B3. Now run a one-way ANOVA with hprobgrp as the independent (group) variable, using the variable satovrl (overall satisfaction with material well-being, Variable #43) as the outcome variable. This variable is a summated rating scale variable for women’s responses to their degree of satisfaction with four aspects of their material well-being—their housing, food, furniture, and clothing for themselves and their children. Each item was coded from 1 (very dissatisfied) to 4 (very satisfied), so the overall score for the four items could range from a low of 4 (4  1) to 16 (4  4). To run the one-way analysis, click Analyze ➜ Compare Means

➜ Oneway. In the opening dialog box, move satovrl into the Dependent List, and hprobgrp into the slot for Factor. Click the Options pushbutton, and click Descriptives and Homogeneity of Variance, then Continue. Next, click the Post Hoc pushbutton and select LSD, Scheffe, and Tukey.

Click Continue, then OK, and answer these questions:

(a) What are the mean levels of satisfaction in the three groups? Which group is most dissatisfied?

(b) Can the null hypothesis regarding equality of variances in the three populations be rejected?

(c) What was the value of the F statistic that tested the equality of mean levels of satisfaction in the three groups?

(d) What were the degrees of freedom?

(e) What was the probability level for the F statistic? Can the null hypothesis for equality of means be rejected?

(f) According to Tukey’s HSD test, were any group means significantly different from any others? If yes, which ones?

(g) Were our statistical decisions sensitive to the particular multiple comparison test we used? In other words, would we have come to different conclusions if we had used the Scheffe or LSD test?