ideology, and frequency of church attendance) used. Belief in the “free choice” model of homosexuality was the strongest predictor of homophobia in the sample with a critical ratio value of 5.91, thus validating hypothesis 3.
Hypothesis 4 was also validated; with a critical ratio of 3.61, a negative correlation between interpersonal contact with gay men and/or lesbians as friends and/or family members and homophobia was statistically significant. Hypothesis 5 was also valid. Support of a nondiscrimination policy protective of gay men and lesbians in the workplace was negatively correlated with homophobia with a critical ratio value of -4.01. Lastly, non- support of a nondiscrimination policy protective of gay men and lesbians in the workplace was positively correlated with homophobia with a critical ratio value of 3.23.
The researcher also assessed the overall goodness of fit for the original model using all of the endogenous variables collected on the demographic survey instrument. After reconstructing a revised measurement model (removing all statistically insignificant endogenous variables), the chi- square, probability, comparative fit index (CFI), Tucker-Lewis index (TLI), root mean squared error of approximation (RMSEA), CMIN/degrees of freedom values and squared multiple correlations were compared. The overall goodness of fit for the revised model was improved, indicating a much stronger measurement model to assess the overall homophobia of the sample.
Following the statistical analysis, the researcher provided a discussion based on the results of the data. Comparing the