ANOVA Assumptions - Analysis of Variance Designs - page 68

The normality assumptions are
not violated; the probabilities
(Sig.) are > .05.

Tests of Normality

Kolmogorov-Smirnov a Shapiro-Wilk

therapy Statistic df Sig. Statistic df Sig.

GAFscore 1

2
3

.180 5 .200* .962 5 .819

.200 * .164 5 .975 5 .907

.237 5 .200* .961 5 .814

*.

This is a lower bound of the true significance.

a.

Lilliefors Significance Correction

Figure 5.16 Normality tests produced by Explore .

5.5.5 SPSS HOMOGENEITY OF VARIANCE ANALYSIS
There are a variety of ways to assess homogeneity of variance (homoscedas-
ticity) within SPSS. We will demonstrate how to assess homogeneity of
variance (equal variance) of the dependent variable across levels of the
independent variable by using the SPSS One-Way ANOVA procedure to
obtain Levene's, Welch's, and the Brown-Forsythe tests. Click Analyze ➜
Compare Means ➜ One-Way ANOVA , which produces the One-Way
ANOVA dialog window (see Figure 5.17). As can be seen from Figure
5.17, we have clicked over GAFscore to the Dependent List: and therapy
to the Factor: window.
By clicking the Options pushbutton, we produce the One-Way
ANOVA: Options dialog window (see Figure 5.18) where we checked the
Homogeneity of variance test (i.e., the Levene test), Brown-Forsythe for
the Brown-Forsythe test, and Welch for the Welch test. Clicking Continue
brings you back to the main dialog window and clicking OK allows the

Figure 5.17 Main dialog window of One-Way ANOVA .