Analysis of Variance and

Experimental Designs

Analysis of Variance and Experimental Designs 1

 Learning Objectives

Upon completion of this chapter, you will be able to:

 Understand the concept of ANOVA and experimental designs

 Compute and interpret the result of completely randomized design
(one- way ANOVA)
 Compute and interpret the result of randomized block design
 Compute and interpret the result of factorial design
(two-way ANOVA)

Analysis of Variance and Experimental Designs 2

Introduction to Experimental Designs

 An experimental design is the logical construction of an experiment

to test hypothesis in which the researcher either controls or
manipulates one or more variables.
 Independent variable: In an experimental design, the independent
variable may be either a treatment variable or a classification
variable.
 Treatment variable: This is a variable which is controlled or modified
by the researcher in the experiment. For example, in agriculture, the
different fertilizers or the different methods of cultivation are the
treatments.
 Classification variable: Classification variable can be defined as the
characteristics of the experimental subject that are present prior to
the experiment and not a result of the researcher’s manipulation or
control.

Analysis of Variance and Experimental Designs 3

Introduction to Experimental Designs

 Experimental Units: The smallest division of the experimental

material to which treatments are applied and observations are made
are referred to as experimental units.
 Dependent variable: In an experimental design, a dependent
variable is the response to the different levels of independent
variables. This is also called response variable.
 Factor: A factor can be referred to as a set of treatments of a single
type. In most situations, a researcher may be interested in studying
more than one factor.

Analysis of Variance and Experimental Designs 4

 Analysis of Variance

 Analysis of variance or ANOVA is a technique of testing hypotheses

about the significant difference in several population means.
 In analysis of variance, the total variation in the sample data can be
on account of two components, namely, variance between the
samples and variance within the samples.
 Variance between the samples is attributed to the difference among
the sample means. This variance is due to some assignable causes.
 Variance within the samples is the difference due to chance or
experimental errors.

Analysis of Variance and Experimental Designs 5

 Completely Randomized Design
(One-way Anova)

Completely randomized design contains only one independent variable,

with two or more treatment levels or classifications.

Analysis of Variance and Experimental Designs 6

 Figure 12.1: Partitioning the total sum of squares of the variation for
completely randomized design (one-way ANOVA)

SST (total sum of squares) = SSC (sum of squares between columns) + SSE (sum of squares within samples)

Analysis of Variance and Experimental Designs 7

 Applying the F -Test Statistic

 In case of ANOVA, F value is obtained by dividing the treatment

variance (MSC) by the error variance (MSE).
 F test statistic in one-way ANOVA
 The F test statistic follows F distribution with k – 1 degrees of
freedom corresponding to MSC in the numerator and n – k degrees
of freedom corresponding to MSE in the denominator.

Analysis of Variance and Experimental Designs 8

Figure 12.2: Rejection and non-rejection region (acceptance region) when
using ANOVA to test null hypothesis

Analysis of Variance and Experimental Designs 9

The ANOVA Summary Table

Analysis of Variance and Experimental Designs 10

 Example 12.1
Vishal Foods Ltd is a leading manufacturer of biscuits. The company
has launched a new brand in the four metros; Delhi, Mumbai,
Kolkata, and Chennai. After one month, the company realizes that
there is a difference in the retail price per pack of biscuits across
cities. Before the launch, the company had promised its employees
and newly-appointed retailers that the biscuits would be sold at a
uniform price in the country. The difference in price can tarnish the
image of the company. In order to make a quick inference, the
company collected data about the price from six randomly selected
stores across the four cities. Based on the sample information, the
price per pack of the biscuits (in rupees) is given in Table 12.5:

Analysis of Variance and Experimental Designs 11

 Example 12.1: Continued

Use one-way ANOVA to analyse the significant difference in the prices. Take 95% as the
confidence level.

Analysis of Variance and Experimental Designs 12

Table 12.7: ANOVA table for Example 12.1

Analysis of Variance and Experimental Designs 13

 Using MS Excel, Minitab and SPSS
for Hypothesis Testing with F Statistic
for Difference in Means of More Than
Two Populations

 Solved Examples\Excel\Ex 12.1.xls

 Solved Examples\Minitab\Ex 12.1.MPJ
 Ch 12 Solved Examples\SPSS\Ex 12.1.sav
 Ch 12 Solved Examples\SPSS\Output 12.1.spv

Analysis of Variance and Experimental Designs 14

 Randomized Block Design

 Like the completely randomized design, the randomized block

design also focuses on one independent variable of interest
(treatment variable).
 Additionally, in randomized block design, we also include one
more variable referred to as “blocking variable.”
 Blocking variable is a variable which a researcher wants to control
but is not a treatment variable of interest.

Analysis of Variance and Experimental Designs 15

Figure 12.11: Partitioning the SSE in randomized block design

Analysis of Variance and Experimental Designs 16

Null and Alternative Hypotheses in
a Randomized Block Design

Analysis of Variance and Experimental Designs 17

ANOVA Summary Table for
Two-Way Classification

Analysis of Variance and Experimental Designs 18

 Example 12.2

A company which produces stationary items wants to diversify into

the photocopy paper manufacturing business. The company has
decided to first test market the product in three areas termed as
the north area, central area, and the south area. The company
takes a random sample of five salesmen S1, S2, S3, S4, and S5 for
this purpose. The sales volume generated by these five salesmen
(in thousand rupees) and total sales in different regions is given in
Table 12.9:

Analysis of Variance and Experimental Designs 19

Example 12.1: Continued

Use a randomized block design analysis to examine:

(1) Whether the salesmen significantly differ in performance?
(2) Whether there is a significant difference in terms of sales capacity
between the regions?
Take 95% as confidence level for testing the hypotheses.

Analysis of Variance and Experimental Designs 20

Table 12.11: ANOVA Summary table for Example 12.2

Analysis of Variance and Experimental Designs 21

 Using MS Excel and Minitab for
Hypothesis Testing with F Statistic in
Randomized Block Design

 Solved Examples\Excel\Ex 12.2.xls

 Solved Examples\Minitab\Ex 12.2.MPJ

Analysis of Variance and Experimental Designs 22

Factorial Design (Two-Way Anova)

 In some real-life situations, a researcher has to explore two or

more treatments simultaneously. This type of experimental
design is referred to as factorial design.
 Factorial design provides an opportunity to study the interaction
effect of two treatment variables.
 Randomized block design does not provide the opportunity to
study the interaction effect of treatment and block. This facility is
available only in factorial design.

Analysis of Variance and Experimental Designs 23

Null and Alternative Hypotheses
in a Factorial Design

Analysis of Variance and Experimental Designs 24

ANOVA Summary Table for
Two-Way ANOVA

Analysis of Variance and Experimental Designs 25

 Example 12.2

Chhattisgarh Steel and Iron Mills is a leading steel rod manufacturing

company of Chhattisgarh. The company produces 8-metre long steel
rods, which are used in the construction of buildings. The company
has four machines which manufacture steel rods in three shifts. The
company’s quality control officer wants to test whether there is any
difference in the average length of the iron rods by shifts or by
machines. Data given in Table 12.13 is organized by machines and
shifts obtained through a random sampling process. Employ a two-
way analysis of variance and determine whether there are any
significant differences in effects. Take alpha = 0.05.

Analysis of Variance and Experimental Designs 26

Table 12.13: Length of the iron rod in different shifts and produced
by different machines

Analysis of Variance and Experimental Designs 27

Table 12.15: ANOVA Summary table for Example 12.3

Analysis of Variance and Experimental Designs 28

 Using MS Excel and Minitab for
Hypothesis Testing with F Statistic
in a Factorial Design

 Solved Examples\Excel\Ex 12.3.xls

 Solved Examples\Minitab\Ex 12.3.MPJ

Analysis of Variance and Experimental Designs 29