RV Institute of Technology & Management ®

MODULE - V
DESIGN OF EXPERIMENTS & ANOVA

Introduction
Design of Experiments (DOE) is a systematic approach for planning and conducting experiments
to understand relationships between factors and responses. Analysis of Variance (ANOVA) is a
statistical technique for assessing differences in means between multiple groups. DOE helps
optimize processes, while ANOVA quantifies the significance of these optimizations. It's a
powerful combination in scientific research, manufacturing, and quality control to make informed
decisions. DOE designs experiments, and ANOVA analyzes results, enhancing decision-making
across diverse fields.

Topic Learning Objectives:

Experimental Design and Hypothesis Testing: Gain proficiency in selecting appropriate

experimental designs and using ANOVA for hypothesis testing to efficiently investigate the impact
of factors on a response variable.

Factor Analysis and Data Interpretation: Develop the skills to identify significant factors,
understand their interactions, and optimize processes or systems. Learn how to interpret and derive
meaningful insights from experimental data.

Practical Application: Apply DOE and ANOVA techniques in real-world scenarios, such as
quality improvement, process optimization, and scientific research, to make data-driven decisions
and enhance outcomes.

5.1.2 Upon Completion of this module, students will be able to:

Effective Experimental Planning: Develop the ability to plan and execute experiments
systematically, considering factors, levels, and experimental designs for efficient data collection.

Hypothesis Testing Proficiency: Gain expertise in formulating hypotheses and applying ANOVA
to determine statistically significant differences among groups or treatments and understand the
impact of factors.

Data Analysis and Interpretation: Learn to interpret and draw meaningful insights from
experimental data, making informed decisions and improvements in quality, processes, and
research based on empirical evidence.

Design and Analysis of Experiments

Planning an experiment to obtain appropriate data and drawing inference out of the data with
respect to any problem under investigation is known as design and analysis of experiments. This
might range anywhere from the formulations of the objectives of the experiment in clear terms to
the final stage of the drafting reports incorporating the important findings of the enquiry. The

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 1 | 24
 RV Institute of Technology & Management ®

structuring of the dependent and independent variables, the choice of their levels in the experiment,
the type of experimental material to be used, the method of the manipulation of the variables on
the experimental material, the method of recording and tabulation of data, the mode of analysis of
the material, the method of drawing sound and valid inference etc. are all intermediary details that
go with the design and analysis of an experiment.

Principles of Experimentation

Almost all experiments involve the three basic principles, viz., randomization, replication and
local control. These three principles are, in a way, complementary to each other in trying to
increase the accuracy of the experiment and to provide a valid test of significance, retaining at the
same time the distinctive features of their roles in any experiment.

Randomization: Assigning the treatments or factors to be tested to the experimental units

according to definite laws or probability is technically known as randomization.

This provides a basis for making a valid estimate of random fluctuations which is so essential in
testing of significance of genuine differences.

Replication: Replication is the repetition of experiment under identical conditions but in the
context of experimental designs, it refers to the number of distinct experimental units under the
same treatment. Replication, with randomization, will provide a basis for estimating the error
variance. In the absence of randomization, any amount of replication may not lead to a true
estimate of error. The greater the number of replications, greater is the precision in the experiment.

Local control: Local control means the control of all factors except the ones about which we are
investigating. Local control, like replication is yet another device to reduce or control the variation
due to extraneous factors and increase the precision of the experiment.

Note: In short, it may be mentioned that while randomization is a method of eliminating a

systematic error (i.e., bias) in allocation thereby leaving only random error component of variation,
the other two viz., replication and local control try to keep this random error as low as possible.
All the three however are essential for making a valid estimate of error variance and to provide a
valid test of significance.

Completely randomized design

A completely randomized design (CRD) is one where the treatments are assigned completely at
random so that each experimental unit has the same chance of receiving any one treatment. For
the CRD, any difference among experimental units receiving the same treatment is considered as
experimental error. Hence, CRD is appropriate only for experiments with homogeneous
experimental units, such as laboratory experiments, where environmental effects are relatively
easy to control. For field experiments, where there is generally large variation among experimental
plots in such environmental factors as soil, the CRD is rarely used.

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 2 | 24
 RV Institute of Technology & Management ®

Randomized complete block design

The randomized complete block design (RCBD) is one of the most widely used experimental
designs in forestry research. The design is especially suited for field experiments where the number
of treatments is not large and there exists a conspicuous factor based on which homogenous sets
of experimental units can be identified. The primary distinguishing feature of the RCBD is the
presence of blocks of equal size, each of which contains all the treatments.

Analysis of variance (ANOVA)

Analysis of variance (abbreviated as ANOVA) is an extremely useful technique concerning
researches in the fields of economics, biology, education, psychology, sociology,
business/industry and in researches of several other disciplines. This technique is used when
multiple sample cases are involved. As stated earlier, the significance of the difference between
the means of two samples can be judged through either z-test or the t-test, but the difficulty arises
when we happen to examine the significance of the difference amongst more than two sample
means at the same time. The ANOVA technique enables us to perform this simultaneous test and
as such is considered to be an important tool of analysis in the hands of a researcher. Using this
technique, one can draw inferences about whether the samples have been drawn from populations
having the same mean.

The ANOVA technique is important in the context of all those situations where we want
to compare more than two populations such as in comparing the yield of crop from several varieties
of seeds, the gasoline mileage of four automobiles, the smoking habits of five groups of university
students and so on. In such circumstances one generally does not want to consider all possible
combinations of two populations at a time for that would require a great number of tests before we
would be able to arrive at a decision. This would also consume lot of time and money, and even
then, certain relationships may be left unidentified (particularly the interaction effects). Therefore,
one quite often utilizes the ANOVA technique and through it investigates the differences among
the means of all the populations simultaneously.

The basic principle of ANOVA

The basic principle of ANOVA is to test for differences among the means of the populations by
examining the amount of variation within each of these samples, relative to the amount of variation
between the samples. In terms of variation within the given population, it is assumed that the values
of (𝑋𝑖𝑗 ) differ from the mean of this population only because of random effects i.e., there are
influences on (𝑋𝑖𝑗 ) which are unexplainable, whereas in examining differences between
populations we assume that the difference between the mean of the 𝑗 𝑡ℎ population and the grand
mean is attributable to what is called a ‘specific factor’ or what is technically described as treatment
effect. Thus, while using ANOVA, we assume that each of the samples is drawn from a normal
population and that each of these populations has the same variance. We also assume that all
factors other than the one or more being tested are effectively controlled. This, in other words,
means that we assume the absence of many factors that might affect our conclusions concerning
the factor(s) to be studied. In short, we have to make two estimates of population variance viz.,
one based on between samples variance and the other based on within samples variance. Then the
said two estimates of population variance are compared with F-test, wherein we work out.

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 3 | 24
 RV Institute of Technology & Management ®

𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒 𝑜𝑓 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑏𝑎𝑠𝑒𝑑 𝑜𝑛 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒

𝐹 =
𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒 𝑜𝑓 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑏𝑎𝑠𝑒𝑑 𝑜𝑛 𝑤𝑖𝑡ℎ𝑖𝑛 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒
This value of F is to be compared to the F-limit for given degrees of freedom. If the F value we
work out is equal or exceeds* the F-limit value (see the F table) we may say that there are
significant differences between the sample means.

ANOVA TECHNIQUE

Under one-way (or single factor) ANOVA, we consider only one factor and then observe that the
reason for said factor to be important is that several possible types of samples can occur within
that factor. We then determine if there are differences within that factor. The technique involves
the following steps:
̅1 , 𝑋
(i) Obtain the mean of each sample i.e., obtain 𝑋 ̅2, 𝑋
̅3, … , 𝑋
̅ 𝑘 when there are k samples.
ii) Work out the mean of the sample means as follows:
̅ 1 +𝑋
𝑋 ̅ 2 +𝑋
̅ 3 +⋯+𝑋
̅𝑘
𝑋̿ =𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠(𝑘)
iii) Take the deviations of the sample means from the mean of the sample means and calculate
the square of such deviations which may be multiplied by the number of items in the
corresponding sample, and then obtain their total. This is known as the sum of squares for
variance between the samples (or SS between). Symbolically, this can be written:
2 2 2 2
𝑆𝑆 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 = 𝑛1 (𝑋̿ − 𝑋
̅1 ) + 𝑛2 (𝑋̿ − 𝑋
̅ 2 ) + 𝑛3 (𝑋̿ − 𝑋
̅ 3 ) + ⋯ + 𝑛𝑘 (𝑋̿ − 𝑋
̅𝑘 )

iv) Divide the result of the step (iii) by the degrees of freedom between the samples to obtain
variance or mean square (MS) between samples.
𝑆𝑆 𝑏𝑒𝑡𝑤𝑒𝑒𝑛
𝑀𝑆 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 =
(𝑘 − 1)
Where 𝑘 − 1 represents degrees of freedom (d.f) between samples.
v) Obtain the deviations of the values of the sample items for all the samples from corresponding
means of the samples and calculate the squares of such deviations and then obtain them
total. This total is known as the sum of squares for variance within samples (or SS within).
𝑆𝑆 𝑤𝑖𝑡ℎ𝑖𝑛 = ∑(𝑋1𝑖 − 𝑋̅1 )2 + ∑(𝑋2𝑖 − 𝑋̅2 )2 + … + ∑(𝑋𝑘𝑖 − 𝑋̅𝑘 )2

vi) Divide the result of step (v) by the degrees of freedom within samples to obtain the variance or
mean square (MS) within samples.
𝑆𝑆 𝑤𝑖𝑡ℎ𝑖𝑛
𝑀𝑆 𝑤𝑖𝑡ℎ𝑖𝑛 =
(𝑛 − 𝑘)
Where (𝑛 − 𝑘) represents degrees of freedom within samples,
n=total number of items in all the samples i.e., 𝑛1 + 𝑛2 + ⋯ + 𝑛𝑘
k= number of samples.
vii) For a check, the sum of squares of deviations when the deviations for the individual items in
all the samples have been taken from the mean of the sample means.
2
𝑆𝑆 𝑓𝑜𝑟 𝑡𝑜𝑡𝑎𝑙 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = ∑(𝑋𝑖𝑗 − 𝑋̿) , 𝑖 = 1,2,3. . & 𝑗 = 1,2,3 …
This should be equal to total of the result of the (iii) and (v) steps explained above i.e.,
𝑆𝑆 𝑓𝑜𝑟 𝑡𝑜𝑡𝑎𝑙 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = 𝑆𝑆 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 + 𝑆𝑆 𝑤𝑖𝑡ℎ𝑖𝑛

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 4 | 24
 RV Institute of Technology & Management ®

The degrees of freedom for total variance will be equal to the number of items in all samples minus
one i.e., (n – 1). The degrees of freedom for between and within must add up to the degrees of
freedom for total variance i.e., (n – 1) = (k – 1) + (n – k)
This fact explains the additive property of the ANOVA technique.
(viii) Finally, F-ratio may be worked out as under:
𝑀𝑆 𝑏𝑒𝑡𝑤𝑒𝑒𝑛
𝐹 − 𝑟𝑎𝑡𝑖𝑜 =
𝑀𝑆 𝑤𝑖𝑡ℎ𝑖𝑛
This ratio is used to judge whether the difference among several sample means is
significant or is just a matter of sampling fluctuations. For this purpose, we look into the table*,
giving the values of F for given degrees of freedom at different levels of significance. If the
worked-out value of F, as stated above, is less than the table value of F, the difference is taken as
insignificant i.e., due to chance and the null-hypothesis of no difference between sample means
stands. In case the calculated value of F happens to be either equal or more than its table value, the
difference is considered as significant (which means the samples could not have come from the
same universe) and accordingly the conclusion may be drawn. The higher the calculated value of
F is above the table value, the more definite and surer one can be about his conclusions.
Note:
i) It should be remembered that ANOVA test is always a one-tailed test, since a low calculated
value of F from the sample data would mean that the fit of the sample means to the null hypothesis
̅1 = 𝑋
(viz., 𝑋 ̅2 = ⋯ = 𝑋 ̅ k ) is a very good fit.

Table 1: Analysis of Variance for one-way ANOVA Technique (there are k sample having n items)
Problems
1. Set up an analysis of variance table for the following per acre production data for three varieties
of wheat, each grown on 4 plots and state if the variety differences are significant.

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 5 | 24
 RV Institute of Technology & Management ®

Solution: First we calculate the mean of each of these samples:

6+7+3+8
̅̅̅1 =
𝑋 =6
4
5+5+3+7
̅̅̅2 =
𝑋 =5
4
5+4+3+4
̅̅̅3 =
𝑋 =4
4
𝑋̅1 +𝑋̅2 +𝑋̅3 6+5+4
Mean of Samples means, 𝑋̿ = = =5
𝑘 3

Now we work out SS between and SS within samples:

2 2 2
SS between = 𝑛1 (𝑋̿ − 𝑋
̅1 ) + 𝑛2 (𝑋̿ − 𝑋
̅ 2 ) + 𝑛3 (𝑋̿ − 𝑋
̅3)

= 4(5 − 6)2 + 4(5 − 5)2 + 4(5 − 4)2 = 4 + 0 + 4 = 8

SS within = ∑(𝑋1𝑖 − 𝑋̅1 )2 + ∑(𝑋2𝑖 − 𝑋̅2 )2 + ∑(𝑋3𝑖 − 𝑋̅3 )2
= {(6 − 6)2 + (7 − 6)2 + (3 − 6)2 + (8 − 6)2 }
+ {(5 − 5)2 + (5 − 5)2 + (3 − 5)2 + (7 − 5)2 }
+ {(5 − 4)2 + (4 − 4)2 + (3 − 4)2 + (4 − 4)2 }
= {0 + 1 + 9 + 4} + {0 + 0 + 4 + 4} + {1 + 0 + 1 + 0}
= 14 + 8 + 2 = 24
2
𝑆𝑆 𝑓𝑜𝑟 𝑡𝑜𝑡𝑎𝑙 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = ∑(𝑋𝑖𝑗 − 𝑋̿)

= (6 − 5)2 + (7 − 5)2 + (3 − 5)2 + (8 − 5)2 + (5 − 5)2 + (3 − 5)2

+ (7 − 5)2 + (5 − 5)2 + (4 − 5)2 + (3 − 5)2 + (4 − 5)2
= 1 + 4 + 4 + 9 + 0 + 0 + 4 + 4 + 0 + 1 + 4 + 1 = 32
Alternatively, it (SS for total variance) can also be worked out as,
SS for total = SS between + SS within=8+24=32

We can now set up the ANOVA table for this problem

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 6 | 24
 RV Institute of Technology & Management ®

Source of SS d.f MS F-ratio 5%

variation F-limit(from
the F-table)
Between 8 (3-1)=2 8/2=4.00 4.00/2.67=1.5 F(1,2)=4.26
sample
Within 24 12-3=9 24/9=2.67 4.00/2.67=1.5 F(1,2)=4.26
sample
Total 32 12-1=11

The above table shows that the calculated value of F is 1.5 which is less than the table value of
4.26 at 5% level with d.f. being v1 = 2 and v2 = 9 and hence could have arisen due to chance. This
analysis supports the null-hypothesis of no difference is sample means. We may, therefore,
conclude that the difference in wheat output due to varieties is insignificant and is just a matter of
chance.

2. Three different kinds of food are tested on three groups of rats for 5 weeks. The objective is to
check the difference in mean weight (in grams) of the rats per week. Apply one-way ANOVA
using a 0.05 significance level to the following data:
Food I Food II Food III
8 4 11
12 5 8
19 4 7
8 6 13
6 9 7
11 7 9

Solution: Using the same procedure as explained in problem1 we set null hypothesis as
H0: μ1= μ2=μ3
H1: The means are not equal
The other computed values as follows
Since, 𝑋̅1 = 5, 𝑋̅2 = 9, 𝑋̅3 = 10
Total mean = 𝑋̅= 8

SSB = 6(5 – 8)2 + 6(9 – 8)2 + 6(10 – 8)2 = 84

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 7 | 24
 RV Institute of Technology & Management ®

SSE = 68

MSB = SSB/df1 = 42
MSE = SSE/df2 = 4.53
f = MSB/MSE = 42/4.53 = 9.33
Source of SS d.f MS F-ratio 5%
variation F-limit(from
the F-table)
Between 8 (3-1)=2 8/2=4.00 4.00/2.67=1.5 F(1,2)=4.26
sample
Within 24 12-3=9 24/9=2.67 4.00/2.67=1.5 F(1,2)=4.26
sample
Total 32 12-1=11

Since f > F, the null hypothesis stands rejected

3. Four brands of flashlight batteries are to be compared by testing each brand in five flashlights.
Twenty flashlights are randomly selected and divided randomly into four groups of five
flashlights each. Then each group of flashlights uses a different brand of battery. The lifetimes
of the batteries, to the nearest hour, are as follows.

Brand A Brand B Brand C Brand D

42 28 24 20
30 36 36 32
39 31 28 38
28 32 28 28
29 27 33 25

Preliminary data analyses indicate that the independent samples come from normal populations
with equal standard deviations. At the 5% significance level, does there appear to be a difference
in mean lifetime among the four brands of batteries?

Solution: H0: μ1= μ2=μ3

H1: The means are not equal
Significance level 𝛼 = 0.05
SS total = 560.2
SS between samples = 68.2
SS within samples = SS total -SS between samples=492.0

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 8 | 24
 RV Institute of Technology & Management ®

Source of SS d.f MS F-ratio 5%

variation F-limit(from
the F-table)
Between 68.2 3 22.7333 0.7393 3.24
sample
Within 492.0 16 30.75
sample
Total 560.2 19

At the 𝛼 = 0.05 level of significance, there is not enough evidence to conclude that the mean
lifetimes of the brands of batteries differ, thus it is failed to reject the null hypothesis.

4. Data on Scholastic Aptitude Test (SAT) scores are published by the College Entrance
Examination board in National College-Bound Senior. SAT scores for randomly selected
students from each of four high-school rank categories are displayed in the following table.
Top Tenth Second Tenth Secon-fifth Third fifth
528 514 649 372
586 457 506 440
680 521 556 495
718 370 413 321
532 470 424
330
Construct the one-way ANOVA table for the data. Compute SSC and SSE using the defining
formulas.
̅̅̅1 = 628.0, 𝑋
Solution: 𝑋 ̅̅̅2 = 478.8, 𝑋
̅̅̅3 = 518.8, 𝑋
̅̅̅4 = 397.0, 𝑋̿ = 494.1
SS between samples (SSC) = 132508.2
SS Within samples (SSE) = 95877.6 and SS total = SSC+SSE=228385.8
Source of SS d.f MS F-ratio
variation
Between 132508.2 3 44169.40 7.37
sample
Within 95877.6 16 5992.35
sample

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 9 | 24
 RV Institute of Technology & Management ®

Total 560.2 19

Exercise
1. Manufacturers of golf balls always seem to be claiming that their ball goes the farthest. A
writer for a sports magazine decided to conduct an impartial test. She randomly selected 20
golf professionals and then randomly assigned four golfers to each of five brands. Each golfer
drove the assigned brand of ball. The driving distances, in yards, are displayed in the following
table.
Brand 1 Brand 2 Brand 3 Brand 4 Brand 5
286 279 270 284 281
276 277 262 271 293
281 284 277 269 276
274 288 280 275 292

Preliminary data analyses indicate that the independent samples come from normal populations
with equal standard deviations. Do the data provide sufficient evidence to conclude that a
difference exists in mean weekly earnings among nonsupervisory workers in the five industries?
Perform the required hypothesis test using α = 0.05.

2. The U.S. Bureau of Prisons publishes data in Statistical Report on the times served by prisoners
released from federal institutions for the first time. Independent random samples of released
prisoners for five different offense categories yielded the following information on time served,
in months. At the 1% significance level, do the data provide sufficient evidence to conclude
that a difference exists in mean time served by prisoners among the five offense groups?

Major
𝑛𝑖 𝑋𝑖 𝑠𝑖
Counterfeiting 15 14.5 4.5
Drug Laws 17 18.4 3.8
Firearms 12 18.2 4.5
Forgery 10 15.6 3.6
Fraud 11 11.5 4.7

TWO-WAY ANOVA
Two-way ANOVA technique is used when the data are classified on the basis of two factors. For
example, the agricultural output may be classified on the basis of different varieties of seeds and
also on the basis of different varieties of fertilizers used. A business firm may have its sales data
classified on the basis of different salesmen and also on the basis of sales in different regions. In a
factory, the various units of a product produced during a certain period may be classified on the
basis of different varieties of machines used and also on the basis of different grades of labour.
Such a two-way design may have repeated measurements of each factor or may not have repeated

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 10 | 24
 RV Institute of Technology & Management ®

values. The ANOVA technique is little different in case of repeated measurements where we also
compute the interaction variation. We shall now discuss the two-way ANOVA technique in the
context of both the said designs with the help of examples.

(a) ANOVA technique in context of two-way design when repeated values are not there: As
we do not have repeated values, we cannot directly compute the sum of squares within samples as
we had done in the case of one-way ANOVA. Therefore, we have to calculate this residual or error
variation by subtraction, once we have calculated (just on the same lines as we did in the case of
one-way ANOVA) the sum of squares for total variance and for variance between varieties of one
treatment as also for variance between varieties of the other treatment.
The various steps involved are as follows:

(i) Take the total of the values of individual items in all the samples and call it T.
𝑇2
(ii) Work out the correction factor as under correction factor = 𝑛

(iii) Find out the square of all the item values (or their coded values as the case may be) one by
one and then take its total. Subtract the correction factor from this total to obtain the sum of squares
2 𝑇2
of deviations for total variance. Total SS = ∑ 𝑋𝑖𝑗 − 𝑛

(v) Take the total of different columns and then obtain the square of each column total and divide
such squared values of each column by the number of items in the concerning column and take the
total of the result thus obtained. Finally, subtract the correction factor from this total to obtain the
sum of squares of deviations for variance between columns or (SS between columns).
(vi) Take the total of different rows and then obtain the square of each row total and divide such
squared values of each row by the number of items in the corresponding row and take the total of
the result thus obtained. Finally, subtract the correction factor from this total to obtain the sum of
squares of deviations for variance between rows (or SS between rows).
(vii) Sum of squares of deviations for residual or error variance can be worked out by subtracting
the result of the sum of 5th and 6th steps from the result of 4th step stated above. In other words,
Total SS – (SS between columns + SS between rows) = SS for residual or error variance.
(viii) Degrees of freedom (d.f.) can be worked out as under:
d.f. for total variance = (c . r – 1)
d.f. for variance between columns = (c – 1)
d.f. for variance between rows = (r – 1)
d.f. for residual variance = (c – 1) (r-1) where c = number or columns, r = number of rows
ix) ANOVA table cab be setup in the usual fashion as shown below

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 11 | 24
 RV Institute of Technology & Management ®

Table 2: Analysis of variance for Two-way ANOVA

In the table c = number of columns, r = number of rows and SS residual = Total SS – (SS between
columns + SS between rows).
Thus, MS residual or the residual variance provides the basis for the F-ratios concerning variation
between columns treatment and between rows treatment. MS residual is always due to the
fluctuations of sampling, and hence serves as the basis for the significance test. Both the F-ratios
are compared with their corresponding table values, for given degrees of freedom at a specified
level of significance, as usual and if it is found that the calculated F-ratio concerning variation
between columns is equal to or greater than its table value, then the difference among columns
means is considered significant. Similarly, the F-ratio concerning variation between rows can be
interpreted.
(b) ANOVA technique in context of two-way design when repeated values are not there:
In case of a two-way design with repeated measurements for all of the categories, we can
obtain a separate independent measure of inherent or smallest variations. For this measure we can
calculate the sum of squares and degrees of freedom in the same way as we had worked out the
sum of squares for variance within samples in the case of one-way ANOVA. Total SS, SS between
columns and SS between rows can also be worked out as stated above. We then find left-over sums
of squares and left-over degrees of freedom which are used for what is known as ‘interaction
variation’ (Interaction is the measure of inter relationship among the two different classifications).
After making all these computations, ANOVA table can be set up for drawing inferences.
Problems
1. Set up an analysis of variance table for the following two-way design results:
Per Acre Production Data of Wheat

Also state whether variety differences are significant at 5% level.

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 12 | 24
 RV Institute of Technology & Management ®

Solution: As the given problem is a two-way design of experiment without repeated values, we
shall adopt all the above stated steps.
𝑇2 60×60
Step (i) T=60, n=12, ∴ 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 = 𝑛
= 12
= 300
(60× 60)
Step (ii) 𝑇𝑜𝑡𝑎𝑙 𝑆𝑆 = (36 + 25 + 25 + 49 + 25 + 16 + 9 + 9 + 64 + 49 + 16) − 12
24×24 20×20 16×16 60×60
Step (iii) SS between columns treatment = [ 4
+ 4
+ 4
]−[ 12
] = 144 + 100 +

64 − 300 = 8
16×16 16×16 9×9 19×19 60×60
Step (iv) SS between rows treatment = [ 3
+ 3
+ 3
+ 3
]−[ 12
] = 85.33 +

85.33 + 27.00 + 120.33 − 300 = 18

Step (v) SS residual or error = Total SS-(SS between columns + SS between rows) = 33-(8+18) =
6.
Setting up the ANOVA table
Source of variation SS d.f MS F-ratio 5%
F-limit(from
the F-table)
Between columns (i.e., 8 (3-1)=2 8/2=4.00 4/1=4 F(2,6)=5.14
between varieties of
seeds)
Between rows(i.e., 18 (4-1)=3 18/3=6 6/1=6 F(3,6)=4.76
between varieties of
fertilizers)
Residual or error 6 (3 − 1) × (4 − 1) = 6/6=1
6
Total 32 (3 × 4) − 1 = 11

From the said ANOVA table, we find that differences concerning varieties of seeds are
insignificant at 5% level as the calculated F-ratio of 4 is less than the table value of 5.14, but the
variety differences concerning fertilizers are significant as the calculated F-ratio of 6 is more than
its table value of 4.76.
2. Set up ANOVA table for the following information relating to three drugs testing to judge the
effectiveness in reducing blood pressure for three different groups of people:
Amount of Blood Pressure Reduction in Millimeters of Mercury

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 13 | 24
 RV Institute of Technology & Management ®

Do the drugs act differently? Are the different groups of people affected differently? Is the
interaction term significant? Justify your answer at 5% level of significance.
Solution: As the given problem is a two-way design of experiment with repeated values, we shall
adopt all the above stated steps.
Step (i) T=187, n=18,
187×187
𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 = 18 = 1942.7
2 2 2 2 2 2 2 2 2 2
Step (ii) 𝑇𝑜𝑡𝑎𝑙 𝑆𝑆 = (14 + 15 + 12 +11 + 10 + 9 + 72 + 8 + 11 + 11 + 11 +
(1872 )
112 + 102 + 112 + 82 + 72 ) − 18
= 2019 − 1942.72 = 76.28
73×73 56×56 58×58 1872
Step (iii) SS between columns (i.e., between drugs) = [ 6
+ 6
+ 6
]−[ 18
]=

888.16 + 522.66 + 560.67 − 1942.72 = 28.77

70×70 59×59 58×58 1872
Step (iv) SS between rows (i.e., between people) = [ 6
+
6
+
6
]−[
18
] = 816.67 +

580.16 + 560.67 − 1942.72 = 14.78

2 2 2 2
Step (v) SS within samples=(14 − 14.5) + (15 − 14.5) + (10 − 9.5) + (9 − 9.5) + (11 −
11)2 + (11 − 11)2 + (12 − 11.5)2 + (7 − 7.5)2 + (8 − 7.5)2 + (10 − 10.5)2 + (11 −
10.5)2 + (10 − 10.5)2 + (11 − 10.5)2 + (11 − 11)2 + (11 − 11)2 + (8 − 7.5)2 + (7 −
7.5)2 = 3.50
Step(vi) SS for interaction variation = Total SS-(SS between columns + SS between rows) = 29.33
Setting up the ANOVA table
Source of SS d.f MS F-ratio 5%
F-limit(from
variation
the F-table)
Between 28.77 3-1=2 28.77/2=14.385 14.385/0.389=36.9 F(2,9)=4.26
columns (i.e.,
between drugs)
Between 14.78 3-1=2 14.78/2=7.390 7.390/0.389=19.0 F(2,9)=4.26
rows(i.e.,
between people)

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 14 | 24
 RV Institute of Technology & Management ®

Interaction 29.33 4 29.33/4 7.308/0.389 F(4,9)=3.63

Within 3.50 18-9=9 3.50/9=0.389
samples(error)
Total 76.28 18-1=17

(NOTE: These figures are left-over figures and have been obtained by subtracting from the
column total the total of all other value in the said column. Thus, interaction SS = (76.28) – (28.77
+ 14.78 + 3.50) = 29.23 and interaction degrees of freedom = (17) – (2 + 2 + 9) = 4).

The above table shows that all the three F-ratios are significant of 5% level which means that the
drugs act differently, different groups of people are affected differently and the interaction term is
significant. In fact, if the interaction term happens to be significant, it is pointless to talk about the
differences between various treatments i.e., differences between drugs or differences between
groups of people in the given case.

3. The following data show the number of worms quarantined from the GI areas of four groups
of muskrats in a carbon tetrachloride anthelmintic study. Conduct a two-way ANOVA test.
I II III IV
338 412 124 389
324 387 353 432
268 400 469 255
147 233 222 133
309 212 111 265
Solution: Using the same procedure as explained in problem1
Source of Sum of Squares Degrees of Mean Square Ratio F
variation freedom
Between the 62111.6 8 9078.067 F = MST / MSE
groups = 9.4062 / 3.66
F = 2.57
Within the groups 98787.8 16 4567.89
Total 167771.4 24

Since F = MST / MSE

= 9.4062 / 3.66 F = 2.57
Exercise:

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 15 | 24
 RV Institute of Technology & Management ®

1. The following data represents the number of units of tablet production (in thousands) per day
by five different technicians by using four different types of machines.
Workers A B C D
P 54 48 57 46
Q 56 50 62 53
R 44 46 54 42
S 53 48 56 44
T 48 52 59 48
a) Test whether the mean productivity of different machines is same?
b) Test whether the 5 technicians differ with respect to mean productivity?

ANOVA IN LATIN-SQUARE DESIGN

Latin-square design is an experimental design used frequently in agricultural research. In such a
design the treatments are so allocated among the plots that no treatment occurs, more than once in
any one row or any one column. The ANOVA technique in case of Latin-square design remains
more or less the same as we have already stated in case of a two-way design, except the fact that
the variance is split into four parts as under:
(i) variance between columns;
(ii) variance between rows;
(iii) variance between varieties;
(iv) residual variance.
All these above stated variances are worked out as under:

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 16 | 24
 RV Institute of Technology & Management ®

Note:
1. CODING METHOD: Coding method is furtherance of the short-cut method. This is based on
an important property of F-ratio that its value does not change if all the n item values are either
multiplied or divided by a common figure or if a common figure is either added or subtracted from
each of the given n item values. Through this method big figures are reduced in magnitude by
division or subtraction and computation work is simplified without any disturbance on the F-ratio.
This method should be used specially when given figures are big or otherwise inconvenient. Once
the given figures are converted with the help of some common figure, then all the steps of the
short-cut method stated above can be adopted for obtaining and interpreting F-ratio.
2. In place of c we can as well write r or v since in Latin-square design c = r = v

Problems
1. Analyze and interpret the following statistics concerning output of wheat per field obtained as
a result of experiment conducted to test four varieties of wheat viz., A, B, C and D under a
Latin-square design

Solution: Using the coding method, we subtract 20 from the figures given in each of the small
squares and obtain the coded figures as under:

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 17 | 24
 RV Institute of Technology & Management ®

Squaring these coded figures in various columns and rows we have:

𝑇2 (−12)×(−12)
Step (i) T=60, n=12, ∴ 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 = 𝑛
= 16
=9
2 𝑇2
Step (ii) 𝑇𝑜𝑡𝑎𝑙 𝑆𝑆 = ∑(𝑋𝑖𝑗 ) − 𝑛
= 122 − 9 = 113
2
(𝑇𝑗 ) 𝑇2 66
Step (iii) SS between columns = ∑ 𝑛𝑗
− 𝑛
= 4
− 9 = 7.5

(𝑇𝑖 )2 𝑇2 222
Step (iv) SS between rows treatment = ∑ 𝑛𝑖
− 𝑛
= 4
− 9 = 46.5

Step (v) For finding SS for variance between varieties, we would first rearrange the coded data in
the following form:

Now we can work out SS for variance between varieties as under:

𝑇𝑣2 𝑇2 (−12)2 12 62 (−7)2
SS for variance between varieties=∑ 𝑛𝑣 − 𝑛
={ 4
+ 4
+ 4
+ 4
} − 9 = 48.5

Sum of square of residual variance will work out to, 113-(7.5+46.5+48.5) =10.50
d.f. for variance between columns = (c – 1) = (4 – 1) = 3
d.f. for variance between rows = (r – 1) = (4 – 1) = 3
d.f. for variance between varieties = (v – 1) = (4 – 1) = 3
d.f. for total variance = (n – 1) = (16 – 1) = 15
d.f. for residual variance = (c – 1) (c – 2) = (4 – 1) (4 – 2) = 6
ANOVA table in Latin-Square design can now be set up as shown below

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 18 | 24
 RV Institute of Technology & Management ®

Source of variation SS d.f MS F-ratio 5%

F-limit(from the F-
table)
Between columns 7.50 3 7.50/3=2.50 2.50/1.75=1.43 F(3,6)=4.76
Between rows 46.50 3 46.50/3=15.50 15.50/1.75=8.85 F(3,6)=4.76
Between varieties 48.50 3 48.50=16.17 16.17/1.75=9.24 F(3,6)=4.76
Residual or error 10.50 6 10.50/6=1.75
Total 113.00 15

The above table shows that variance between rows and variance between varieties are significant
and not due to chance factor at 5% level of significance as the calculated values of the said two
variances are 8.85 and 9.24 respectively which are greater than the table value of 4.76. But variance
between columns is insignificant and is due to chance because the calculated value of 1.43 is less
than the table value of 4.76.
2. Below are given the plan and yield in kgs/plot of a 5x5 Latin square experiment on the wheat
crop carried out for testing the effects of five, manorial treatments A, B, C, D, and E. ‘A’ denotes
control.

Analyze the data and state your conclusions.

Solution:
𝑇2
Step (i) ∴ 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 = 𝑛
= 6177.96
2 𝑇2
Step (ii) 𝑇𝑜𝑡𝑎𝑙 𝑆𝑆 = ∑(𝑋𝑖𝑗 ) − 𝑛
= 483.04
2
(𝑇𝑗 ) 𝑇2
Step (iii) SS between columns (SSC) = ∑ 𝑛𝑗
− 𝑛
= 14.24

(𝑇𝑖 )2 𝑇2
Step (iv) SS between rows treatment (SSR) = ∑ 𝑛𝑖
− 𝑛
= 3.04

Step (v) To get SS due to treatments, first find the totals for each treatment using the given data
as follows:

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 19 | 24
 RV Institute of Technology & Management ®

𝑇𝑣2 𝑇2
SS for variance between treatments=∑ 𝑛𝑣 − 𝑛
= 454.64

Sum of square of residual variance(error)= TSS-SSR-SSC-SST=11.12

ANOVA table in Latin-Square design can now be set up as shown below
Source of SS d.f MS F-ratio 5%
variation F-limit(from the F-table)
Between 4 3.04 0.76 3.26
columns
Between 4 14.24 3.56 5.41
rows
Between 4 454.24 113.66 123.34
varieties
Residual or 12 0.92
11.12
error
Total 24 484.04
The observed highly significant value of the variance ratio indicates that there are significant
differences between the treatment means. S.E. of the difference between the treatment means is
𝐸𝑀𝑆 0.92
given by 𝑆𝐸𝐷 = √2 × 𝑟
= √2 × 5
= 0.61 and Critical difference=SED * t 5% at df=1.33

Summary of results
Treatment means will be calculated from the original table on treatment totals.
Treatments A B C D E CD 5%
Mean yield 9.0 13.6 17.6 21.8 16.6 1.33
in Kgs/plot
The treatment has been compared by setting them in the descending order of their yields.
Treatments D C E B A CD 5%

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 20 | 24
 RV Institute of Technology & Management ®

Mean yield 21.8 17.6 16.6 13.6 9.0 1.33

in Kgs/plot
The treatment ‘D’ is the best of all. The treatments ‘C’ and ‘E’ do not differ significantly each
other. The yield obtained by applying every one of the manorial treatments is significantly higher
that obtained without applying any manure.
Analysis of co-variance (ANOCOVA)
The object of experimental design in general happens to be to ensure that the results observed may
be attributed to the treatment variable and to no other causal circumstances. For instance, the
researcher studying one independent variable, X, may wish to control the influence of some
uncontrolled variable (sometimes called the covariate or the concomitant variables), Z, which is
known to be correlated with the dependent variable, Y, then he should use the technique of analysis
of covariance for a valid evaluation of the outcome of the experiment. “In psychology and
education primary interest in the analysis of covariance rests in its use as a procedure for the
statistical control of an uncontrolled variable.

While applying the ANOCOVA technique, the influence of uncontrolled variable is usually
removed by simple linear regression method and the residual sums of squares are used to provide
variance estimates which in turn are used to make tests of significance. In other words, covariance
analysis consists in subtracting from each individual score (Yi) that portion of it Yi´ that is
predictable from uncontrolled variable (Zi) and then computing the usual analysis of variance on
the resulting (Y – Y´)’s, of course making the due adjustment to the degrees of freedom because
of the fact that estimation using regression method required loss of degrees of freedom.

ASSUMPTIONS IN ANOCOVA
The ANOCOVA technique requires one to assume that there is some sort of relationship between
the dependent variable and the uncontrolled variable. We also assume that this form of relationship
is the same in the various treatment groups. Other assumptions are:
(i) Various treatment groups are selected at random from the population.
(ii) The groups are homogeneous in variability.
(iii) The regression is linear and is same from group to group.

Problems
1. The following are paired observations for three experimental groups:

Y is the covariate (or concomitant) variable. Calculate the adjusted total, within groups and
between groups, sums of squares on X and test the significance of differences between the adjusted

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 21 | 24
 RV Institute of Technology & Management ®

means on X by using the appropriate F-ratio. Also calculate the adjusted means on X.

Solution: We apply the technique of analysis of covariance and work out the related measures as

(∑ 𝑋)2
Correction factor for 𝑋 = 𝑁
= 7616.27 and
(∑ 𝑌)2
∑ 𝑌 = 33 + 72 + 105 = 210 and correction factor for Y= = 2940
𝑁
∑𝑋∑𝑌
∑ 𝑋 2 = 9476,∑ 𝑌 2 = 3734, ∑ 𝑋𝑌 = 5838 and correction factor for XY= = 4732
𝑁

Hence, total SS for X = X2 - correction factor for X= 9476 – 7616.27 = 1859.73

492 1142 1752
SS between for X = { + + } − {𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 𝑓𝑜𝑟 𝑋}=480.2+2599.2+6125-
5 5 5

7671.27=1588.13
SS within X = (total SS for X) - (SS between for X) = 1859.73-1588.13 = 271.60
Similarly, we work out the following values in respect of Y
Total SS for Y= Y2 - correction factor for Y =3734-2940=794
332 722 1052
SS between for Y ={ 5
+ 5
+ 5
} −correction factor for Y=519.6

SS within for Y = (total SS for Y) – (SS between for Y) = (794) – (519.6) = 274.4
Then, we work out the following values in respect of both X and Y
Total sum of product of XY = ∑ 𝑋𝑌 – 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 𝑓𝑜𝑟 𝑋𝑌 = 5838 – 4732 = 1106
49×33 114×72 175×105
SS between for XY = { + + } − 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 𝑓𝑜𝑟 𝑋𝑌
5 5 5

= (323.4 + 1641.6 + 3675) – (4732) = 908

SS within for XY = (Total sum of product) – (SS between for XY) = (1106) – (908) = 198
ANOVA table for X, Y and XY can now be set up as shown below

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 22 | 24
 RV Institute of Technology & Management ®

𝑇2 11062
Adjusted total SS= 𝑇𝑋𝑋 − 𝑇𝑋𝑌 = 1859.73 − = 1859.73 − 1540.60 = 319.13
𝑌𝑌 794

𝐸2 1982
Adjusted SS within group= 𝐸𝑋𝑋 − 𝐸𝑋𝑌 = 271.60 − 274.40 = 128.73
𝑌𝑌

Adjusted SS between groups= (adjusted total SS) – (Adjusted SS within group) = (319.13 –
128.73) = 190.40
ANOVA table for adjusted X

At 5% level, the table value of F for v1 = 2 and v2 = 11 is 3.98 and at 1% level the table value of
F is 7.21. Both these values are less than the calculated value (i.e., calculated value of 8.14 is
greater than table values) and accordingly we infer that F-ratio is significant at both levels which
means the difference in group means is significant.
Adjusted means on X will be worked out as follows:
𝑆𝑢𝑚 𝑜𝑓 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑤𝑖𝑡ℎ𝑖𝑛 𝑔𝑟𝑜𝑢𝑝 198
Regression coefficient for X on Y i.e.,𝑏 = 𝑆𝑢𝑚 𝑜𝑓 𝑠𝑞𝑢𝑎𝑟𝑒𝑠 𝑤𝑖𝑡ℎ𝑖𝑛 𝑔𝑟𝑜𝑢𝑝𝑠 𝑓𝑜𝑟 𝑌 = 274.40 = 0.7216

Adjusted means of groups in X = (Final mean) – b (deviation of initial mean from general mean
in case of Y). Hence,
Adjusted mean for Group I = (9.80) – 0.7216 (–7.4) = 15.14
Adjusted mean for Group II = (22.80) – 0.7216 (0.40) = 22.51
Adjusted mean for Group III = (35.00) – 0.7216 (7.00) = 29.95

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 23 | 24
 RV Institute of Technology & Management ®

Video Links:
1. Design of experiments
2. Basic of ANOVA
3. Problems on one-way ANOVA
4. Problems on two-way ANOVA
5. Problems on two-way ANOVA
6. Problems on ANOVA for Latin-Square design
7. Problems on ANOCOVA

Disclaimer: The content provided is prepared by department of Mathematics for the specified
syllabus by using reference books mentioned in the syllabus. This material is specifically for the
use of RVITM students and for education purpose only.

III-Semester: Mathematics for Computer Science(MCS) (BCS301)

P a g e 24 | 24