ANALYSIS

OF
VARIANCE
ANOVA

INTRODUCTION
ANOVA is used for examining the
differences in the mean value of
dependent variables associated with the
effect of the controlled independent
variables , after taking into account the
influence of uncontrolled independent
variables
Essentially ANOVA is used as a test of
means for two or more populations
The null hypothesis typically is that all
means are equal

INTRODUCTION
E.g.,- Suppose the researcher
was interested in examining
whether heavy, medium, light
and non users of cereals differed
in their preference for Total
cereal measured on a 9 point
scale. The null hypothesis that
the four groups were not
different in preference for Total
could be tested using ANOVA

INTRODUCTION
In its simplest form ANOVA must have a
dependent variable ( preference for Total
cereal) that is metric ( measured using an
interval or ratio scale)
There must be also be one or more
independent variable (product use: heavy,
medium, light and non users). The
independent variables must be all categorical
(non metric). Categorical independent
variables are called factors.
A particular combination of factor levels or
categories is called a treatment.

ONE WAY ANOVA Some

E.g.,
A marketing researchers are often
interested in examining the difference in
the mean value of the dependent variable
for several categories of a single
independent variable or factor .E.g. are
Do the various segments differ in terms of their
volume of product consumption?
Do the brand evaluations of groups exposed to
different commercials vary?
Do retailers , wholesalers and agents differ in
their attitudes towards and the firms
distribution policies?

ONE WAY ANOVA Some

E.g.,
The production level in three shifts in a factory can
be compared to answer of the question of types
Is the production level higher/lower on any day of the
week?
Is Wednesdays morning shifts production better/worse
than any other shift?

Effectiveness of different promotional devices in

terms of sales
Quality of product produced by different
manufacturers in terms of an attribute
Production volume in different shifts in a factory
Yields from plots due to varieties of seeds,
fertilizers & cultivation method

INTRODUCTION
One way ANOVA involves only one categorical
variable or a single factor .The difference in
preference of heavy, medium , light and non
users would be examined by one way ANOVA
In one way analysis the treatment is same as a
factor level ( medium users constitute a
treatment).
If two or more factors are involved the analysis
is termed n way analysis of variance. E.g., if in
addition to product use the researcher also
wanted to examine the preference for Total
cereal of customers who are loyal and those
who are not ,an n way ANOVA would be
conducted

INTRODUCTION
If set of independent variables consists
of both categorical and metric variables
the technique is called ANALYSIS OF
COVARIANCE (ANCOVA).
E.g., ANCOVA would be required if the
researcher wanted the preference of
product use groups and loyalty groups
taking into account the respondents
attitudes towards nutrition and the
importance they attached to breakfast
as a meal. The latter two would be
measured on a 9 point Likert scale

INTRODUCTION
In this case the categorical
independent variables ( product
use and brand loyalty) are still
referred to as factors whereas
the metric independent variables
( attitude towards nutrition and
importance attached to
breakfast) are referred to as
covariates

STATISTICS ASSOCIATED WITH ONE

WAY ANOVA
Eta ()-The strength of effects of
X(independent variable) on
Y(dependent variable) is measured by
eta () . This value is between 0 and 1
F statistic- The null hypothesis that the
category means are equal in the
population is tested by an F statistic
based on the ratio of mean square
related to X and mean square related
to error

STATISTICS ASSOCIATED WITH ONE

WAY ANOVA
Mean Square The mean square
is the sum of squares divided by
the appropriate degree of
freedom
SS between or SSx- This is the
variation in Y related to the
variation in the means of the
categories of X. This represents
variation between the categories
of X or the portion of the sum of

STATISTICS ASSOCIATED WITH ONE

WAY ANOVA
SSwithin or SSerror - This is the
variation in Y due to variation
the variation within each
categories of X. This variation is
not accounted for by X
SSy The total variation in Y is
SSy

Assumptions in ANOVA
1. Ordinarily the categories of
independent variable are assumed to be
fixed. This is referred to fixed effect
model. In random effect model the
categories or treatment are considered
to be random sample from universe of
treatments
2. The error term is normally distributed
with zero mean and constant variation
3.The error terms are uncorrelated.

Short cut method for

ANOVA

Short cut method for ANOVA

The values of SSx and SSerror can be
calculated by applying short cut method
Step No 1- Arrange the data in a table(as
shown in next slide)
Step No 2- Calculate the grand total of all
samples T
T = X1 + X2+. +Xn
Step No 3- Calculate the correction factor
CF = T / n ; n= n1+n2 +n3+ + nr

Short cut method for ANOVA

Obser
vation
s
1
2
..
4
SUM

Sample Sample
No1
No 2
x11
x12
..
xk1
x1

x12
x22
..
xk2
x2

Sample
No r
..
..

x1r
x2r
..
xkr
xr

Short cut method for ANOVA

Step No4 - Find the sum of the
squares of all the observations in
samples from each of r samples
and subtract CF from this sum to
obtain the total sum of the
squares of deviations
SSy = X1 + X2+. +Xn CF
SSx= (xj) - CF
SSerror = SSy - SSx

Short cut method for ANOVA

Step No 5- Calculate MSx and
MSerror by
MSx=SSX/df1
;
MSy=SSy/df2
[df1=r-1; df2=n-r]
Make ANOVA Table and calculate
F statistic
=MSx/MS
error
Source
of Sum of Degree of Mean
Test
variation

square freedom
s

squares(MS statistic
)

Between
samples

SSx

df1

MSx =
SSx/df1

Within
sample

SSerror

df2

MSerror =
SSerror/df2

F=
MSx/MSerror

Short cut method for ANOVA

Step No6 Compare with critical
value of F for df1 and df2 at .If
Fcalculated is less than Fcritical then
accept hypothesis , otherwise
reject
NOTE
We will use the short cut method
for calculation for carrying out
the ANOVA test

Thank You