Reported prevalence rates of sexual violence range widely in the literature, even in the same sample when using two different questionnaires of the same construct. These discrepancies are concerning as they indicate we may be underestimating the rate of sexual violence and, therefore, resources and treatment for victims. Thus, the current study aimed to investigate two mechanisms that may contrib-ute to differences in reported prevalence rates across the literature and discrepancies within studies: the tactic-first and item-order hypotheses. Participants were 265 MTurk workers whom all complet-ed the Post-Refusal Sexual Persistence Scale (PRSPS), then were randomly assigned to one of two versions of a tactic-first Sexual Experiences Survey (T-SES). Experimental conditions varied in the item order of the T-SES, one condition received the traditional hierarchical item order (n = 130) while the other received a randomized item order (n = 135). Our results suggest strong support for the tactic-first hypothesis; victimization prevalence rates on the T-SES were double compared to the traditional SES (54.1 vs. 19.8%) in prior research. Further, in both conditions, victimization prevalence rates were statistically equivalent between the PRSPS (62.6%) and the tactic-first SESs (56.2 and 54.1%), χ2(1) < 2.5, p ≈ .1 -.7, contrary to prior research. We did not find support for the item-order hypothe-sis; there were few differences between item-order conditions. Our findings indicate that uninten-tional underreporting remains a threat to validity in sexual violence assessment, and continued re-search into the mechanisms of measurement is warranted.
An Initial Test of the Tactic-First and Item-Order Hypotheses: Accounting for Response Discrepancies in Sexual Victimization Questionnaires
