Last month we wrote about what we learned from the Bureau of Justice Statistics new report, “Rape and Sexual Assault Among College-Age Females, 1995-2013.” We noted that the rates reported by the BJS, which were based on the National Crime Victimization Survey (NCVS), differ from other widely cited statistics about the prevalence of rape among college-age women. In this post, we’ll dive deeper into why these numbers are so different.
What Are the Other Reports?
The NCVS is one of three recent surveys that researchers have used to study rape and sexual assault among college students and in the general population.
The other two are:
The National Intimate Partner and Sexual Violence Survey (NISVS)
The Campus Sexual Assault Study (CSA)
Other surveys worth mentioning are the National Violence Against Women Survey and The Sexual Victimization of College Women. Though both studies date to the late 90s, their findings have also been widely cited and can provide useful context and background for anyone who wants to understand this complicated issue.
They Do Different Things
The NISVS, CSA, and NCVS differ in purpose and methodology.
NCVS is a survey about crime. The survey grew out of the realization that many crimes were not reported to police and that a more accurate measure of victimization was needed. Hence, unlike the FBI’s Uniform Crime Reporting, the NCVS surveys respondents about both reported and unreported crimes.
CSA and NISVS approach rape and sexual assault from a public health perspective. The purpose of the CSA was “To examine the prevalence, nature, and reporting of various types of sexual assault experienced by university students in an effort to inform the development of targeted intervention strategies.” The NISVS’s primary objectives are to measure the prevalence of intimate partner violence and the impact and health consequences of this violence on victims.
They Employ Different Survey Methods
The NCVS follows a group of households over several years, interviewing them every six months. In contrast, the CSA and NISVS are surveys that capture responses from a single point in time. The NCVS asks respondents about events that happened since the last interview, whereas the CSA and NISVS ask about events that occurred during a specified reference period.
The problem in the CSA and NISVS’s approach is that respondents may unintentionally over report the experiences by including events that fell outside the time frame as if they fell within the time frame. According to NCVS, the reporting of traumatic events may be particularly prone to this effect (called telescoping). Thus cross-sectional studies (like the CSA and NISVS) may end up with higher rates than longitudinal studies like the NCVS.
In the NCSV and NISVS, the researchers interview the respondents. This allows them to clarify any confusion around questions but also introduces the possibility that the interviewer might steer or otherwise affect the subject. The CSA, on the other hand, was a web-based survey, which eliminated the influence the interviewer might exert on the respondents but also prevented the respondents from clarifying any confusion they may have had.
They Use Different Definitions
Because the NCVS is a survey about crime, it uses definitions of rape and sexual assault that are “shaped from a criminal justice perspective.” CSA and NISVS use broader definitions of sexual assault that may include incidents that do not rise to the level of a crime. See definitions below (warning: the definitions include explicit language).
The NCVS defines rape as “the unlawful penetration of a person against the will of the victim, with use or threatened use of force, or attempting such an act.” Sexual assault is defined more broadly and generally involves unwanted sexual contact.
The CSA measures rape due to force and incapacitation (that is, when the victim is incapacitated by drugs or alcohol). It also measures unwanted sexual contact.
The NISVS measures five types of sexual violence: rape (including due to incapacitation), sexual coercion (“unwanted vaginal, oral, or anal sexual penetration that occurs after a person is pressured in a nonphysical way”), being made to penetrate someone else, unwanted sexual contact (such as kissing or fondling), and non-contact unwanted sexual experiences (such as flashing or harassment).
They Ask Different Questions
NCVS asks directly about rape, whereas CSA and NISVS both use behavioral cue questions.
For example, the NCVS asks, “has anyone attacked or threatened you in any of these ways…any rape, attempted rape, or other type of sexual attack.” Whereas the NISVS and CSA avoid the terms rape and focus instead on describing events that would qualify as sexual assault or rape, “When you were drunk, high, drugged, or passed out and unable to consent, how many people ever…had vaginal sex with you?”
The way these questions are asked influences how respondents answer. Critics of the NCVS suggest that by asking directly about rape, it fails to measure victims who have experienced rape but may not realize it or may not wish to acknowledge it. Critics of the CSA and NISVS’s questions suggest that they over report sexual assault by using broader and potentially confusing definitions.
They Survey Different People
The CSA only surveys students, and the NISVS does not ask respondents whether or not they are students. Thus, NCSV is the only one of the three surveys that allows researchers to reliably compare rates between students and non-students.
It should also be noted that while both the NCVS and NISVS survey the general population, the CSA only surveyed undergraduate students at two large public universities (one in the South and one in the Midwest).
Interested in More Information?
The National Crime Victimization Survey offers its own discussion of why rates of sexual violence vary between different surveys. We recommend that you read their analysis.