Table 4.1: Analysis of Variance for Mean MHCS Sub-scale Scores
Source df F rp2 p
Between subjects
Category 4 25.15 .10 .001*
Gender 1 17.79 .02 .001*
Category X Gender 4 .73 <.01 .570
S within-group 893 (19.77)
error
Note: Values enclosed in parentheses represent mean square errors. rp2 denotes partial eta squared. An
asterisk indicates significant results p <.01.
MHCS scores. The ANOVA indicated no significant interaction between protected
category and gender F(4,893) = .733, p > .05, partial eta square (rip2) = .003, but
significant main effects for protected category F(4,893) = 25.15, p < .001, rip2 = .10, and
gender, F(1,893) = 17.79, p < .001, rip2 = .02 (See Table 4.1). The multiple correlation
coefficient squared or R2 is a measure of the strength of relationship. The multiple
correlation coefficient squared for the 5 x 2 two-way ANOVA model was R2 = .29 and
the adjusted R2 for the model was adjusted R2 = .28.
A Levene's test of unequal variance was conducted to determine whether the
assumption of equal error variance had been violated. Levene's test was significant
F(9,893) = 9.87, p < .001 indicating the assumption had been violated concluding the
error variance is not equal across groups. A one-way ANOVA was conducted for
protected category and gender separately to explore whether the violation of equal
variance occurred in both analyses.