You are on page 1of 3

NOTE 8

ANOVA is designed specifically to test the differences in population means; it requires

examining the variances in the samples used.

More specifically, the procedure can be used to determine if a particular treatment when
applied to a population will have a significant impact on its mean.

Consider, as an example, the desire to measure the relative effects of three different
training programs on employee output. These different types of training may be (1) self-
taught, (2) computer instructed, or (3) taught by a supervisor.

In ANOVA study, the experimental units are the objects receiving the treatment, in our
training example; the employees constitute the experimental units. The factor is force or
variable whose impact on these experimental units we wish to measure. In this case,
training is the factor of interest. Finally, the three types of training constitute the
treatments, or factor levels, of the factor training.

The most common application of ANOVA is the completely randomized design. The
experimental units are assigned randomly to various treatments. Each unit randomly
chosen for the study has an equal chance of being assigned to any treatment.

The analysis of variance technique previously is called a one-way analysis of variance

since there is only one independent variable.

The two-way analysis of variance is an extension of the one-way analysis of variance

already discussed; it involves two independent variables. The independent variables are
also called factors.

Using the two-way analysis of variance, the researcher is able to test the effects of two
independent variables or factors on one dependent variable.

Variables or factors are changed between two levels (i.e., two different treatments).

The benefit of considering other factors is that we can reduce the error variance.

That is, if we can reduce the denominator of the F statistic (reducing the error variance,
the SSE term) the value of F will be larger, causing us to reject the hypothesis of equal
treatment means.

If we can explain more of the variation, then there is less error.

A second treatment variable than when included in the ANOVA analysis will have the
effect of reducing the SSE term, is referred to as a blocking variable.

1
Two-Way ANOVA Statement of Hypothesis

There is one for each independent variable.

For a test of the difference among treatments, the following hypotheses should be used:
H0: 1 2 a, where a is the number of groups
H1: At least one treatment mean is different from the others.

For a test of the difference among blocks, the following hypotheses should be used:
H0: 1 2 b, where b is the number of groups
H1: At least one block mean is different from the others.

The Sum of Squares

The treatments sum of the squares, denoted SSA, is found using the following formula:

x x
2 2
a
b xi x
2

i ij

SSA = i 1 or SSA = b ab

The blocks sum of the squares, denoted SSB, is found using the following formula:
x x
b 2 2

a x j x
2

j ij

SSB = j 1
or SSB = a ab

The total sum of the squares, denoted SST, is found using the following formula:

xij
2
a b

x x x
2
ij
2
ij
SST = i 1 j 1
or SST = ab
The error sum of the squares, denoted SSE, is found using the following formula:
a b

xij xi x j x
2

i 1 j 1
SSE = = SST SSA - SSB

2
Analysis of Variance Summary Table

Source Sum of d.f Mean F

Squares Squares
Treatments SSA a1 MSA MSA
F
MSE
Blocks SSB b1 MSB MSB
F
MSE
Error SSE (a 1)(b1) MSE

Total SST n1

Assumptions for the Two-Way ANOVA

1. The population from which the samples were obtained must be normally or
approximately normally distributed.
2. The samples must be independent.
3. The variances of the population from which the samples were selected must be equal.
4. The groups must be equal in sample size.

Example 2

Rudduck Shampoo sells three shampoos, one each for dry, normal and oily hair. Sales, in
millions of dollars, for the past five months are given in the following table. Using the
0.05 significance level, test whether the mean sales differ for the three types of shampoo
or by month.

Sales (\$ million)
Month Dry Normal Oily
June 7 9 12
July 11 12 14
August 13 11 8
Sept 8 9 7
Out 9 10 13