Lecture Note 1

STA 433
Statistical Methods &
Field Experiments
Dr. M. K. Garba Caution
This learning material is NOT FOR SALE by
Department of Statistics, anybody, not even by the Class
University of Ilorin, Nigeria Representative. Appropriate sanctions
08035568461 await the violators.
garba.mk@unilorin.edu.ng
Course Contents
1.Introduction to Field Experiments
2.Selection of Designs for Specific Situations
3.Collection and Analysis of Data
4.Introduction to Analysis of Variance Techniques
5.Design of Experiments: Completely Randomized,
Randomized Complete Block, Latin Squares, Pair
Plot, Split-plot and Factorial Designs
This learning material is NOT FOR SALE 2
 Introduction
Experiment can be defined as a test or series of
tests in which purposeful changes are made to the
input variables of a process or system so that we
may observe and identify reasons for changes that
may occur in the output/response.
Virtually in all fields of inquiry, experiments are
performed by investigator (experimenter) usually
to discover something about a particular process
or system.
Experiment involves the planning and collection of
measurements or observations according to a pre-
arranged plan.
This learning material is NOT FOR SALE 3
 Illustration
In an agricultural experiment where two varieties
of a particular crop are planted on two similar
plots and given similar management, a clear-cut
difference in yields between the two plots can
reasonably be attributed to the difference in the
yielding abilities of the two varieties.

This learning material is NOT FOR SALE 4

 Kinds of Experiment
There are basically two kinds of experiment;
laboratory and field experiments.
1. Laboratory Experiment – This is an experiment
that is conducted under a controlled environment
of which its outcomes are predictable ahead.
2. Field Experiment – This is an experiment that
takes place outside a laboratory, in a real-world
setting. In this experiment, an experimenter
applied treatment to the subjects to examine if a
causal relationship exists.
This learning material is NOT FOR SALE 5
Field experiments are designed to compare the
outcomes of applying one or more treatment to
experimental units, then comparing the results to a
control group that does not receive a treatment.
Designing a good experiment starts with
identifying the question(s) an investigator wants to
answer, followed by carefully planning all the
details of the experiment in advance. It is essential
that the experimenter reduces the probability of
errors and increases the likelihood that the
experiment will produce good data that will
enhance reasonable conclusion after the analysis.
This learning material is NOT FOR SALE 6
 Needs for Statistics in Field Experiment
Statistical approach to design of experiments is
imperative if the experimenter wishes to draw
meaningful conclusions from the data. When the
problem involves data that are liable to experimental
errors, statistical methodology is the only objective
approach to the analysis.
Hence, there are two aspects to any experimental
problem: the design of the experiment and the
statistical analysis of the emanated data from the
experiment. The two aspects are closely related
because method of analysis depends directly on the
design employed.
This learning material is NOT FOR SALE 7
To use the statistical methods to design
experiments and consequently analyze the
emanated data from such experiments, it is
necessary to consider in advance what
exactly is to be investigated or studied, how
the data would be generated and collected,
and then how such data would be analyzed
so that valid conclusions can be drawn.
This learning material is NOT FOR SALE 8
 Definitions of Some Terms
Experimental design refers to the statistical
procedure of imposing certain treatments on
subjects or materials (experimental units) so as to
be able to monitor the effects of these treatments
on the units.
An experimental design is a plan and a structure
to test hypotheses in which the researcher either
controls or manipulates one or more variables.

This learning material is NOT FOR SALE 9

Experimental unit is the smallest unit to which a
treatment is applied and on which an observation
is made. The experimental unit may be an animal
or plant, a number of birds in a pen, a number of
rats in a cage, individual leaves on growing
plants, half-leaves, or plot of land, etc.
For conducting an experiment, the experimental
material is divided into smaller parts and each
part is referred to as experimental unit.

This learning material is NOT FOR SALE 10

Treatment is an entity or phenomenon under study in an
experiment whose effect is to be measured and
compared to others. Treatment may be a standard ration,
a spraying schedule, a temperature-humidity
combination, and so on. Treatment could be classified as
absolute or comparative.
✓ Absolute Experiment An absolute or single phenomenon
experiment is one which contains a single treatment.
✓ Comparative Experiment A comparative experiment is
one designed specifically to compare two or more
treatments, which might be different varieties of maize,
different diets for cows, different drugs, or different acid
concentrations in an industrial process.
11
This learning material is NOT FOR SALE
Factor is a possible cause of variation / response in
an experiment. In other words, it is a controllable
experimental variable that can influence the
observed values of the response variable.
A factor may be qualitative (for example, different
concentrations of nitrogen in a fertilizer
experiment) or quantitative (different kg of
animal feed, as an instance).

This learning material is NOT FOR SALE 12

Levels of a factor are individual conditions or
values that make up a factor. A factor usually has
at least two levels which represent the treatments
of the experiment.

Response Variable is the characteristics of

interest that are measured or observed from the
experimental units. This is sometimes referred to
as the explanatory variable of the experiment.

This learning material is NOT FOR SALE 13

Basic Principles of Experimental Design
There are three basic principles of experimental
design. These are replication, randomization
and local control.
1. Replication – This implies a repetition of the
basic experiment or application of a treatment
in more than one unit of the experimental
material.
When a treatment appears more than once in
an experiment, the treatment is said to be
replicated.
This learning material is NOT FOR SALE 14
Replication allows experimenter to have reliable
estimate of the experimental error. The more units
used for each treatment, the lower the standard error
of the estimates of treatment effects and hence the
more precise the estimates from the experiment are.
Therefore, increasing replication increases precision,
although it could be costly. However, replication
beyond the limit may be impractical because it
depends on the availability of experimental units,
time, cost or ability of the experimenter.
This learning material is NOT FOR SALE 15
 Importance of Replication
Replications allow experimenter to:
1. Estimate the error variance due to
uncontrollable or assignable causes in the
experiment.
2. Make experimental result “powerful” enough
to recognize true differences between groups
(statistical significance) by increasing the
accuracy of estimates of means, and other
functions of the response variable.

This learning material is NOT FOR SALE 16

2. Randomization – By this principle, it is
required that both the allocation of the
experimental material and the order in which
the individual trials of the experiment are to be
performed are randomly determined. That is to
say, without any influence of the experimenter.
In other words, randomization is the allocation
of treatments to units such that the probability
that a particular treatment will be allocated to a
particular unit is the same for all treatments.
17
This learning material is NOT FOR SALE
 Remarks
Randomization is an objective or fair method of
random allocation of the experimental material or
treatments in an experiment to the experimental
units.
Randomization is also used to ensure that the order
in which the trials are performed are random.
Thus, randomization subjects all treatments to as
nearly equal conditions as possible by utilizing
chance allotments.
18
This learning material is NOT FOR SALE
 Importance of Randomization
1. It protects the experimenter from “systematic”
error that might be caused by subjectively
assigning the treatments to the experimental
units.
2. It assists to obtain unbiased estimates of
differences among treatment responses
(means or effects).
3. It helps to obtain an unbiased estimate of the
error variation in the experiment.
4. It is used to validate the underlying statistical
assumption that the errors are randomly
independently distributed
19
This learning material is NOT FOR SALE
3. Local Control – This is a procedure or
technique for reducing experimental error and
improving precision with which comparisons
among the factors of interest are made.
Local control is also referred to as blocking or
grouping or error control. It is required that
experimenter should try as much as possible to
group units in such a way that units within the
group are as similar as possible but may be
different from units in other groups.
20
This learning material is NOT FOR SALE
For instance, an investigator might decide whether to
group experimental units by gender, age, litters,
breed, systolic blood pressure, or by any other factor
that may be deemed to influence the response
variable.
The principle of blocking helps the experimenter
to reduce experimental error by identifying
assignable causes of variation and takes care of
them accordingly.

21
This learning material is NOT FOR SALE
Error control can be accomplished by the
following techniques;
1. Blocking – Appropriate grouping of the units
such that units in a group or block are as
homogeneous as possible.
2. Use of concomitant observations or variable as
in analysis of covariance (ANCOVA).
3. Plot technique as in field experiment, that is,
use of appropriate size and shape of the
experimental units.

22
This learning material is NOT FOR SALE
 Objectives of an Experiment
In designing an experiment, it is necessary
to state the objectives clearly as questions
to be answered, hypotheses to be tested,
and effects to be estimated.

This learning material is NOT FOR SALE 23

 Guideline for Designing a Field Experiment
1. Statement of the problem
2. Choice of factors and their levels
3. Selection of the response variable
4. Choice of experimental design
5. Performing the experiment
6. Carry out the statistical analysis
7. Drawing conclusions and recommendations

24
This learning material is NOT FOR SALE
 Class Activity I
An agricultural experimental station manager wished to
examine yield differences of 5 varieties of cowpea with
each variety planted on 6 different plots. At the end of
farming season, the yields in ton from the plots were
measured.
1. Identify the factor in this experiment.
2. How many levels does the factor have?
3. Identify the treatment and state whether it is absolute
or comparative.
4. What is the response variable in the experiment?
5. Identify the experimental unit.
6. How many experimental units are there? 25
This learning material is NOT FOR SALE
 Class Activity II
Consider a situation where an investigator was interested to evaluate
effects of polluted stream water on fish lesions. He set up 3 aquaria
each with 30 fish and randomly applied pollutants labeled Toxin A,
Toxin B and Control (no contaminant) to the aquaria, one for each.
After 21 days, 10 fish were caught from each aquarium and the
number of lesions were counted and recorded for analysis.
1. Identify the factor in this experiment and state how many levels
it has.
2. Identify the treatment and state whether it is absolute or
comparative.
3. What is the response variable in the experiment?
4. Identify the experimental unit in this experiment.
5. How many replications are there in the experiment?
26
This learning material is NOT FOR SALE
 Class Activity III
An experiment involves investigating the performances of 4
brands of spark plugs intended to be used for motorbikes. Five
plugs of each brand were tested and the number of kilometers
covered at a constant speed until failure was observed.
1. Identify the factor in this experiment and state how many
levels it has.
2. Identify the treatment and state whether it is absolute or
comparative.
3. Identify the experimental unit.
4. What is the response variable in the experiment?
5. How many replications are there in the experiment?

27
This learning material is NOT FOR SALE
 Class Activity IV
A researcher conducted a study on effects of two different diets and
three different exercise programs on the glucose level in patient’s
blood. Five randomly selected patients were assigned to each group
for one month after which the patient’s glucose level was measured in
milligrams per deciliter (mg/dl) as shown below
Diet Program
A B
Program
Exercise

I 50, 61, 49, 71, 39 69, 67, 68, 70, 65

II 38, 46, 80, 79, 81 92, 86, 91, 80, 69
III 64, 54, 56, 62, 57 90, 83, 66, 80, 93

1. Identify the factors in this experiment and state how many levels
each has.
2. Identify the treatment and the experimental unit in this experiment .
3. What is the response variable in the experiment?
4. How many replications are there in the experiment? 28
This learning material is NOT FOR SALE
#1
Practice Questions
Students were randomly assigned to three classes during summer school
and each class was taught by a different method. At the end of the three
weeks of intensive teaching, a comprehensive examination was given and
the results are shown below
Class A Class B Class C
87 82 97
92 78 90
61 41 83
83 65 92
47 63 91
1. Identify the factor in this experiment and state how many levels it has.
2. What is the treatment in this experiment?
3. Identify the experimental unit.
4. What is the response variable in the experiment?
5. How many replications are there in the experiment? 29
This learning material is NOT FOR SALE
#2
A researcher wished to try different techniques to lower the blood pressure of
individual diagnosed with high blood pressure. The subjects were randomly
assigned to three groups; the first group takes medication, the second group
exercises, and the third group diets. After four weeks, the reduction in each
person’s blood pressure is recorded as shown below
Medication Exercise Diet
10 6 7
12 8 9
9 4 8
15 0 5
6 12
13

1. Identify the factor in this experiment and state how many levels it has.
2. Identify the treatment and the experimental unit in this experiment .
3. What is the response variable in the experiment?
4. How many replications are there in the experiment?

30
This learning material is NOT FOR SALE
#3
A researcher wished to test the effects of four different types of
plant food and three different types of soil on the growth of
certain plants planted in experimental pots in a greenhouse
while other factors like water, temperature and sunlight were
held constant. Five experimental pots were randomly assigned
to each of the treatment combinations.
1. Identify the factors in this experiment and state how many
levels each has.
2. Identify the treatment and the experimental unit in this
experiment .
3. What is the response variable in the experiment?
4. How many replications are there in the experiment?
5. How many pots were used for the experiment?
31
This learning material is NOT FOR SALE