Description
Policies at both the state and federal level are made based on data gathered from national and state surveys that examine sexual-related cognitions and behaviors through participants' self-report of attitudes and actions. However, there has yet to be any study examining their factorial invariance--more specifically, whether the relationship between the items assessing sexual behavior and their underlying construct differ depending on gender, ethnicity, or age. The following study examined the factor structure of four items selected from the sexual behavior questionnaire part of the National Health and Nutrition Examination Survey (NHANES), with three items examining type of sexual intercourse, and one item looking at the gender of sexual partner. Factorial invariance was examined across gender (men [N = 2,294], women [N = 2,352]) and ethnicity/race (Latinos/Latinas [N = 1,045], Non-Hispanic Whites [N = 1,548], Non-Hispanic Blacks [N = 1,336], Non-Hispanic Asians [N = 717]). Factorial invariance was also examined across generation, used as a proxy for age: participants aged 18-27 were categorized as Millennials (N = 1,020), those 28-47 as a GenerationX'ers (N = 1,757), those 48-63 as Baby Boomers (N = 1,438), and those 64-69 as part of the Silent Generation (N = 431). A chi-square test was also conducted, establishing that there are significant differences (all chi-square values were statistically significant: p < .001) across gender, ethnicity/race, and generation in rates of sexual activity (oral, vaginal, anal, and same gender sex). As NHANES provided different versions of the sexual behavior survey depending on participant gender, invariance was first tested across gender to determine whether subsequent invariance tests across ethnicity/race and generation could be done across gender. The items did not achieve configural invariance so the female and male data were not collapsed together. Across ethnicity for both men and women, and across generation for women, the items were configurally invariant--the same parameters exist for the groups being compared--although items exhibited metric invariance across Latino/Latina and Non-Hispanic Black participants for both men and women. However, the items did not achieve configural invariance across generation for men. On the test of invariance across generation for men, the configural invariance model could not be identified so the baseline models for each generational group were examined, revealing that while the four item one factor model fit well for the Millennial and GenerationX groups, it was a poor fit for the Baby Boomer and Silent Generation groups. The results suggest that gender moderated the invariance across generation, and that comparisons between ethnic and generational groups should not be made between men and women, or even within each gender. These findings highlight the need to develop programs and interventions that promote a more inclusive definition of "having had sex".
