Abstract

Differences in individual behaviors have failed to explain racial disparities between Black and White men who have sex with men (MSM). However, reporting of behaviors and partner characteristics are assumed to be non-differentially reported by race. From 314 participants, this study used the two-sided data—where sexual partners provide information on each other and their relationship—of 127 dyads of Black and White MSM from Atlanta, GA, to assess the reliability of partner-reported demographic characteristics and the concordance of sexual behaviors and partnership attributes by race. We compared proportions of concordance by race using a modified kappa (K m) to assess chance-corrected agreement. The median difference in age between self- and partner-reports was 0 (0–1) years. Compared to self-reports, 97 % of the partners of Black participants and 96 % of the partners of White participants correctly classified their race. We observed poor agreement on pre-sexual discussion (K m = 0.18) and being in an ongoing relationship (K m = 0.13), with no differences by race (p = 0.11). Although not statistically significant, Black MSM dyads had lower levels of concordance for unprotected anal intercourse in the previous 12 months (68 %) compared to White dyads (90 %), with fair agreement among Black dyads (K m = 0.26). Measures of partner-reported age and race are likely accurate; however, certain self-reported sexual behaviors and partnership attributes may be unreliable and differentially reported by race. Our findings highlight the need to assess the validity of measures used to estimate HIV transmission and inform racial disparities research.