CHAPTER : 6

ESTIMATION OF RELIABILITY AND VALIDITY

6.1 RELIABILITY

6.1.1 TYPES OF RELIABILITY

6.1.1.1 TEST-RETEST RELIABILITY
6..1.1.2 ALTERNATE OR PARALLEL FORMS RELIABILITY

6..1.1.3 SPLIT-HALF RELIABILITY

6,.1.1.4 METHOD OF RATIONAL EQUIVALENCE

(RUDER RICHARDSON FORMULAS)
6.1.2 ESTIMATING RELIABILITY OF VII

6.1.3 PREDICTION OF COMBINED RELIABILITY OF

FORMS A AND B
6.1.4 DIFFERENT RELIABILITIES OF THE VII IN
GUJARATI VERSION

6.2 VALIDITY
6.2.1 TYPES OF VALIDITY
6..2.1.1 OPERATIONAL AND PREDICTIVE VALIDITY
6,.2.1.2 FACE VALIDITY

.6..2.1.3 CONTENT VALIDITY

6,.2.1.4 FOCTORIAL VALIDITY
6,.2.1.5 CONSTRUCT VALIDITY
6,.2.1.6 CONCURRENT VALIDITY

6.2.2 VALIDITY OF THE VII

248
6.1 RELIABILITY

Reliability and validity represent two extremely

essential psychometric characteristics of a sound measuring

instrument. Tests can predict, diagnose or measure a trait

accurately and efficiently only if they are dependable and

trustworthy.

1
According to Anastasi :

Reliability refers to the consistency of scores

obtained by the same individuals when re-examined
with the same test on different occasions, or with
different sets of equivalent items, or under other
variable examining conditions.

This concept of reliability underlies the computation

of coefficients of test-retest reliability as well as of two

alternate forms of the samq test or the standard error of

measurement of a single score, whereby one can predict the

range of fluctuations likely to occur in a single

individual's score as a result of irrelevant chance factors.

Whenever anything is measured, whether in the physical,

biological or behavioural sciences, there is some

possibility of chance errors. This is true of psychological

tests as well. Variations of results obtained with the same

test administered more than once, using the same persons as

subjects, or within the parts of a test given only once, are

due not only to chance factors, which should be eliminated

so far as possible, but also to actual differences among the

individuals taking the test and to whatever defects may be

inherent in the instrument itself.

249
 In psychological testing, the term reliability has

been used to denote two closely related though some what

different aspects. In the first place, reliability refers to

the consistency of scores obtained from a single

administration of the test and is an index of internal

consistency. In the second place, it refers to the extent to

which the test yields the same results on several occasions

This is known as an index of temporal stability and is

associated with the predictive value of the instrument.

6.1.1 Types of Reliability :

Four types that are commonly used to estimate

reliability of a test are :

1 Test - retest method

2 Alternate - forms or parallel forms method •

3 The split-half method

4 The method of Rational Equivalence

6.1.1;1 Test-Retest Reliability : The simplest way to find

the reliability of the test is by means of a retest, or

repetition of the identical test on a second occasion. The

reliability coefficient which is also known as the

coefficient or stability is nothing but correlation between

the scores obtained by the same subjects on the two

administrations of the test. This corresponds to the concept

of temporal stability. The main advantage of this method is

the inclusion, of a new dimension - one that is external to

the test and independent of *it - time and that is perhaps

250
why Helim considers this method the most satisfactory means

of assessing test consistency.

When retest reliability is reported in a test manual,

the interval over which it was measured should always be

specified, since retest correlations decrease progressively

as this interval lengthens. There is n o t one, but an

infinite number of retest reliability coefficients for any

test. It is also desirable to give some indications of

relevant in ter ve nin g experiences of the subjects on whom

reliability was measured.

"Test reliability (by test-retest method) is an

« extent to which the same group of people get (sic)
the same scores (therefore, rank the same) on
reported testings... It is important because if a
test isn't consistent in the scores and ranking
o b t a i n e d from it, it c a n ' t be c o n s i d e r e d an
accurate measuring device. A test on which persons
scored very high (low) the first time, but very
low (high) the s e c o n d time, is like a b r o k e n
ruler, it's not useful in measuring anything."2

•6.1.1.2 Alternate or Parallel Forms Reliability : In case

of alternate or parallel forms reliability, retesting of the

same group is done with a parallel or equivalent form of the

test. The statistics thus obtained is known as the

coefficient of equivalence of the two forms.

Like test-retest reliability, alternate forms

reliability should always be accompanied by a statement of

the length of the interval between two test administrations,

251
as well as a description of relevant intervening experiences.

If the two forms are administered in immediate succession or

at essentially the same time, the resulting correlation is

designated as the coefficient of equivalence.

In the development of alternate forms, catfe should, of

course, be exercised to ensure that they are truly parallel.

Fundamentally, parallel forms of a test should be

independently constructed tests designed to meet same

specifications. The test should contain the same number of

items, and items should be expressed in the s a m e - form and

should cover the same type of content. The range and level

of difficulty of the items should also be equal.

Instructions, time-limits, illustrative examples, format and

all other aspects of the parallel form must likewise be

checked for comparability.

6.1.1.3 Split-half Reliability : From a single

administration of one form of a test, it is possible to

arrive at a measure of reliability by various split half

procedures. In such a way, two scores are obtained for each

individual by dividing the test into comparable halves. It

is apparent that split half reliability provides a measure

of consistency with regard to content sampling. Temporal

stability ' of the scores does not enter into such

reliability, because only one test session is involved. This

type of reliability coefficient is sometimes called . a

coefficient 'of internal consistency, since only a single

administration of a single form is required.

252
 To find split half reliability, the first problem is

how to split the test in order to obtain the most nearly

comparable halves. Any test can be divided in many different

ways. In- most tests, the first half and the second half

would not be comparable, owing to differences in nature and

difficulty level of items, as well as of the cummulative

effects of warming up, practice, fatigue, boredom and any

other factors varying progressively from the beginning to

the end of the test. A procedure that is adequate for most

purposes is to find the scores on the odd and even items of

the test. If the items were originally arranged in an

approximate order of difficulty, such, a division yields very

nearly equivalent half scores.

Once the two half scores have been obtained for each

subject, they may be correlated by the usual method and the

reliability of the full test can be computed by using

Spearman-Brown prophency formula.

6.1.1.4 Method of Rational Equivalence (Kuder-Richardson

Formulas): A fourth method for finding reliability, also

utilizing a single administration of a single form is based

on the consistency of subjects' responses to all items in

the test. This inter-item consistency is influenced by two

sources of error variance. (1) content sampling (as in

al t e r n a t e form and split ha l f re li ab il it y) ; and (2)

heterogeneity of the behavoiur domain sampled.. The more

homogeneous the domain, the higher the inter-item

consistency.

253
 The most common procedure for finding inter-item

consistency is that developed by Kuder and Richardson. As in

the split .half method, inter-item consistency is found from

a single administration of a single test. Rather than

requiring two half scores, however, such a technique is

based on an examination of performance on each item.

The method of rational equivalence stresses the inter

correlations of the items in the test and the correlations

of the items with the test as a whole. Four formulas for

determining test reliability have been derived of which the

one given below is perhaps the most useful

■VI £ jo<v
Y1 1 't
on - 1
in which
'fc

reliability coefficient of the whole test

^11 ~

Y) number of items in the test

the SD of the test scores

the proportion of the group answering a test

Y =
item correctly

% = (1-p) = the proportion of the group

a n s w e r i n g a t e s t i t e m i n c o r r e c t

The product of p and q is computed for ecah item, and

these products are then added for all items, to give pq.

Since in the process of test construction p is often

routinely recorded in order to find the difficulty level of

each item, this method determining reliability involves

little additional computation.

254
 A simple approximation to the above formula is often

useful to teachers and others who w a n t to determ ine the

reliability quickly.

The f o r m u l a i s :
v* - M Cn - ^ J
*YoII
Wher e : cvi - i ;

- reliability of th e whole t e s t

vi = numbe r o f i t e m s i n t h e t e s t

SD o f t h e t e s t s c o r e s

M = t h e mean o f t h e t e s t s c o r e s

As H. E. G a r r e t t r e m a r k s :. 4

"The r a t i o n a l e q u iv a l e n c e fo r m u la s a r e n o t
s t r i c t l y co m p ar ab le to the' t h r e e methods a l r e a d y
out lin e d . Like the s p i l t - h a l f te c h n iq u e , th e se
fo rm u la s p r o v id e an e s t i m a t e of th e i n t e r n a l
c o n s i s t e n c y of the t e s t and t hu s of the
d e p e n d a b i l i t y of t e s t s c o r e s . Rational
eq uivalence is s u p e r io r to the s p i l t - h a l f
technique in c e r t a i n t h e o r e t i c a l a s p e c ts , but the
a c tu a l d iff e r e n c e in r e l i a b i l i t y c o e f f ic ie n ts
f o u n d b y t h e t wo m e t h o d s i s n e v e r l a r g e a n d i s
often n e g lig ib le ."

6.1.2 E s tim a tin g R e l i a b i l i t y of VII :

In the p r e s e n t stu d y , the problem of r e l i a b i l i t y of VII

h as been a p p ro a ch e d from the f o l l o w i n g d i f f e r e n t a n g l e s :

(a) Coefficient of Temporal Stablility by test-retest

me t h o d

(b) Coefficient o f E q u i v a l e n c e by p a r a l l e l f o r m s m e t h o d

255
(c) Coefficient of Internal Consistency by spilt-half

method

(d) Coefficient of Inter-item Consistency by method of

rational equivalance

(a) Coefficient of Temporal Stability by test-retest

method : Tne coefficient of temporal stability was

determined by test retest method. If the time

interval between the two administrations of the

test is short, the immediate memory effect,

practice and the confidence created by

familiarity with the test material may vitiate the

realibility of the test. On the other hand, if

the time interval is pretty long, the real changes

in behaviour in terms of growth may again vitiate

the reliability of the test.

To avoid these difficulties, the present

investigator administered the retest after 4

weeks. Test-retest method was applied for both

the groups : (1) Secondary school pupils and (2)

Higher Secondary school pupils. Sanghavi High

School from the urban area and Sheth C.K. .High

School from the semi-urban area were selected at

random. The proportianate number of pupils from

each grade (IX, X, and XI & XII) was selected by

systematic random sampling. All- the pupils of

256
 different grades were administered the same

inventory again, in one group, as a retesting.

The total number of boys, thus, selected were 86

while the total girls were 81. The test-retest

reliability for all the ten areas of interest was

found out using product-moment ' Y" ’ for boys and

girls seperately and, the average Ys were

found out using Fisher's Z.

Table 6.1 presents Ys of test-retest realibility.

■ TABLE 6.1
AREA OF INTEREST-WISE TEST-RETEST REALIBILITY OF VII
(%.= 86 Boys, Nx= 81 Girls)
(Time - Interval 4 weeks)

Coeff.of temporal stability

Sr. Area of Interest -------------------------------
No. Boys Girls Average
U J N 5 ' 0 ' . P ' O O O O > - i C T ' H ‘ 4>-
i.n t_n £Ti I n l.n U l / i U i U i U l

U i 0 " J W O 'O 'O M ^ O 'O '

1 Outdoor
o H
H U i O O ' f ' U N O ' O O v J

2 Mechanical
3 Computational
4 Scientific
w Nj ot>

5 Persuasive
6 A r tis tic
7 Literary
8 Musical
O ' -v! I

9 Clerical
vi

10 Social Service

It can be read from Table 6.1 that throughout

all areas of interest, boys are having higher

coeffiet&nts of temporal stability ranging from

.56 (for literary interest) to .80 (for musical

interest). The girls' results of test-reliability

257
 are in ferio r, the minimum being .51 (for

m echanical interest) and the ma x i mu m, .66 (for

musical interest). Would it be that girls woul d

be less serious than boys in taking this

inventory; a t l e a s t in th e second a d m i n i s t r a t i o n ?

The a v e r a g e of b o th boys and girls together were

computed by c®iver t i n g T into F ish er’s Z,

coefficients and reconverting them again into

Yi.The a v e r a g e Ts have been shown in the last

coloumn; they range f r om .55 to .68. One s h o u l d

not forget the fact that one c a n n o t expect very

high coefficients of reliability as those found

w it h i n t e l l i g e n c e and a p t i t u d e tests. Personality

traits and interests are not com paratively so

stable and hence generally the later has low

reliability.

(b) C oefficient of Equivalence by Parallel Forms

Met hod : P a r a l l e l f o r m s r e a l i b i l i t y or c o e f f i c i e n t

of equivalence was determine'd for the 10

vocational in te re s t variables.

The sample of 100 subjects were selected out of

1237 subjects by a systematic proportionate

( g r a d e w i s e ) me t h o d o f s a m p l i n g a n d i n c l u d e d p u p i l s

of a l l the grades IX, X, X I , and XI I . The t o t a l

n umbe r o f p u p i l s selected from these four grades

was 23, 24 , 28 a n d 25 r e s p e c t i v e l y . These s t u d e n t s

w e r e a l r e a d y a d m i n i s t e r e d b o t h t h e Forms A a nd B.

258
 The areawise coefficients o f equivalence, thus,

obtained have been presented in Table 6.2.

TABLE 6.2
AREA Of INTERESTWISE COEFFICIENTS OF EQUIVALENCE

(N-100)

Sr. AREA OF INTEREST COEFFICIENT OF EQUIVALENCE

NO.

1 Outdoor

O 'O O f C O O
2 Mechanical
3 Computational
4 Scientific
5 Persuasive

O M U i C N P
6 Artistic
7 Literary
8 Musical
9 Clerical
10 Social service h-4
0-*

O'
hO
cr

nr
(D
44

ft
<D
Da

It can
H
a)

U
O
03

£
CD

coefficients of equivalence by parallel forms

method range from 0.60 to 0.96; the least value

being for 'Literary' interest while the highest

value being for 'Scientific' interest. Thus, all

the coefficients of equivalence are 0.60 and

above.

(c } Coefficient of Internal Consistency by Split-half

method :

Split half reliability coefficients or

coefficients of' internal consistency were

determined for the 10 vocational interest areas.

259
 The sample of 100 pupils selected at random

was the same as that used in Parallel Forms

method. The internal consistency 'coefficients,

corrected by the Spearman-Brown formula are

presented in Table 6.3.

TABLE : 6.3
M E A OF INTERESTWISE COEFFICIENTS ON INTERNAL CONSISTENCY
(N=100)
Sr. AREA OF INTEREST COEFFICIENT OF
No. INTERNAL CONSISTENCY

1 Outdoor .62
2 Mechanical .88
3 Computational .74
4 Scientific .79
5 Persuasive .70
6 Artistic .82
7 Literary .78
8 Musical .72
9 Clerical .79
10 Social service .79

Table 6.3 reveals that the coefficients of

internal consistency vary from 0.62 to 0.88; only one

coefficient of internal consistency is less than 0.70

(for 'Outdoor* interest).

The internal consistency of the VII developed in

Gujarati version seems to be quite moderate. Yet, one

point should not be forgotten, and that is, the format

of the schedule is totally different wherein one has to

show one's preference for each and every statement on a

five point scale; while in the original schedule out of

the triad one has to select one activity for liking and

the other for disliking.

260
(d ) The Method of Rational Equivalence . (Kuder-Richardson
Cofflcient :

Besides the test-retest reliability, Parallel

forms reliability and split-half reliability, the

rational equivalence method was also used to estimate

reliability of VII.

The Kuder-Richardson formulas 20 and 21 which are

very commonly used in estimating reliability by the

method of rational equivalence are not applicable for

VII. The K.R. formulas are "applicable only to those

tests whose items are scored

wrong, right or or
5
according to some other all-or-none system. "

Some tests, however, may have multiple scored

items. In this VII, the testee receives a different

numerical score on an item from 4 to 0, depending on

whether he or she checks 'likes the most', 'likes it',

'indifferent', 'dislikes it' and 'dislikes the most',

respectively. In such a case, 'a generalised formula

has been derived, known as coefficient alpha (Cronback

1951.... )f It is as follows :

2
Wherein, pq is replaced by € (SDt’)> the sum of
6 •

the variance of item scores" . In this study, the

above Cronbach alpha formula has been applied to

estimate reliability by the method of rational

equivalence.

261
 Table 6.4 presents the data about coefficients of

inter-item consistency.

TABLE : 6.4
AREA OF INTEREST-WISE RELIABILITY BY THE METHOD
OF RATIONAL EQUIVALENCE
(N=l00)

Sr. AREA OF INTEREST COEFFICIENT OF

No. • Cronbach alpha

-v l
1 Outdoor
2 Mechanical
3 • Computational

U 1 GO t D v l
4 Scientific
5 Persuasive
6 Artistic
7 Literary

K
8 Musical

OO U U O
9 Clerical
10 Social service

It can be read from Table 6.4 that the range of

Cronbach alpha coefficients is from 0.71 to 0.89 which can

be considered quite satsifactory looking to the number of

items (that is, 12 items), in each area of interest. It is a

very open fact that reliability of a test depends upon its

length. Simply by increasing the length of a test (adding

more items into it), the reliability of a test can be

boosted.

6.1.3 Prediction of Combined Reliability of Forms A and B:

It was mentioned in Chapter 5 that the standardisation

262
of the Form B would be taken up immediately as a post

doctoral work. However, coefficient of equivalence by

parallel forms A and B each containing 12 items has already

been estimated.

Taking these data as a split half reliability and

applying the very well known Spearman-Brown Prophecy

formula, one can predict the reliability of both the forms A

and B together. "As we increase the length of the test,

7
chance plays a less and less important role. "

Table 6.5 presents the data about coefficient of

equivalence of parallel forms which is now split-half

reliability and the predicted reliability of the full

in'entory (Forms A and B ) .

TABLE s 6.5
PREDICTED RELIABILITY OF THE FULL VII
(FORMS A AND B COMBINED )
(N=100)

Sr. AREA OF INTEREST Split half Predicted

No. Reliability Reliability
(From Table of the full
6,2) VII (Forms A and B)

1 Outdoor .68 .81

2 Mechanical .72 84
3 Computational .68 .81
4 Scientific • .96 .98
5 Persuasive .71 .83
6 Artistic .72 .84
7 Literary .60 .75
8 Musical .65 .79
9 Clerical .62 .77
10 Social service .70 .82

It is naturally visualised from Table 6.5 that

predicted reliability of the whole VII (Forms A and B) is

263
boosted up, area of interestwise. The range from 0.60 to

0.06 has been . increased to from 0.75 t,o 0.98. It can,

therefore, be concluded that whenever time permits, both the

forms A and B developed by the present investigator should be

administered to the subjects, after the Form B is

standardised.

6.1.4 Different Reliabilities of the VII in Gujarati

Version-

To have the gestalt of different types of reliability,

the present investigator has added one more Table 6.6

wherein all types of coefficients of reliability are shown

together to have a clear-cut picture, at sight; of the

reliability of the VII, in Gujarati Version.

2 64
 TABLE 6.6
DIFFERENT TYPES OF RELIABILITIES OF GUJARATI VII
TYPES OF COEFFI. OF C0EFF. OF C0EFF. OF COEFF. OF INT. PREDICTED COEFF.
CORRELATION TEMPO. STABILITY EQUIVALENCE IOT.‘IT CONSISTENCY T TPM CONSISTENCY OF INT.CONSISTENCY
Method Test Retest Parallel Split Half Cronbach Split half method
N=427 Forms Method alpha method (N=100)
N=100 N=100 N=100 Corrected By S.B.
(Corrected by prophecy formula
Sr. Area of Intersal S.B.Prophecy
No. Interest 4 weeks Formula

1 Outdoor 0.61 0.68 0.62 0.77 0.81

2 Mechanical 0.60 0.72 0.88 0.84 0.84
3 Computational 0.61 0 .6 8 0.74 0.77 0.81
4 Scientific 0.64 0.96 0.79 0.89 0.98
5 Persuasive 0.68 0.71 0.70 0.78 0.83
6 Artistic 0.67 0.72 0.82 0.75 0.84
7 Literary 0.55 0.60 0.78 0.71^ 0.75
8 Musical 0.74 0.65 0.72 0.73 0.79
9 Clerical 0.65 0.62 0.79 0.75 0.77
10 Social 0.57 . 0.70 0.79 0.78 0.82
Service
Vo
 Table 6.6 is self explanatory and needs no discussion.

Reliabilities estimatd by four different methods have

different specifications and they can, in no way, be

averaged, a well-known fact for the psychometr-icians.

6.2 VALIDITY ;
The validity of a test is related to what the test

purports to measure and how well it does so. The validity of

a test or of an inventory is frequently defined as, "The

extent to which the test or inventory actually measures what

it purports to measure." The determination of the validity

of an inventory would involve the correlation between scores

on the inventory and some pure criterion measure of what the

inventory purports to measure.

The construction and use of a test imply that

instrument has been evaluated against criteria regarded by

experts as the best evidence of the traits to be measured by

the test.

The first essential quality of a valid test is that it

should be highly reliable. If a reliability coeffieient of a

test is very low, it will not be correlated with anything. A

test that yields inconsistent results (low reliability)

cannot correlate well with a measure of another variable; in

this instance, a criterion.

266
 8 ,

Glenn and others observes :

Test validity is an extent to which scores on a

test correspond to the actual behaviour the test
is designed to measure.. If it turns out that the
behaviour the test was designed to measure has
little or no relation to the test scores, the test
is useless.

6.2.1 Types of Validity :

6.2.1.1 OPERATIONAL AND PREDICTIVE VALIDITY : Operational

validity means that the tasks required by the test are

adequate for the measurement and evaluation of certain

defined psychological traits. The predictive validity of a

test is the extent to which it is efficient in forecasting

and differentiating behaviour or performance in a specified

area under actual working and living conditions.

Predictive validity is dependent, at least in part,

upon the operational validity of the test. The reason is

that the psychological operations required by the test were

included because they were found essential for testing, in

certain actual situations. Hence, if the psychological

operations, or the information, or the. specific skills are

not measured truly, prediction or later performance will be

adversely affected.

6.2.1.2 Face validity : This form of validity now, more or

less neglected, has been disparaged since more sophisticated

procedures have been devised. As a matter of fact, however,

face validity in the earlier days of test development was

the criterion used by many competent psychologists as a

267
first step. Validation of their test-content at face value

was not as carpricious, haphazard, or casual as some critics

have said. On the contrary, their content was based upon

whatever psychological knowledge and insight could, then, be

utilised. Face validity was claimed most often with tests of

educational achievement and of personality and to a lesser

extent with tests of specific aptitudes.

6.2.1.3 Content Validity: It is the most appropriately

applied only to tests of proficiency and of educational

achievement, although such validity may be and should be

supplemented by several types of statistical analysis;

validity of content, however, should not depend upon the

subjective judgement of only one specialist. It should be

.based upon careful analysis, by several specialists, of

instructional objectives and of the actual subject-matter

studied.

6.2.1.4 Factorial validity : In the statistical method

called fator analysis, the items of a large number of tests

are examined and if possible, accounted for in terms of a

much smaller number of more general 'factors' or 'trait

categories'. The factors presumably run through the often

complex abilities measured by the individual tests. It is

some times found, for example, that 3 or 4 factors will

account for the inter correlations obtained among fifteen or

more tests. The validity of a given test is defined by its

factor loadings and these are given by the correlations of

268
the test with each factor. Factor analysis is a specialized

mathematical technique widely used and highly important in

modern test construction.

6.2.1.5 Construct Validity : It is concerned with the

meaning and interpretation of the test scores obtained in

terms of psychological or theoretical constructs. Construct

validity is concerned not only with the test itself but also

with the theory which seeks to explain, or to account for

the results which are obtained when the test is used.

Construct validity is established through a long

continued experimentation based on imagination, reasoning

and observation. First imagination suggests that a construct

X accounts for the test performance. The researcher, then

reasons, '^It that is so, individuals with a high score

should have characteristic or attribute Y.” An experiment is

performed and observations are made, and the expectations

are confirmed; the interpretation is supported. But as

^arrus; deductions are tested, some of them prove to be

inaccurate. The proposed interpretation is, then, altered by

involving a different concept,' by introducing an additional

concept, or by altering the theory of the concept itself.

The process of construct validation is the same as that by

which scientific theories are developed.

269
6.2.1.6 Concurrent Validity : Originally, psychologists

spoke to validation with other tests, validation with a

proficiency rating, validation with school grades, or, in

the case of a personality test, validation with a recent

diagnosis. At present, psychologists prefer the term,

"concurrent validity" to indicate the process of validating

the new test by correlating it, or otherwise comparing it

for agreement, with some present source of information. This

source of information might have been obtained shortly

before or after the new test has been administered.

9
Dougless and Choen rightly remark :

Two methods that demonstrate validity by

correlating tests with independent criteria are
predictive and concurrent validity. It is assumed
that psychological tests predict independent,
external criteria that might include specific
behaviour (e.g. suicide attempts), general traits
(e.g. scores on other psychological tests), or
certain future outcomes (e.g. graduation from
college).... the higher the correlation, the
greater test's predictive power, and more valid
the test is in placement. This is the essence of
predetive validity : current test scores foretell
some independent criterion in future.

Concurrent validity is nearly identical validity,

except that the outside criterion is assessed at
the same time the test is administered. Often a
pre-existing or well established test is used as a
criterion with which a new test is correlated ....
Obviously, concurrent validtion depends on having
identified a suitale external criterion, such as
the Wechsler scales.

270
6 . 2.2 VALIDITY OF THE VII

Face Validity : The definition of vocational interests

acc epted for this study was that of Kuder - "Vocational

Interests, in general, tell you what kind of duties you like to

perforin. " (p.lll). To be more specific, the investigator would

like to give an operational definition as follows :

"A Vocational Interest is an expressed preference for

engaging in a given occupational activity. The degree of a

person’s apparent occupational interests is reflected in the

extent to which the person expresses a greater preference for one

activity (or collection of activities) than for another activity

(or collection of activities)".

As the present inventory assesses students' "expressed

interest", it can be claimed that it possesses "face validity".

Content Validity : The rationale for the selection of

activities that are represented in the items on the inventory was

three-fold.

a day-to-day / seasonal activities

b curricular / co-curricular activities run in the schools,

and

c certain duties or activities carried out in different

occupations commonly found in this region.

The activities under (a) and (b) were such that they

were bases for future complicated or sophisticated types of

271
occupational duties or activities; they might be considered as

miniature activities for future occupations.

The newly coined items presented on pages 139-146 were

also devised on this rationale.

On page 158, the investigator has enlisted the newly

coined items actually selected in the present inventory by giving

serial numbers that can be found in Appendix H (English version -

pp. 412-16).

These thirty seven items are divided into three parts

as per the rationale.

a 18,20,21,46,53,58,64,103 -> 8 items

b 1,27,28,39,44,54,57,74,112,117,119 -> 11 items

c 3,11,19,25,62,65,67,71,76,78,79,80,85,92,94,96

97,120 -> 18 items.

Thus, the s e l e c t i o n of a c t i v i t i e s in the p r e s e n t

inventory representiviable options available in the state of

Gujarat. The present inventory, therefore, samples in a

reasonable way the 'content* of the labour market available to

the students, in this regij,. It can be concluded that the present

inventory possesses content validity.

Construct validity : Generally, "Vocational Choice" is

taken up as one of the 'constructs" for validation purposes. The

higher scores on areas of interest are pitted against the actual

vocational choice of a person. In India, there is a very acute

272
problem of unemployability; even graduates and some post­

graduates are either unemployed or grasp any vocation available

without taking into account the interest profile. There is a

question of one's survival. Again the "social status" accompanied

with an occupation is being given undue importance. Most of the

unemployed persons crave for "White Collar Jobs" only, though

there are a v a i l a b i l i t y of blue collar jobs. Under these

circumstances, it would be futile to take vocational choice as a

Construct.

Howsoever,' the investigator has taken up "teachers'

rating" as a construct and has estimated construct validity.

Teachers' Rating : The comparison of three highest

scores achieved by 100 pupils of Pravin Vidya Vihar of Ahmedabad

city was carried out with three highest ratings of teachers out

of the list of ten areas of interest. Teachers were instructed to

give first three ranks 1, 2 and 3 to those areas of interest

which teachers thought, the pupil has shown such interest in

different activities carried out in the school. Pravin Vidya

Vihar was selected purposefully as some guidance activities like

career talks, celebration of Guidance Week, keeping cumulative

record cards, Parent - Teachers' a s s o c i a t i o n , ■etc. were

implemented in that school. In short, the school was guidance -

oriented that has been rarely found in Gujarat, except in those

missionary schools.

273
 Then, the number of frequencies was counted : (a)

Common interest in three areas (pupils' scores and teachers'

rating) (b) common interest in any two areas of interest (c)

common interest in only one area and (d) no common area of

interest between the two. Equal probability hypothesis was

applied to the data and chi-square was found out.

To find out coefficient of contingency from chi-square,

11
the following formula was applied :
Z =
v N + X 1-
Table : 6.8 presents the necessry data.

274
 TABLE : 6.8
PERCENTILE NORMS AND TEACHERS’ RATING

Congruence of Frequencies in

ALL THE ANY TWO ANY ONE NONE TOTAL

THREE AREAS AREAS AREA AREA

11 14 63 12 100
25 25 25 25 100

O C 2 - 77.2 df - 3 P < 0.001

(Table Value 16.266)

Coefficient of Contingency = C = 0.66

It can be read from Table 6.8 that chi-square value is 77.2

which is significant beyond 0.001 level. The null hypothesis of

equal probability is, therefore, not accepted. This chi-square

value is, then, converted into contingency coefficient by the

formula mentioned earlier. This coefficient clearly indicates

t here is a significant relationship between first three

percentile scores of pupils and three areas of interest in which

pupils show their interest as per rating of teachers.

Importance of Measurement Of Interest in Vocational Counselling :

The measurement of interests is not merely justified

and desirable but is a necessity in a career planning as well as

in vocational counselling. Two very effective ways in which a

student engaged in a career planning can use his interest profile

are as a means of uncovering some new and promising vocational

leads within a vast network of existing vocations and as a device

for narrowing the wide range of job opportunities to those most

worthy of his intensive and serious investigation. The assumption

underlying the importance of the survey of students' or people's

interests is that expressed interest of th e V I I is on e

significant factor determiningj

275
a the types of occupations people seek to enter,

b the extent to which they remain in those occupations, and

c the level of success they enjoy in those occupations.

These three aspects can be taken up as 'predictive

validity' only when the pupils selected in a sample would finish

their study upto higher secondary level, certificate or diploma

courses, graduation or post graduation.

Of course, one has to be cautious not to give undue

weightage to 'interest' only; the other equally important factors

that should be taken into consideration while doing counselling

are aca d e m i c a c h i e v e m e n t , mental ability (intelligence),

aptitudes of pupils, personality traits, socio-economic level of

the parents, job-opportunities available, the experiences a

person undergoes, etc.

Two important aspects of validity, namely (1) what the

VII truly m e a s u r e s and (2) why it is i mportant to measure

interest in vocational counselling have been discussed.

In conclusion, it can be said that a new inventory has

sufficiently satisfactory reliability as well as validity.

The present investigator is quite aware that she need

not get inflated in any way. Mursell-^ very rightly remarks,

"The ultimate validation of any test is to be found only in its

wide and serviceable use". The present inventory must stand on

its own feet in its day-to-day trial and evaluation by one's

peers would be the final arbitrator.

275 A
 REFERENCES

1 A. Anastasi, : Psychological Testing (The Macmillan

Publishing Co. Inc., New York, 1968.)
p.103.

2 Sigrid S. Glenn : Introduction to the Science of Psychology

and et.al., (Michigan : Behaviourdelia, Inc., 1976.),
pp. 145-146.

3 H.E.Garrett, : Statistics in Psychology and Education,

(Vakils, Feffer and Simons Ltd., Bombay,
1981.) p.341.

4 , : Ibid., p .342.

5 A. Anastasi, : Op.cit., p.116.

6 , : Ibid., p.117.

7 A.K.Kurtz and : Statistical Methods in Education and

S .T.Mayo Psychology, (Narosa Publishing House,
New Delhi, 1980.) p.271.
8 Sigrid S. Glenn: Op.cit., p .146.
and et.al.

9. Shelley^ Dougless: Testing Psychological Tests.

anf David Cohen, (Croom Helm Ltd., Sydney, 1986.)
pp. 13-14.

10 Kurtz and Mayo : Op.cit., p.491.

1] ........ , : Ibid., p .390.

12 J.L.Mursell, : Psychological Testing (Longmans, Green &

Co., Newyork, 1950.) p.45.

276