A psychometric evaluation of the Sex-Role Egalitarianism Scale with African Americans

Sex Roles: A Journal of Research,  May, 2004  by LaVerne A. Berkel

• E-mail
• Print
• Link

RESULTS

The means, standard deviations, and reliability coefficients for scores on the SRES-KK are presented in Table I. Also included in Table I are the means, standard deviations, and reliability coefficients for scores on the SRES-KK obtained by the European American comparison group. The reliability coefficients for the African American students' scores ([greater than or equal to].84) suggest that the instrument is a reliable measure of gender role attitudes.

To assess construct validity, the men's scores were compared to the women's scores; it was expected that women would report more egalitarian attitudes than would men. A one-way analysis of variance was conducted to test this hypothesis. Because of unequal sample sizes in the groups, Levene's Test of Equality of Error Variance was conducted. The results of the Levene's test were significant: Levene (1, 164) = 28.87, p < .001. Because of the inequality of the variance, in comparing means between men and women, the Welch test, a robust test of equality of means was used instead of the F statistic. Statistically significant differences were found between men's and women's SRES scores, Welch (1, 63.94) = 22.11, p < .001, which supports the first hypothesis.

Because age, religious affiliation, and level of education have also been found to predict gender role attitudes (e.g., Brogun & Kutner, 1976), additional analyses were conducted. A one-way ANOVA to examine religious affiliation (scored dichotomously; 1 = affiliated with a particular religious group; 2 = not affiliated) was not significant, F(1, 161) = 1.93, p = .17. Another one-way ANOVA that examined three educational groups (1 = freshman, sophomores, and juniors, 2 = seniors and beginning advanced graduate students, 3 = advanced graduate students) was not significant, F(1, 148) = 1.07, p = .35. To examine the association between age and gender role egalitarianism scores, a correlation analysis was conducted, which also was not significant, r = .04, p > .05.

A secondary purpose of the study was to explore whether the African American students' scores on the SRES would differ from those of European American students. First, a one-way ANOVA was conducted to determine if the two samples differed in age, religious attendance, or education level. No significant differences were found between the two groups on the number of services they attended annually, F(48, 460) = 1.05, p = .38; however, significant differences were found for both age, F(1, 476) = 11.80, p < .001, and years of education, F(1, 449) = 16.33, p < .001. On average, the African American students were about 1 year older and had more education. Because of these differences, both age and education level were included as covariates in a 2 (race) X 2 (sex) factorial ANOVA to examine whether race, sex, or the interaction of race and sex affected SRES-KK scores. A main effect for sex was found, F(1, 429) = 62.16, p < .001. No main effect for race, F(1, 428) = 2.84, p = .09, or interaction effect, F(1, 428) = 1.34, p = .25, was found, which supports the second hypothesis of the study.