Information Technology Reference
In-Depth Information
between cases to be appropriate. Kolmogorov-Smirnov and Shapiro-Wilk
tests of normality of the dependent variable (GAF) across the levels of type
of therapy were not statistically significant (
p
05), indicating no normality
assumption violations. The Bartlett and Brown-Forsythe tests of the equality of
error variances assumption were statistically significant (
p
>.
05), indicating
heterogeneity (unequal variances) across independent variable groups. This
heterogeneity was attributed to the small sample size in the present example.
<.
CHAPTER 5 EXERCISES
5.1.
Three groups of tae kwon do black belt martial artists were taught a
new form that consisted of twenty complicated kicks, punches, and blocks,
and under conditions of no feedback, positive feedback and encouragement,
and criticism and yelling. Participants were randomly assigned to treatment
conditions. After two weeks of “training,” students performed their forms
and were rated by a Master instructor on a 1
=
poor
to 10
=
excellent
scale.
No feedback
Positive feedback Criticism
3
7
10
6
8
9
2
8
8
3
9
4
4
10
9
Check for normality and homogeneity of variance violations by using either
SPSS or SAS.
5.2.
A random sample of five adult community mental health clients were
randomly assigned to the following four treatment conditions. Their GAF
scores after six weeks of treatment served as the dependent measure. Assume
that clients were equated on initial GAF scores at the beginning of the
treatment.
Group Medication only
Brief Cognitive-behavioral
58
65
50
70
55
60
50
65
60
68
45
60
48
58
40
60
50
60
50
65
Check for normality and homogeneity of variance violations using either
SPSS or SAS.
