Measuring sexual violence perpetration acknowledgment: Testing the effects of label and response format

Abstract

Individual acknowledgment of sexual assault and rape perpetration is extraordinarily low in prior research. Only about 1% of individuals report perpetrating rape, in contrast to the 6% perpetrating rape as estimated by using behaviorally specific items that exclude stigmatized words such as rape. The goal of this study was to examine two possible measurement mechanisms for increasing perpetration acknowledgment: label choice and response format. In Sample 1 (N = 291), participants completed two acknowledgment items which varied in label choice. One item used the term rape; one used the term sexual assault. Acknowledgment of perpetration using the label sexual assault was significantly higher than when using the term rape (6.38%-1.71%, p = .01, Cohen's d = 0.44). In Sample 2 (N = 438), participants were presented with a scaled and a dichotomous sexual assault item at different parts of the overall survey. Sexual assault acknowledgment was higher on the scaled item compared to the dichotomous item (15.75% vs. 3.2%, p < .0001, Cohen's d = 0.64). Rates of sexual perpetration as measured behaviorally were higher for ambiguous acknowledgment types ("might or might not," "probably not") than for those reporting "definitely not," (76.81% vs. 29.0%, p < .0001, Cohen's d = 0.59). The two different measurement strategies tested here, using a less stigmatized label such as sexual assault and using a scaled response format, both increased rates of perpetration acknowledgment 3-15x greater than rates documented in prior research.

Keywords: assessment; gender; measurement; rape; sexual assault.

Figure 1

Overlap between the use of the perpetration terms in Sample 1, N=291