Predictors of Participation for Sexuality Items in a U.S. Population-Based Online Survey

Online surveys are a popular tool in sex research and it is vital to understand participation bias in these surveys to improve inferences. However, research on this topic is limited and out of date given the increase in online survey methodology and changes in sexual attitudes. This study examined whether demographics and sexual abuse and assault history predict opting into online survey questions about sex. The sample was recruited for a longitudinal mental health study using a probability-based sampling panel developed to represent the US household population. Participants were masked to the inclusion of sexual content and given a choice to opt into sex questions. Analyses were run on raw and weighted responses to adjust for sampling bias. Of the total sample (n = 476, 62.6% female), 69% opted into sex questions. Raw analysis showed that participants were more likely to be younger, have higher education and income, and have a history of sexual abuse or assault. No racial, gender, relationship status, or regional differences were found. After weighting, effect sizes were reduced for most predictor variables, and only a history of sexual abuse or assault still significantly predicted participation. Results suggest that key demographic features do not have a strong association with participation in sex survey questions. Reasons for participation bias stemming from sexual abuse or assault history should be examined further. This study demonstrates how researchers should continue to monitor participation bias in sex survey research as online methodologies and sexual attitudes evolve over time.