TABLE 2—

Adjusted Multivariate Correlates of Sexual Risk: Heterosexual and Sexual Minority Youths (n = 822): Johannesburg, South Africa; October 2008–November 2009

Variable	Ever Had Vaginal/Anal Sexa (n = 822), OR (95% CI)	Age at Sexual Debut (n = 366), B (SE)	Lifetime No. of Partners (n = 374), B (SE)	Sex Trading (n = 388), OR (95% CI)	Condom Useb (n = 360), OR (95% CI)
Sociodemographics
Sexual minority orientationc		−0.52 (0.29)	3.90*** (0.95)	3.11* (1.12, 8.62)
Male	2.61*** (1.87, 3.64)	−0.89*** (0.18)	2.62*** (0.60)	5.22** (2.04, 13.37)
Age	1.71*** (1.40, 2.08)	0.72*** (0.11)			0.51* (0.27, 0.96)
Black race	1.41* (1.01, 1.97)				0.23** (0.08, 0.65)
Parents unemployed		−0.40 (0.28)
Psychosocial risk factorsd
Depression				2.42 (0.72, 8.14)
Substance use	2.00*** (1.38, 2.90)	−0.16 (0.22)	1.36 (0.78)	4.46 (0.95, 20.89)
Partner-perpetrated violence	1.66* (1.11, 2.49)	−0.30 (0.19)		4.90*** (2.26, 10.63)
Older partner	4.03*** (2.15, 7.57)	−0.06 (0.26)
Psychosocial protective factors
Self-esteem				0.89* (0.80, 0.99)	0.88* (0.78, 0.99)
Parent–child communication		0.02 (0.02)		0.94 (0.85, 1.03)	0.81** (0.70, 0.94)

Note. CI = confidence interval; OR = odds ratio. We conducted bivariate analyses (not shown) in which each sociodemographic and psychosocial variable and each outcome variable were used. Only variables with P < .1 in the bivariate tests were included in each adjusted multivariate regression.

^aAs a result of small sample sizes, we combined the vaginal sex and anal sex variables into this single variable.

^bDichotomized as always versus sometimes/never.

^cWe classified participants as sexual minority youths if they self-identified as homosexual, bisexual, or undecided; reported a history of same-sex sexual intercourse; or reported being involved in a same-sex relationship.

^dNone of the traumatic stress bivariate analyses had a P < .1; thus, traumatic stress was excluded from the multivariate analyses.

*P < .05; **P < .01; ***P < .001.