What’s the point of asking? A review of the reliability of self-reported measures of sexual behaviour

In this survey methods webinar, independent researcher Ignacio Franco Vega will review the reliability of self-reported measures of sexual behaviour. The presentation will cover reliability in general, how it is measured, the subtleties involved in using biological markers and the consequences that unreliable results might have in research and policy. Self-reported measures are the primary source of information to study sexual behaviour.

However, they present various issues that can potentially bias the results of our research (e.g., social desirability, recall errors, inadequate question and answer structures). This study aims to establish the size of such bias.

Ignacio Franco Vega conducted a rapid review to identify studies that assessed the reliability of self-reported measures. This focused on studies that contrast self-reported condom use with biological markers of semen exposure. The investigation reviewed 33 studies with contrasting self-reported measures and biomarkers (Prostate-specific antigen and Y-chromosome DNA). Studies were conducted worldwide, but most focused on at-risk groups in sub-Saharan Africa. All studies showed substantial inconsistencies between the two sources. Discrepancies range between 9% to 70% of participants.

However, there are concerns about the validity of the PSA and YcDNA as a biomarker of semen exposure. Several confounding factors might create both false positives and negatives, including the estimated rate of biomarker decay, endogenous presence of PSA in women and non-standardised thresholds for a positive result.

The study found that self-reported measures of condom use are unreliable; scientists and policymakers should be cautious when concluding anything from them. Biomarkers, albeit useful, are far from being a perfect measure.