told is that the pit and the cherry are
both in the mouth of the fox. There is no way to be certain tether the pit is in the
cherry or not.

Here is the last sample question.

They can stand for anything This time the letters "X" and mr" are
used.

you like. Circle your answer:

Suppose you know that

6.

X is next to Y.
Then would this be true?
Y is next to X.

The correct answer is A, "YES", no matter what X and Y stand for.

A. YES
B.

NO

C.

MAYBE

If X is

next to Y, then Y must be next to X.

Now that you have done the practice
expected.

questions you probably understand what is

If you have any questions, ask them nov.

DO 1)T TURN THE PAGE UBTIL YOU ARE

IDIDTODO SO.

The Cornell Class - Reasoning Test, Form X

Page 5

Here is a reminder of the meaning of the possible answers.

YESIt must be true.

NOIt can't be true.
C.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

T.

Suppose you know that

All the cars in the garage are Mr. Smith's.

All W. Smith's cars are Fbrds.

Then would this be true?
All of the care in the garage are Fbrds.
8.

C.

MAYBE

Suppose you know that

All John's pencils are blue.

YES

8. A.

Then would this be true?

At least some of John's pencils are not blue.
9.

B.

YES
NO

T. A.

B.

760

C.

MAYBE

Suppose you know that

All the books about sailing are Bill's.

All the green books are Bill's.

A.

YES

Then wild this be true?

B.

BO

C. MADE

At least same of the green books are about sailing.

Suppose you know that

None of Jane's dolls have hats.
Then would this be true?
None of the dolls that have hats are Jane's.

10.

YES
NO

C.

MAYBE

The Cornell Class-Reasoning Test, Form X

111110111111111100,..0111111

1.111111111111111

Page 6
41111111111101111111111111111011110111.1.1111011111.16,

Here is a reminder of the meaning of the possible answers.

YESIt must be true.
NOIt can't be true.
MAYBE It maybe true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

11.

Suppose you know that

All the red books are John's.
Then vould this be true?
All John's books are red.

12.

11. A.
B.
C.

YES
NO
MAYBE

Suppose you know that

All of Mary's books are about horses.

None of the books on the shelf are about horses.
Then would this be true?
At least some of Mary's books are on the shelf.
13.

B.

YES
NO

C.

MAYBE

12. A.

Suppose you know that

All Jean's pencils are red.

All the pencils on the table are red.
Then would this be true?
At least some of the pencils on the table are

13. A.
B.
C.

YES
NO
MAYBE

Jean's.

Suppose you know that

At least some of the children in the Martin family
take out books from the library.
All people who take out books from the library
14. A.
have library cards.
B.
Then would this be true?
At least some of the children in the Martin
family have library cards.

C.

/ES
NO
MAYBE

The Cornell Class- Reasoning Test, Form X

Here is a reminder of the

Page

meaning of the possible answers.

YES It must be true. B. NO

It can't be true.

C., MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

15.

Suppose you know that

.All X's are Y's.
No Z's are Y's.

Then would this be true?

At least some X's are Z's.
16.

Then would

this be

true?

None of Fred's pencils are green.

Suppose you know

MAYBE

16. A.

YES

B.

BO

C.

MAYBE

that

None of Sue's books are about animals.

Then would this be true?
None of the books about animals are Sue's.
18.

C.

Suppose you know that

At least some of Fred's pencils are green.

17.

B.

YES
NO

15. A.

I17. A.

YES

B.

NO

C.

MAYBE

Suppose you know that

At least some of Kate's pencils are blue.

All the pencils in the box are blue.
Then would this be true?
At least some of Kate's pencils are in the box.

B.

YES
NO

C.

MAYBE

18. A.

The Cornell Class-Itylsoning Test, Fora

Page

0101.101111
=11111111111111110111110M111111111,

Here is a reminder of the meaning of the possible answers.

YESIt must be true.
NOIt can't be true.
MAYBE It may be true or it may not be true.

You weren't told.

enough to be certain whether it is "YES" or "NO".

.011.111101111111111,

Suppose you know that

All Z's are Vs.

All Y's are X's.

A.

Then would this be true?

YES

BO

All Z's are X's.

MAYBE

Suppose you know that

None of the fifth grade boys are on the football
team.
John is a fifth grade boy.
A.

Men would this be true?

YES

BO

John is not on the football team.

MAYBE
Suppose you know that

AU the members of the

Boston.

school band have been in

No one in Frank's class has been in Boston.

'Ben would this be true?

A.

At least

NO

some ambers of the school band

are in Prank's class.

2?.

YES

MAYBE

Suppose you know that

All X's are Vs.

Men would this be true?

At least

sone X's are not Vs.

A.

YES

ND

MADE

Page

The Cornell ClassReasoning Test, brm X

Here is a reminder of the meaning of the possible answers.

YESIt must be true.
NOIt can't be true.
MAYBE It may be true or it may not be true.

You weren't told

enough to be certain Whether it is "YES" or "NO".

23.

Suppose you know that

All boys are painters.
All children are painters.
Then would this be true?

23.

At least some children are boys.

24.

A.
B.

YES
NO

C.

MAYBE

A.
B.

YES
NO

C.

MAYBE

A.
B.

YES
NO

C.

MAYBE

Suppose you know that

All the second grade children are out on the

playground.
%hen would this be true?

24.

All the children out on the playground are

25.

in the second grade.

Suppose you know that

At least some of the books on the table are

about stars.
None of Bob's books are about stars.
Then would this be true?

All of the books on the table are Bob's.

26.

25.

Suppose you know that

All the boys in John's class are football players.

Fred is a football player.
Then would this be true?

Fred is not in John's class.

26. A.
B.

C.

YES
BO

MAYBE

Page 10

The Cornell Class-.Reasoning Test, Form X

IIMI11.1111=.1.

Here is a reminder of the meaning of the possible

answers.

YESIt must be true.

NOIt can't be true.
C.
MAYBE It maybe true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

27.

Suppose you know that

All the pets of the Greens' won some prize in the
pet show.
Fido is one of the Greens' pets.

27. A.

Then would this be true?

Fido won a prize in the pet show.
28.

NO

C.

MAYBE

28. A.

Then would this be true?

No dogs are animals.

YIPZ

B.

NO

C.

MAYBE

Suppose you know that

Eileen is one of the children on the playground
Then would this be true?

30

B.

Suppose ycu know that

No animals are clop.

29.

YES

Eileen is not one of the children on the

playground.

29.

A.

YES

B.

AO

C.

MAYBE

Suppose you know that

All X's are Y's.
Then would this be true?
All Y's are X's.

30. A.

YES

B.

NO

C.

MAYBE

The Cornell Class-Reasoning

Page 11

Test, Ibrm X

Here is a reminder of the meaning

of the possible answers.

YESIt must be true.

BOIt can't be true.
MAYBE It maybe true or it may not be true.

You weren't told

enough to be certain whether it is

31.

"YES" or "non.

Suppose you know that

All oats can fly.

31.

All animals that can fly are black.

Then would this be true?
All cats are black.
32.

A.

YES

B.

NO

C.

MAYBE

Suppose you know that

Yaw

All the things in the trunk are Bill's.

The brown baseball bat is Bill's.
Then would thistle true?

YES
32. A.
B.

NO

MAYBE

The brown baseball bat is in the trunk.

Suppose you know that

None of Bob's books are on the table, but there
are books on the table.

A.

YES
B. NO C..

Then would this be true.

MAYBE

At least some of the books on the table are

not Bob's.

11111!,
Suppose you know that
All X's are Y's.
All Z's are Y's.
Then would this be true?
At least some Z's are X's.

34.

YES

A.
B. NO

C. MAYBE

Page 12
The Cornell Class-Reasoning

Test, Form X

Here is a reminder of the meaning of the possible

answers.

YESIt must be true.

NOIt

can't be true.

You weren't told

MAYBE It may be true or it may not be true.
"NO ".
enough to be certain whether it is "YES" or

35. Suppose you know that

All Mary's pencils are

yellow.

35.

Then would this be true?

At least some of Mary's pencils are not yellow.

A.

YES

B.

NO

C.

MAYBE

Suppose you know that

All pencils are heavy.
Nothing made of wood is heavy.
Then would this be true?
At least some pencils are
37.

36.

C.

made of wood.

MAYBE

Suppose you know that

At least some of the green

pencils are Dick's.

37.

A. YES
B. NO

Then would this be true?

C.

All Dick's pencils are green.

38.

A. YES
B. NO

MAYBE

Suppose you know that

No X's are Y's.

Then would this be true?

No Y's are X's.

38.

A. YES
B. NO
C. MAYBE

The Cornell Class-Reasoning Test, Form X

Page 13

Here is a reminder of the meaning of the possible answers.

YESIt must be true.
NOIt can't be true
MAYBE It maybe true or it may not be true. You weren't told
enough to be certain Whether it is "YES" or "NO".

39.

Suppose you know that

All dogs are brown.
Then would this be true?
At least some dogs are not brown.

4o.

39. A.
B.
C.

YES
NO
MAYBE

Suppose you know that

All the cookies Jane made for the fair had nuts
in them.
All the cookies with nuts in them were sold.
Then would this be true.

All the cookies Jane made for the fair were

41.

YES
NO

C. HAM

sold.
Suppose you know that

All brown animals have four legs.

Then would this be true?
All animals with four legs are brown.
42.

40. A.
B.

41. A.
B.

YES
NO

C. MAYBE

Suppose you know that

All members of the football team weigh over

150 pounds.
Henry does not weigh over 150 pounds.
Then would. this be true?

Henry is on the football team.

42. A.
B.

YES
NO

C.

MAYBE

The Cornell Class-Reasoning Test, Dorm X

Page lei

Here is a reminder of the meaning of the possible answers.

YESIt must be true.

NOIt can't be true.
C. MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "no".

43.

Suppose you know that

All of John's candy is in the box.
All of the candy that is not chocolate is also
not in the box.
Then would this be true?

43.

At least some of John's candy is not chocolate.

44.

A.
B.

YES
NO

C.

MAYBE

A.
B.

YES
NO

C.

MAYBE

A.

YES

B.

NO

C.

MAYBE

Suppose you know that

All the papers in the box are torn.

None of John's papers are in the box.
Then would, this be true?

44.

None of John's papers are torn.

45.

Suppose you know that

All of the boys are singing.

Then would this be true?
All of the people who are not singing are

45.

also not boys.

46.

Suppose you know that

All the math homework is due today.

None of Joan's homework is due today.
All the homework for Mr. Miller's class is
math homework.
Then would this be true?
None of Joan's homework is for W.
class.

46. A.
B.

YES
1)

C.

MAYBE

The Cornell Class-Reasoning Test, Form X

Page 15

Here is a reminder of the

YES

meaning of the possible answers.

It must be true.

NO

It can't be true.
MAYBE It maybe true

or it may not be true.

You weren't told

enough to be certain whether it is "TES" or "N3".

Suppose you know that

All the pencils in

the box are green.

All Sue's pencils are sharp.

All the green pencils are Sue's.

A.

Then would this be true?

At least some of the

YES

NO
pencils in the box are

MAYBE

not sharp.

Suppose you know that

None of n

shirts are wool.

None of the shirts

are wool.

hanging up in the closet

A.

Then would this be true?

NO

At least

some of my shirts are hanging

in the closet.

49.

YES

up
MAYBE

Suppose you know that

All X's are Y's.
Then would this be true?

All things that

A.

YES
B.

are not Y's are also not X's.

C.

Suppose you know that

NO

MAYBE

All four- legged animals can fly.

No horses can fly.

All fast

runners are four-legged animals.

Then would this be true?

No horses

are fast runners.

A.

YES

NO
MAYBE

The Cornell Class-Reasoning Test, Form X

Page 16

41111
Here is a reminder of the meaning of the possible answers.

YESIt must be true.

NOIt can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

51.

Suppose you know that

All of the boys in the class collect stamps.
All students who are not members of the Stamp
Club also do not collect stamps.
Then would this be true?

51. A.
B.

At least some of the boys in the class are not

52.

C.

YES
N3
MAYBE

members of the Stamp Club.

Suppose you know that

All of the boys are running, but not everyone

is running.
Then would this be true?

53.

52. A.

At least some of the people not running are

not boys.
Suppose you know that

B.

YES
N3

C.

MAYBE

Bone of Tam's books are on the shelf.

No science books are on the shelf.
Then vould this be true?

53. A.

At least some of Tom's books are science books.

54.

B.

YES
NO

C.

MAYBE

A.
B.

YES
ND

C.

MAYBE

Suppose you know that

All of Bill's five uncles are allowed to drive.

An people who have a license have passed a

driving test.
All people who are allowed to drive have a
license.
Then would this be true?
At least one of Bill's uncles has not passed a
driving test.

54.

The Cornell Class- Reasoning Test, ibrm X

Page 17

Here is a reminder of the meaning of the possible answers.

YESIt must be true.

NOIt can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

55.

Suppose you know that

All of the band members are working.

55.

A. YES
B. NO

Then would this be true?

C. MAYBE

Everyone who is not working is also not in the

band.
Suppose you know that

All the books on the shelf belong to the

library. No science books belong to the library.
At least some of the books that Elmer likes
are on the shelf.

A.

Then would this be true?

YES

NO

At least some of the books that Elmer likes

are not science fiction.

MAYBE

Suppose you know that

All the people Who live on Main Street were born
in Milltown.
None of the students in Room 352 live on
Mein Street.

57 A.

Men would this be true?

NO

None of the students in Boom 352 were born

in Milltown.

58.

YES

MAYBE

Suppose you know that

At least some of Mr. Jones' students ride the bus to
school. All students who live on Route 55 own dogs.
All students who ride the bus to school live

on Route 55.
Then would this be true?

None of Mr. Jones' students own dogs.

58.

A.

B. NO

C. MOM

the Cornell Class

Page 18

- Reasoning Test, Tbrm X

Here is a reminder of the meaning of the possible answers:

YESIt must be true.
NOIt can't be true.

MAYBE It maybe true or it may not be true.

You weren't told

enough to be certain whether it is "YES" or "NO".

Suppose you know that

All Y's are X's.
No Z's are Y's.
A.

YES

Then would this be true?

N3
No Z's are X's.

MAYBE
Suppose you know that

All teachers

are college graduates.

All people Who have

gone to high school are men

All college graduates have

gone to high school.

A.

Then would this be true?

YES

NO

At least some teachers

are not men.

MAYBE

Suppose you know that

All Z's are Y's.
No X's are
All T's are Z's.
A.

YES

Then would this be true?

B.

No X's are T's.

C.

NO
MAYBE

Suppose you know that

All students who do

not have a star are also

not swimmers.
Frances is a swimmer.
Then would this be true?

Frances does not have a star.

A.

YES

NO

BUBB

The Cornell Class- Reasoning Test) librm X

Page 19

Here is a reminder of the meaning of the possible answers.

YESIt must be true.
BOIt can't be true.

MAYBE

It may

be true or it may not be true. You weren't told

enough to be certain whether it is "YES" or "ND ".

Suppose you know that

All the people in the auditorium are watching a movie.
All students in the senior gay are in the
auditorium.
Bather is a student in the senior platy.
63.
men would this be true?
Esther is not watching a movie.

64.

A. /ES
B. ND

C. Man

Suppose you kmcmrthat

All birds have three

eyes. No ducks are birds.

A. US

Then would this be true?

BD

NO ducks have three eyes.

MAYBE

Suppose you know that

Jo Vs are I's.
No I's are Ils.
Then would this be true?
At least some Z's are X's.

65.

A.

!NB

JO

)SAM

66. Bupirose you know that

All 'ot the red pencils are broken.
pencil is not broken.

Then would this be true?

moil's pencil

is not red.

A. =3
NO

MADE

The Cornell Class-Reasoning

Test, Form X

Page 20

Here is a reminder of the meaning of the possible answers.

YESIt must be true.

Z)It can't be true.
MAYBE It say be true or it may not be true. You weren't told
enought to be certain whether it is "IN" or "NO".

67. Suppose you know that

Then

All Vs are Y's.

All things that are not
would this be true?
At least some Z's

68.

Suppose you know

X's are also not Y's.

C. MAYBE

are not X's.

that

are not roses.

this

be true?

least some of Mrs. Brown's

At

A. YES
B. M)

At least some of Mrs. Brown's flowers are not roses.

At least some of the flowers in the flower show
Then would

67.

flowers

68. A.
B.
C.

EIS
NO
MAYBE

are in the flower show.

69.

Suppose you know

that

An the pencils in the box are yellow.

None of the broken

pencils are yellow.

All Dick's pencils are in the box.

Then would this be true?
None of the broken pencils
70.

are

Dick's.

69. A.

us

B.

NO

C.

MAYBE

Suppose you know that

All the people who live near the lake can swim.
None of the students in Mr. Smith's class live
near the lake.
Then would this be true?
At least

KM

of the students in Mr. Smith's

class cannot swim.

70. A.
B.

YES
NO

C. MADE

The Cornell Class-Reasoning Test, Form X

Page 21

Here is a reminder of the meaning of the possible answers.

YESIt must be true.

NOIt can't be true.
MAYBE It maybe true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

71.

Suppose you know that

None of the houses on Main Street are made
of brick.
Allan's house is not made of brick.
Then would this be true?

71. A.
B.

Allan's house is on Main Street.

72.

C.

YES
NO
MAYBE

Suppose you know that

At least some of the boys in the class have

bicycles.
All those who are not here also do not have
bicycles.
Then would this be true?

72. A.
B.

No boys in the class are here.

C.

YES
NO

man

1
73.

Suppose you know that

All dogs are red.

A.

Then would this be true?

All animals that are not red are also not dogs.

YES
B.

NO

C.

MAYBE

Suppose you know that

All Nr. Smith's cars have polished bumpers.
The red car does not have a polished bumper.

All the cars in the

garage are Mi.. Smith's.

Then would this be true?

the red car is not in the garage.

A.

YES
B.

NO

C. NUN

Page 22

The Cornell Class - Reasoning Test, Pbrm X

Here is a reminder of the meaning of the possible answers.

YESIt must be true.
N)It can't be true.
MAYBE It may be true or it may not be true. You weren't told
enough to be certain whether it is "YES" or "NO".

75.

Suppose you know that

No ducks are birds.
Nothing with large feathers is a bird.
Then would this be true?

75.

At least some ducks have large feathers.

76.

A.
B.

YES
N)

C.

MAYBE

A.
B.

YES
NO

C.

MAYBE

Suppose you know that

All alligators are smart animals.

All animals that cannot sing are also not
smart.
Then would this be true?

76.

At least some alligators cannot sing.

TT.

Suppose you know that

All the students who live in the country have
pets.
Barbara does not live in the country.

Tr.

Then would this be true?

Barbara does not have a pet.

A. =3
B.

Ito

C.

MAYBE

Suppose you know that

All X's are Y's.
All Z's are T's.
All Y's are Z's.
Then would this be true?
At least some X's are not T's.
..NININawsw

END OF TEST.

00 BACK AND =OK YOUR ANSWERS.

78.

A. YES
B. 113

C. MADE

Trial Answer Sheet for Lover Elementary

Students

Name:

ril

P5-]
45 N.

YES

NO
MAYBE
MAYBE

SKIP

SKIP

15

MAYBE

SKIP

YES
NO
MAYBE

NO
MAYBE
MAYBE

SKIP

SKIP

map

NUE
I

a=

SKIP

Cornell Deduction Test, Subtexts on Ordinal and Propositional

Reasoning, Form A
P1 2

34

Class Reasoning Exercises Used in 10th and 12th Grades, Consisting of Three
Used Early in the Instructional Period and One Used at the Ehd.

Name
Exercise 1.
Define "set"

Listed below are five groups of three objects each.

each group belong to?

my pencil, your pencil, the teacher's pencil - b.
a cocker spaniel, a terrier, a bloodhound -

breAd, milk, meat

books, classes, filmstrips

What "set" does

e.

basketball, baseball, wrestling -

List five "sets" of things you encounter every day:

a.

b.
c.
d.
a.

Redefine "set"

- in your own words and without looking back.

Define "element"

Listed below are five "sets".

belong to the set:

books -

For each one, name three elements that

b.

dogs

C-2
Ekercise 1, continued
nations A

subjects taken in school days of the week List five objects in this room which are elements of sets:
a.

b.
c.
d.
e.

Redefine "element" - in your own words and without looking back.

C.3

Miran 2

Name

Each of the following pairs of words includes one set and one element
of that set.

In the space to the right of each pair draw a circle to

represent the set, an x to represent the element, and label both.

I have done the first one for you so that you will understand What
is to be done.

subjects in school
subjects in school
American History
American History
teachers

Mr. Morrow
Monday
days of the week

newscasters
Chet Huntley

this

trout

fish

logicians Leonhard
Euler

you
students of logic
birds
Wes- blooded animals

Name

Exercise 3.

Define "total inclusion"

Below are eight pairs of sets.

Determine the relationship between

them, and draw circle diagrams to represent that relationship.

Diagrams

universities state
universities

things to write with

pent,

American nations
North American nations

summer months
months

suits
clothing

things

*Leh

entertain

movies

states of the United States

southern states of the United States

presidential candidates
Republican presidential candidates
Redefine the relationship of total inclusion, using your own words and

without

looking back

C-5

Quiz

Name

Judge each of the following arguments. Decide whether the conclusion

must follow - if it must, circle the work "VALID".

If the conclusion

either can't follow or may or may not follow, circle the word "INVALID".

At least some voters favor Wallace.

All those who favor Wallace oppose President Johnson.
Therefore, at least some voters oppose President Johnson.

Valid

Invalid

At least some dog.: are carnivorous animals.

Therefore, at least some carnivorous animals are dogs.

Valid

Invalid

At least some ducks are not wild.

All wild animals are protected by law.
Therefore, no ducks are protected by law.

ValidInvalid
At least some Republicans do not support Rockefeller. All
the Cayuga County supervisors are Repbulicans.

Therefore, at least some Cayuga County supervisors do not support

Rockefeller.
Valid

Invalid

At least some Assemblymen are not in favor of higher tam.

No one Who approves civil rights is in favor of higher taxes.
Therefore, some Assemblymen approve civil rights.

Valid

Invalid

At least some patrolmen use radar. All

the men in this room use radar.

TberefOre, at least some of the men in this room are patrolmen.

ValidInvalid

c-6
At least some Finns oppose Communism. No
Marxists oppose Communism. Therefore,
no Finns are Marxists.

ValidInvalid
At least some doctors are not surgeons. All
brain specialists are surgeons.

Therefore, at least some doctors are not brain specialists.

Valid

Invalid

At least some motorcycles are not six-cylindered.

Therefore, at least some six-cylindered things are not motorcycles.

ValidInvalid
At least some cameras are expensive.
At least some Japanese products are not expensive.
Therefore, at least some cameras are not Japanese products.
Valid

Invalid

At least some Cubans are not Communist.

At least some supporters of DeGaulle are not Communist.
Therefore, at least some supporters of DeGaulle are Cubans.
Valid

Invalid

At least some Picasso paintings are valuable. All

valuable things are expensive.

Therefore, at least some Picasso paintings are expensive.

Valid

Invalid

At least some magazines carry world neva.

At least some daily publications carry world neve.
Therefore, at least some magazines are daily publications.

ValidInvalid

C-7

At least some German scientists are in Berlin.

At least some NroAesterners are not in Berlin.
Therefore, no German scientists are pro.Westerners.

ValidInvalid
At least acme cancers can be cured.
Nothing that can be cured is a terminal case.
Therefore, at least acme cancers are not terminal cases.

ValidInvalid

Conditional Reasoning Exercises Used in 11th Grade, Consisting of

Three Used Early in the Instructional Period and One Used at the End.
'Exercise 1'

Find, label and put in parentheses the antecedents and consequents

in each of the following:
If a dog is a beagle, then he can hunt rabbits.
If zombies sing, there is moonlight.
There are phlips if there are phlops.
Parking permits are legal if they are issued by the police

chief.

If it requires real thought, homework is good.

if you should fail all your English tests,
course.

T.

If donne are klobbered, klobs are

then you will definitely

conned.

I will go to the World's Flair if I can get a

hotel reservation in New

That team of wrestlers must be good if they can beat the

Cove recite poetry if the Queen of Hearts plays

c-8

fail the

Ithaca team.

scrabble.

York.

C-9
'ftercise 2'

Try making each of the following arguments

valid by affirming the

antecedent and concluding the consequent.

1.

2.

3.

Premise

Premise

Conclusion

Therefore, (

Premise

Premise

Conclusion

Therefore, (

Premise #1

(These are valuable books) if (they are hand-printed)

Premise #2

Conclusion

Therefore, (

If (Joe makes that free throw), (we will win the game).
)

If (there is a circle), then

(there is a square)

)
)

4.

5.

Premise
-

If (this specimen is an insect), then (it has three pairs

Premise

Conclusion

Therefore, (

Premise

Premise

Conclusion

Therefore, (

Premise-

7.

8.

of legs)

(I will be late for school) if (I eat six

If (John is a junior),

eggs)

(he is in Mr. Brown's homeroom)

Premise
Conclusion

Premise

Premise

Conclusion

Therefore, (

If (a iihoribit is a sabit), (a pringle is a plop)

Premise

Conclusion

Therefore, (

Premise

1
(

)
)

Therefore, (

(The argument is valid) if (I affirm the

)

antecedent)

9.

Premise

Premise
Conclusion

(I will be unhappy) it (I don't win the contest)

Therefore, (

C-11
'ftercise 3'
Study each of these arguments,

and indicite Whether it is valid or

invalid by circling the appropriate word.

If it is Saturday, then you can sleep late.

It is Saturday.
Therefore, you can sleep late.

If

you

pass

all

your

tests,

you

Will

Valid

pass

Invalid

the

course. You have passed all your teats.

Therefore, you have passed the course.

Valid

If Mike is a dog, then he is an animal.

Mike is an animal.

Invalid

Therefore, Mike is a dog.

You must pay a fine if

your library book is overdue.

Valid

Invalid

Valid

Invalid

Your library book is overdue.

Therefore, you must pay a fine
5.

If a car runs out

of gas, it Will stall.

My car has stalled.

Therefore, ray car has

Lemonade is bitter if

Valid

run out of gas.

Invalid

there is no sugar in it.

There is no sugar in this lemonade.

Therefore, this lemonade is bitter.

If

ve

win

the

slalom,

then

we'll

Valid
have

Invalid

gold

medal. We will win the slalom.

Therefore, we won't have a gold medal.
8.

Valid

Invalid

If p, q.
P.
Therefore, q.
Valid

Invalid

C-12
q if p

Therefore, p.

Valid

Invalid

We'll have a hootenanny if ve can get the Rooftop Singers. We

can get the Rooftop singers.

Therefore, ve'll have a hootenanny.

Valid

Invalid

C-13
'Ekercise 14,
Name

Is the underlined conclusion valid?:

If a questionnaire is distributed to all adult members of a community,

if this questionnaire asks whether the person thinks slave labor wrong,
if there is at least 90% response, and if everyone tells the truth - then,
the community really thinks slave labor is wrong, if at least 80% say they

think slave labor is wrong.

A questionnaire was distributed to all adult

members of Smithtown, and it asked whether they thought slave labor was
wrong.

95% of them responded to the questionnaire, and the people of

Smithtown always tell the truth.

Now I know that Smithtown really thinks slave labor is right.

Therefore,

it is false that at least 80% of the adults of Smithtown said that they
think slave labor is wrong.

Steps
1.

2.

3.
4.

5
6.

7.
8.
9.

Reasons:

c-3.4
Reasons:

12.

etc.

A Class Reasoning Exercise Used

toward the End of the Instructional

Period in 4th and 6th Grades'

Grade
Name
Date
Directions: Read the arguments carefully. Then using the Euler circle,

diagram them.

Circle the correct answer for

each question.

Remember, work against the conclusion but without

breaking the rules.
All

Parisians

Frenchmen

are

Frenchmen.

are

Europeans.

All

All

Good 'Blinking!

Valid
Invalid

Parisians are Europeans.

to dogs are cows.

Valid

No cows are dogs.

Invalid

All Frenchmen are Europeans. All

Valid

Parisians are Frenchmen.

Invalid
All Europeans are people.
All Parisians are people.
4.

All t rout are fish.

a) Valid
All rainbows are trout.

b) Invalid

No cats are fish.

No cats are Rainbows.

5.

All Parisians are Frenchmen.

No Ithacans are Parisians

a) Valid

b) Invalid

No Ithacans are henchmen.

C15

C-16
No A's are B's. No
B's are C's. No

Valid
Invalid

A's are C's.

All lions are cats. All

tigers are cats. All
lions are tigers.

Valid
Invalid

All mice have tails. All

rats have tails. No

Valid

mice are rats.

Invalid

All A's are B's. No

C's are A's. No
B's are C's.

Valid
Invalid

All lotatoss ire vegetables.

No,iiiericeas are vagetablei.

No Jimmixems are potatoes

Valid
Invalid

A Conditional Reasoning ftercise Used toward the End of the

Instructional Period in 5th Grade
Name
Date
Grade
Directions:

Using the method of writing the symbols right over the

sentences, decide whether each of the following arguments is valid or

invalid.

Write valid or invalid in the space provided, whichever

correct.

On the line below each argument, write that should be in the

conclusion.

If this is an apple, then it grows on a tree. This is an

apple.
Therefore, it grows on a tree.

If

this

is

Buick,

then

it

is

an

automobile. It is not an automobile.

Therefore, it is a Buick.

If you like ice cream, then you will eat it often. You do eat
ice cream often.
Therefore, you like ice cream.

If

this

is

book,

then

it

has

pages. This is not a book.

Therefore, it does not have pages.

If this is en inclined plane, then it is not a lever.

It is a lever.

Therefore, it is not an inclined plane.

is

C-18
If zebras have stripes, then so do tigers.
Zebras do not have stripes.

Therefore, nothing follows, necessarily.

It has a fulcrum, if it is a lever.

It is a lever.

Therefore, it has a fulcrum.

The man has a hammer, if he is a carpenter.

The man has a hammer.
Therefore, he is a carpenter.

Elephants fly, only if birds have trunks.

Elephants do fly.

Therefore, birds have trunks.

If it rains tonight, we will catch a fish tomorrow. If we catch

a fish tomorrow, then we will eat it. It rains
tonight.

Therefore, we will eat it.

If we passed our grade, then ire passed our subjects.

If we passed our subjects, then we studied.

We did not study.
Therefore, we passed our grade.

C-19
If

vheee

0000m,

then

g000sh

braaack.

Wheee

0000m.

Therefore, g000sh braaack.

Either you like sunny days or you like rainy days. You like
sunny days.
Therefore, you like rainy

days.

If this is a seat, then you sit in it.

This is not a seat.
Therefore, you do not sit in it.

If girls grow beards, then, boys wear lipstick. Boys do

not wear lipstick.
Therefore, nothing follows, necessarily.

I f %, then S.
Not $.
Therefore, not %.

If you are old enough, you may stay up until 9:00 P.X.

You may stay up until 9:00 P.M. only if you behave.

You are old enough.
Therefore, you behave.

C-20
This is a simple machine, if it is a pulley.
This is not a simple machine.
Therefore, nothing follows necessarily.

Sally's dress either has a belt or it is red.

Sally's dress does not have a belt.
Therefore, it is red.

If hee, then ha.

Ha.

Therefore, hee.

:If this is A screw, then it has threads.

It does not have threads.
Therefore, nothing follows, necessarily.

If Jane likes Betty, then she sent Betty a valentine.

Jane likes Betty.

Therefore, she sent Betty a valentine.

If this has a fulcrum, tnen it is a lever.

It is a lever only if it is a simple machine.
It is not a simple machine.
Therefore, it does not have a fulcrum.

If boo, then bee.

Bee.
Therefore, nothing follows, necessarily.

25.

C.21
This is a wedge only if it is not a pulley.
It is a wedge.

Therefore, nothing follows necessarily.

ERRORS

Page and Line No.

11-22, 1.21:

H. P. Grice

111-3, 1.17: Change 'directly' to

'deliberately'.

111-3, 1.22: Quotes around "LNDT -2's"

111-8 & 9: Table No. 111-2 instead of 111-3
IV-10, 1.7: unless

of line 2 with the following:

not been
"...application of the concept when the operation has
performed,
operations have been performed'.

IV -15: Replace line 3 and last two words

in particular when other

IV-27, 1.13: Omit comma after 'see'.

IV-28: Note

"The correlations for CA for grades

combined are based upon a random sample of LDT's. For

g should read:

conditional reasoning N=64; for class reasoning,

V-2, Piaget quote, 1.1: between

V-9, 1.19: Insert a comma after 'question'
V-12, 1.1: "...that arguments with suggestive
VI-11, 1.9:

Cronbach's

VI-24, end of line 25: Add 'be'.

VI-28, 1.17:

inconsistent

content..."

N=82.