Method: The study used Demographic Health Survey (DHS) data for Uganda (2011), considering women who were ever married aged 15-49 years, with the majority of women between the ages of 15-19 years. Factor analysis was conducted to determine the protective factors of sexual violence and HIV risk based on the theory of gender and power. Always using a condom during sex was used as a proxy variable for HIV risk. Sexual violence was operationalized as having experienced any sexual violence perpetrated by husband or partner. Bivariate and multiple regression analyses were performed using SPSS to predict sexual violence and HIV risk. Hayes’ SPSS process macro was used for the mediation analysis.
Results: Results from Kaiser-Meyer-Olkin measure of sampling adequacy was above .7. Communalities were all above.3 and all factor loadings met the criteria of .4 or above generating two latent variables of labor force participation and decision making. There were significant differences in sexual violence, labor force participation, decision making, and HIV risk according to demographic characteristics. Mediation analyses suggested that women labor force participation partially mediated the relationship between sexual violence and HIV risk, but decision making did not. First, sexual violence significantly predicted labor force participation in the first regression analysis (t=-3.75, β= -.11, p<.001) meaning that experience with sexual violence had a negative impact on women labor force participation. Second, sexual violence significantly predicted HIV risk in the second regression analysis (t=-3.22, β=-0.38, p=.001) meaning that women experience with sexual violence decreased condom use. Third, when adding labor force participation, the mediator, into the regression, the beta weight of sexual violence to HIV risk decreased from -0.38 to -0.57 and this relationship was significant (t=-4.17, p<.001), meaning that the effect of sexual violence on HIV risk could be partially decreased if women participate in the labor force market. Finally, labor force participation significantly predicted reducing risk of getting HIV by always using a condom (t=6.53, β=.42, p<.001). The Sobel test was significant.
Implications: Findings from this study indicate the need for development of policies and programs to address the question of sexual violence within institutions of marriage in the African context. Couples counseling programs that are geared towards improving healthy sexual relationships should be enhanced. In addition to conducting routine health check-ups, health providers should also inquire about experience with sexual violence so that protective measures can be instituted.