Assessing the Reliability of Self-Reported Sexual Behavior Among Male Couples

Coupled gay, bisexual, and other men who have sex with men (GBMSM) are at particularly high risk for HIV, and a clear understanding of behavioral risk is key to effective interventions. Accurate behavioral self-reports are a crucial component of valid sexual health research, yet reliability of these data remains understudied. This study aimed to quantify and identify predictors of dyadic discordance in reported 3-month anal intercourse (AI) occurrence and frequency. Using cross-sectional data from 407 male couples in the U.S. (2016-2017), we calculated proportional dyadic concordance and used dyad-level logistic and linear regression to identify demographic, behavioral, and relationship traits associated with the odds of discordant frequency reports and the relative difference between discordant partner reports. Couples had high levels of concordant reports of 3-month anal AI occurrence (97%) but low interpartner agreement in reported frequency (37%). After adjustment, the odds of discordance were significantly associated with dyadic employment and differences on the Communal Coping to Reduce HIV Threat Scale (CCS) (p < .05). Among frequency-discordant couples, the mean relative difference between partner reports was 52.80% ± 35.91% (M ± SD). After adjustment, relative differences between partners' reported AI frequencies were positively associated with interpartner differences in CCS (p < .05). These results indicate that among GBMSM couples in committed partnerships, self-reported sexual behavior data may be impacted by granularity, recall, and relationship characteristics. Further research in this area is warranted to better understand measurement error in self-reported sexual activity data.