U.S. flag

An official website of the United States government, Department of Justice.

Effects of Survey Question Wording on Rape Estimates: Evidence From a Quasi-Experimental Design

NCJ Number
Violence Against Women Volume: 15 Issue: 2 Dated: February 2009 Pages: 133-147
Date Published
February 2009
15 pages

This study examined the measure of rape and other forms of sexual violence from two nationally representative studies, the National College Women Sexual Victimization (NCWSV) study and National Violence Against College Women (NVACW) study.


The results of this study provide four important implications for the measurement of rape. First, the use of behaviorally specific questions not only may produce larger estimates of rape, but they use words and phrases that describe to the respondent exactly what behavior is being measured. Using behaviorally specific screen questions appears to cue more women to recall what they experienced. The use of behaviorally specific questions is not a panacea for addressing measurement error associated with estimating rape, but a step forward in understanding how question wording affects self-report survey responses. Second, drawing on the strength of the two-stage measurement process (screen questions and incident reports) the National Crime Victimization Survey (NCVS) appears to be a promising way to address the measurement error typically associated with a single- stage measurement process; however, further rigorous testing is needed. Third, the context of the two surveys is a factor that might have contributed to significant differences between the NCWSV and the NVACW study estimates; it is plausible that NCWSV respondents were sensitized to report a broad range of sexual victimization incidents they experienced, whereas NVACW respondents limited their reports to incidents they defined as criminal; if so, the contextual difference would mean the NVACW study was measuring a much narrower domain of sexual victimization than was the NCWSV study. Fourth, to further the understanding of rape and a full range of violence, comparative research employing experimental and quasi-experimental designs should not be overlooked. The strength of these types of research designs will allow researchers to manipulate sources of possible measurement error and estimate their effects on estimates of different types of victimization. Tables and references

Date Published: February 1, 2009