ANOVA Assumptions - Analysis of Variance Designs - page 60

Ta b l e 5 . 1 . Hypothetical data set: Global assessment of functioning by type of
therapy

Factor A (type of therapy)

Treatment Brief Psychodynamic Cognitive-behavioral

conditions a 1 a 2 a 3

Raw scores 50 35 72

55 50 69

68 39 70

72 55 70

60 45 71

5.5 SPSS APPLICATIONS: ASSUMPTION VIOLATION
DETECTION AND SOLUTION

Consider the hypothetical data in Table 5.1. Assume we are interested
in the effects of type of therapy (the independent variable) on global
assessment of functioning (GAF), our dependent measure. The three levels
of our independent variable (type of therapy) are brief, psychodynamic,
and cognitive-behavioral therapy; we coded these therapies as 1, 2, and
3, respectively. The GAF dependent variable is a single score that can
vary from 1, indicating severe dysfunction , to 100, indicating good general
functioning . Thus, the higher the score, the better is the mental health
functioning of the client. Assume that a total of fifteen adult community
mental health clients were assigned randomly and in equal numbers to
the three treatment conditions.
In this and other sections throughout Chapters 5-16, we will show
in our figures both dialog windows and output tables displayed on the
screen by SPSS and SAS. We encourage you to read carefully Appendixes
A and B at the back of this topic on the basics of these computer statistical
packages (SPSS and SAS, respectively) and their usage before attempting
these analyses.

5.5.1 SPSS DATA SCREENING
We will begin with a data-cleaning run with SPSS Frequencies .This
analysis will enable us to check the appropriateness of each variable's set
of numerical values. Enter SPSS for Windows and open your SPSS data
file by clicking File ➜ Open ➜ Data and select the SPSS data file you wish
to analyze.
We begin this analysis by clicking Analyze ➜ Descriptive Statistics ➜
Frequencies , which produces the SPSS Frequencies dialog window. There
are only two variables in this file and we have clicked over both of them
( therapy and GAFscore ) by highlighting or activating each variable and
clicking the arrow in the center of the dialog window, as shown in Figure
5.6. The left panel of the dialog window is now empty and the two variables
now reside in the right panel labeled Variable(s) .The Display frequency
tables window that produces frequency tables has been checked (this