State of Florida was randomly selected from the state Board of Nursing licensee database. One-hundred sixty-five (165) surveys were eventually used in the analysis of the data. Using t-tests and one-way analysis of variance (ANOVA) statistical differences were found between age and race/ethnicity. Although males were more homophobic than females, this difference was statistically insignificant; the youngest nurses of the sample were the least homophobic; and Caucasians were the least homophobic among reported ethnicities while African Americans were most homophobic. Differences in ATLG scores based on education were deemed non-statistically significant. The researcher proposed possible confusion based on the categories available for selection on the demographic survey instrument as a potential etiologic source for the lack of statistical significance, while the sample, over-representing females (with only 11 males participating in the study), might have explained the insignificance of the differences between males and females.
To test hypotheses 2, 3, 4, and 5, the researcher applied structural equation modeling (SEM). Confirmatory factor analysis was used to validate the use of the ATLG scale to measure the latent construct of homophobia; all 20 ATLG items were statistically significant indicators to the overall construct with critical ratio values >1.96 while the Cronbach’s alpha was .77. Religious association was a non-significant independent variable in the final analysis of the data. A possible reason for this was failure to capture the significance of religious association based on the three indicators (religion, religious