Testing Measurement Invariance of a Scale of Self-Efficacy for Sexual Health Between Heterosexual and Sexual Minority African American Young Adults

Background and Purpose: Self-efficacy has been shown to be strongly associated with the performance of HIV preventive behaviors. Although several scales of condom use self-efficacy have been previously evaluated, no studies have examined the validity of the scale of self-efficacy for sexual health used in the Cohort 7 Adult Questionnaire, a self-report survey that HIV-prevention programs funded through SAMHSA’s National Minority Substance Abuse and HIV Prevention Initiative are required to administer to program participants. Given the wide-spread use of this scale, it is important that it maintain validity among those most disproportionately affected by HIV, including African American young adults of all sexual orientations. The purpose of this study was to use confirmatory factor analysis (CFA) to determine if the validity of this scale of self-efficacy for sexual health varied between heterosexual and sexual minority (SM) African American young adults.

Methods: Data were drawn from baseline surveys of participants enrolled in HIV prevention programs delivered at a community-based agency located in an urban, Midwestern setting. Participants were 226 African American young adults ages 15-24 years (M = 19). Equal sample sizes of heterosexual and SM participants were used. Participants were asked to rate how confident they were that they could perform a healthy sexual behavior on six Likert-type items with a 4-point scale ranging from “not at all” to “very much.” Three items specifically referred to condom use (e.g. “Ask your partner to use a condom”) and the others referred to other components of negotiating healthy sexual boundaries (e.g. “Refuse to engage in sex practices you didn’t like”). All six items were reflected by a single factor of self-efficacy for sexual health. CFAs were conducted in Mplus using the robust weighted least squares estimator with corrections for means and variances due to the ordered nature of the data.

Results: Gender varied significantly by sexual orientation sub-group. SM participants were more likely to be male, while heterosexual participants were more likely to be female, χ² (3, N=226) = 29.733, p < .001. A CFA conducted with the heterosexual sub-group demonstrated good model fit, χ² (6, N=113) = 9.067, p =.170, (RMSEA = .067, CFI = .996, TLI = .991). All six factor loadings were strong and statistically significant (p<.0001), with standardized estimates ranging from .760-.859. The same model estimated on the SM sub-group yielded poor model fit, χ² (6, N=113) = 13.019, p <.05, (RMSEA = .102, CFI = .993, TLI = .983), indicating factorial non-invariance between heterosexual and SM sub-groups. Since factorial non-invariance was established, no further steps were taken to test for configural and strong invariance.

Conclusions and Implications: The Cohort 7 scale of self-efficacy for sexual health does not perform equally well among heterosexual and SM young adults. Given the disproportionate rates of HIV among SM young people, especially young men of color who have sex with men, these findings highlight a need to develop measures of self-efficacy for sexual health that are valid in multiple populations in order to accurate assess the effectiveness of HIV prevention interventions.