NOTE: Bonnie Fisher at this government office estimates that there are 35.5 rapes per 1,000 female college students in the United States. If all American women are raped at the same rate as female college students, and there is no reason at all to believe that they aren't, then there were 5,168,800 rapes in the US last year. This would be a rape rate per 100,000 population of 1,846, which would make American women the most raped women in the world, bar none. If Bonnie and these other folks at VAWO know what they're doing, then compared to the 660,000,000 women of India, Japan, Italy, and France, American women are raped 1,678 TIMES as often. American women would even be 75 times more likely to be raped than the 4.4 million women of Sweden, who are the second most raped women in the world.
It would also mean that American women are raped 246 times as often as Finnish women, who also report that they are raped much more often than the average woman. But Finnish women are raped one third as often as Swedish women, which raises serious questions about why women in two very similar cultures, geography, and weather, would have such radically different experiences with rape.
The answer seems to be in the amount rape victims receive in the form of victim compensation programs.
The sexual victimization of college women
Series: Research Report
Authors: Bonnie S. Fisher, Francis T. Cullen, and Michael G. Turner
December 2000 NCJ 182369
Figures, charts, forms, and tables are not included in this .asp file. To view this document in its entirety, download the Adobe Acrobat graphic file available from the National Institute of Justice website or order a print copy from NCJRS at 800-851-3420 (877-712-9279 For TTY users).
The results from the college women studies were supported under award 95-WT-NX-0001 from the National Institute of Justice and award 97-MU-MU-0011 from the Bureau of Justice Statistics, U.S. Department of Justice. Points of view in this document are those of the authors and do not necessarily represent the official position of the U.S. Department of Justice. Joanne Belknap, Ph.D., worked with Professors Fisher and Cullen on award 95-WT-NX-0001 in developing and revising the surveys and led the focus groups with Professor Fisher.
The National Institute of Justice and the Bureau of Justice Statistics are components of the Office of Justice Programs, which also includes the Bureau of Justice Assistance, the Office of Juvenile Justice and Delinquency Prevention, and the Office for Victims of Crime.
Table of contents
This study contributes extremely important data to our understanding about the prevalence and nature of violence against women in the United States.
College campuses host large concentrations of young women who are at greater risk for rape and other forms of sexual assault than women in the general population or in a comparable age group. Based on their findings, Bonnie Fisher and her colleagues estimate that the women at a college that has 10,000 female students could experience more than 350 rapes a year-- a finding with serious policy implications for college administrators.
Fisher also found that many women do not characterize their sexual victimizations as a crime for a number of reasons (such as embarrassment, not clearly understanding the legal definition of rape, or not wanting to define someone they know who victimized them as a rapist) or because they blame themselves for their sexual assault. The study reinforces the importance of many organizations' efforts to improve education and knowledge about sexual assault.
A unique feature of this study, brought about by joint funding from our two agencies, is its parallel use of two different survey methods for learning about sexual assault of women. The differences in estimated levels of sexual assault that are associated with the methodological differences will help us design better and more accurate surveys in the future.
Julie E. Samuels
Jan M. Chaiken
During the past decade, concern over the sexual victimization of female college students has escalated. In part, the interest in this problem has been spurred by increasing attention to the victimization of women in general; until the relatively recent past, female victims received very little attention. However, this is no longer true. Terms such as "date rape" and "domestic violence" have entered the public lexicon and signify the unprecedented, if still insufficient, notice given to women who have been victimized.
Attention to the sexual victimization of college women, however, also has been prompted by the rising fear that college campuses are not ivory towers but, instead, have become hot spots for criminal activity. Researchers have shown that college campuses and their students are not free from the risk of criminal victimization. It is noteworthy that large concentrations of young women come into contact with young men in a variety of public and private settings at various times on college campuses. Previous research suggests that these women are at greater risk for rape and other forms of sexual assault than women in the general population or in a comparable age group. College women might, therefore, be a group whose victimization warrants special attention.
Recognizing these risks, the U.S. Congress passed the Student Right-to-Know and Campus Security Act of 1990 (hereafter referred to as the act). This legislation mandates that colleges and universities participating in Federal student aid programs "prepare, publish, and distribute, through appropriate publications or mailings, to all current students and employees, and to any applicant for enrollment or employment upon request, an annual security report" containing campus security policies and campus crime statistics for that institution (see 20 U.S.C. 1092(f)(1)).
Congress has maintained an interest in campus crime issues, passing legislation that requires higher educational institutions to address the rights of victims of sexual victimization and to collect and publish additional crime statistics (e.g., murder and nonnegligent manslaughter, arson). For example, Congress amended the act in 1992 to include the Campus Sexual Assault Victims' Bill of Rights, which requires colleges and universities (1) to develop and publish as part of their annual security report their policies regarding the awareness and prevention of sexual assaults and (2) to afford basic rights to sexual assault victims. The act was amended again in 1998 to include additional reporting obligations, extensive campus security-related provisions, and the requirement to keep a daily public crime log; some States already required a public log (Public Law 105-244). The 1998 amendments also officially changed the name of the act to the Jeanne Clery Disclosure of Campus Security Policy and Campus Crime Statistics Act. In 1999, the U.S. Department of Justice awarded $8.1 million to 21 colleges and universities to combat sexual assault, domestic violence, and stalking. In 2000, 20 additional schools were awarded $6.8 million. Two national-level studies are currently in the field. The first study examines how institutions of higher education respond to the report of a sexual assault. The second one is a multisite evaluation of the programs and policies implemented in the above-mentioned 41 schools.
What we know about sexual victimization of college women
Like government officials, researchers also have given attention to the sexual victimization of college women and have conducted a number of studies. Although illuminating, much of the research is generally characterized by one or more of the following limitations:
The National College Women Sexual Victimization (NCWSV) study, described in this report and funded by the National Institute of Justice (NIJ), attempted to build on, and surmount the limitations of, existing research on the sexual victimization of college students by:
In addition, the research project contained a comparison component designed to assess how rape estimates that use the two-stage process (behaviorally specific questions and incident reports) compared with rape estimates drawn from a sample of college women who completed a survey based on the National Crime Victimization Survey (NCVS). The comparison component was funded by the Bureau of Justice Statistics (BJS).
The resulting data furnish perhaps the most systematic analysis of the extent and nature of the sexual victimization of college women in the past decade.
Who was surveyed?
NCWSV study results are based on a telephone survey of a randomly selected, national sample of 4,446 women who were attending a 2- or 4-year college or university during fall 1996. The questions were asked between February and May 1997. The sample was limited to schools with at least 1,000 students and was stratified by the size of the total student enrollment (1,000-2,499; 2,500-4,999; 5,000-19,999; 20,000 or more) and the school's location (urban, suburban, and rural). Schools were randomly chosen using a probability proportional with the size of the total female enrollment. Students were then randomly selected using a sampling frame provided by the American Student List Company. This company provided the school address and telephone number for each student in the sample.
Each sample member was sent a letter describing the study and research protocol approximately 2 weeks prior to when a trained female interviewer called using a computer-aided telephone interviewing system. The response rate was 85.6 percent.
The comparison component used the same two-stage methodology as the main study except victimization was measured by using the screen questions and the incident report employed by NCVS. One purpose of the comparison component was to conduct a methodological experiment that would provide insight into the extent to which rape estimates are influenced by survey methods.
How was sexual victimization measured?
Measurement of sexual victimization was based on responses to "screen questions" and on a reference period for the victimization. In addition to the victimization measures, survey questions and secondary data sources were used to investigate the factors that potentially placed women at risk of being sexually victimized.
With important exceptions noted later, sexual victimization was measured largely by following the two-stage measurement format of NCVS. NCVS first asks a series of screen questions that seek to determine if a respondent has experienced an act that may possibly be a victimization. If the respondent answers "yes," then for each of the times that the act was experienced, the respondent is asked by the interviewer to complete an "incident report." This report contains detailed questions about the nature of the events that occurred in the incident. The report is used to classify the type of victimization that took place; that is, responses to questions in the incident report--not the screen questions--are used to categorize whether a victimization occurred and, if so, what type.
Some researchers have contended that the screen questions as worded in NCVS are not detailed enough to identify all women who have experienced a rape or another type of sexual assault. A respondent may not answer "yes" to a screen question unless it is worded in a way that reflects the experience the respondent has had. To rectify this limitation, researchers have argued that sexual victimization should be measured with screen questions that are both numerous and detailed enough that respondents will not misunderstand what is being asked.
NCWSV, therefore, used a series of behaviorally specific screen questions that sought to assess whether respondents had experienced a range of sexual victimizations. A behaviorally specific question, for example, is one that does not ask simply if a respondent "had been raped"; rather, it describes an incident in graphic language that covers the elements of a criminal offense (e.g., someone "made you have sexual intercourse by using force or threatening to harm you . . . by intercourse I mean putting a penis in your vagina"). The same logic can be used to ask about other forms of sexual victimization, such as sexual coercion or unwanted sexual contact.
Examples of the screen questions used in the NCWSV study are listed in exhibit 1. Each completed rape screen question asks the respondent about a different form of penetration in which force or the threat of harm was used. A statement then follows each question that defines the type of penetration. For example, anal sex is defined as "putting a penis in your anus or rectum." The other screen questions provide examples of the behaviors that respondents were asked about.
The NCWSV rape screen questions are similar, if not identical, to those used by Kilpatrick and his associates and by Tjaden and Thoennes. The use of behaviorally specific screen questions is an important difference between the current survey and NCVS. The NCVS screen questions begin with a reference to a type of criminal victimization that may have been experienced (e.g., "were you attacked or threatened"), which is then followed by a list of short cue responses about the potential victimization. This list includes cues regarding specific places or situations in which the victimization could have occurred (e.g., "at work or at school"); objects that could have been used (e.g., "with any weapon, for instance, a gun or knife"); actions that could have been associated with the victimization (e.g., "face-to-face threats"); actions that constitute a criminal victimization (e.g., "rape, attempted rape, or other types of sexual attack"); and people who might have perpetrated the criminal act (e.g., "a relative or family member"). There is also a screen question that asks about "incidents involving forced or unwanted sexual acts."
Drawing on the NCVS screen question and incident report methodology, the NCWSV screen questions were followed by a detailed incident report that (1) clarified what type of victimization, if any, had occurred and (2) collected information about various aspects of the incident (e.g., victim-offender relationship, whether the victimization took place on or off the college campus, whether the incident was reported to the police). Responses to the screen questions were not used to classify the type of victimization reported by the respondent. Instead, classification was based on the responses in the incident report to questions about (1) the type of penetration experienced (e.g., penile-vaginal, anal, oral); (2) the type of unwanted sexual contact experienced (e.g., touching, grabbing, or fondling); and (3) the means of coercion used by the perpetrators (e.g., force, threat of force). Like Koss et al. and NCVS, the incidents were classified using a hierarchical algorithm; that is, incidents were classified by the most severe type of sexual victimization that occurred within an incident. For example, if within an incident report the victim answered questions indicating she had experienced a completed rape and attempted sexual coercion, the incident was classified as a completed rape.
To limit potential response bias due to recall or memory decay, the NCWSV survey questions used a reference period that had a clear starting date for students. Thus, respondents were asked if they had experienced a sexual victimization "since school began in fall 1996." The survey was conducted in 1997 between late February and early May. On average, the reference period for the victimization covered almost 7 months (6.91 months). To participate in the study, respondents had to be enrolled in a college or university at the start of the 1996 fall semester.
In addition to the victimization measures, the NCWSV survey contained questions about respondents' demographic characteristics, lifestyles or routine activities, living arrangements, prior sexual victimizations, and so forth. Secondary data sources were used to measure the characteristics of the schools the respondents attended (e.g., size of enrollment, location, crime rate). These individual- and institution-level variables were used in multivariate analyses that investigated which factors potentially placed women at risk of being sexually victimized.
What types of sexual victimization were measured in the NCWSV study?
Measures of 12 types of sexual victimization were constructed; they are defined in exhibit 2. Most important, the NCWSV study included measures of both completed and attempted rape as well as threats of rape. The study also measured completed, attempted, and threatened sexual coercion (penetration with the use of nonphysical forms of coercion) and unwanted sexual contact (sexual contact, but not penetration, with force or threat of force). In addition, the study measured stalking and visual and verbal forms of sexual victimization.
How extensive is rape among college women?
Exhibit 3 reports the extent of rape found in the NCWSV study. As shown, 2.8 percent of the sample had experienced either a completed rape (1.7 percent) or an attempted rape incident (1.1 percent). The victimization rate was 27.7 rapes per 1,000 female students.
We recognize that a hierarchical scoring procedure is not the only way to count victims and incidents, especially because we have multiple victims. Another estimation procedure is to count the total number of completed rape victims and the total number of attempted rape victims separately. For example, suppose there were two incident records for respondent 00: One incident was classified as a completed rape, and the other was classified as an attempted rape (recall that using a hierarchical scoring procedure, respondent 00 would be counted as a completed rape victim). Respondent 00 would now count as a completed rape victim and as an attempted rape victim. Using this "separate" counting procedure, there were 57 attempted rape victims, or 1.3 percent of the sample.
Because some women were victimized more than once, the rate of incidents was higher than the rate of victims (35.3 per 1,000 students). Of the 123 victims, 22.8 percent (n = 28) were multiple-rape victims.
A separate analysis, again using the same hierarchical scoring procedure, found that when rates were computed for only undergraduate students, the percentage of students victimized was 1.8 percent for rape and 1.3 percent for attempted rape. The comparable figures for nonundergraduate students were, respectively, 0.8 percent and 0 percent.
At first glance, one might conclude that the risk of rape victimization for college women is not high; "only" about 1 in 36 college women (2.8 percent) experience a completed rape or attempted rape in an academic year. Such a conclusion, however, misses critical, and potentially disquieting, implications. The figures measure victimization for slightly more than half a year (6.91 months). Projecting results beyond this reference period is problematic for a number of reasons, such as assuming that the risk of victimization is the same during summer months and remains stable over a person's time in college. However, if the 2.8 percent victimization figure is calculated for a 1-year period, the data suggest that nearly 5 percent (4.9 percent) of college women are victimized in any given calendar year. Over the course of a college career--which now lasts an average of 5 years--the percentage of completed or attempted rape victimization among women in higher educational institutions might climb to between one-fifth and one-quarter.
Furthermore, from a policy perspective, college administrators might be disturbed to learn that for every 1,000 women attending their institutions, there may well be 35 incidents of rape in a given academic year (based on a victimization rate of 35.3 per 1,000 college women). For a campus with 10,000 women, this would mean the number of rapes could exceed 350. Even more broadly, when projected over the Nation's female student population of several million, these figures suggest that rape victimization is a potential problem of large proportion and of public policy interest.
How do the NCWSV rape estimates compare with the rape estimates based on the National Crime Victimization Survey?
The sexual victimization literature contains a great deal of discussion about how rape estimates from the Nation's federally sponsored victimization survey, NCVS, compare with estimates from other national surveys. This issue was examined through a comparison component. Like the main NCWSV study, the comparison study was conducted in the 1996-97 academic year, from late March to mid-May. The sample size was 4,432 college women; the response rate was 91.6 percent. 
Every effort was made to ensure that, aside from using different screen and incident report questions, the methodology used in both the main and comparison components was the same. Thus, both components (1) contacted sample members with a letter that explained the purpose of the survey, (2) employed the same sampling design and sampling frame, (3) used the same reference period for victimization ("Since school began in fall 1996 . . ."), and (4) measured victimization using the screen question-incident report methodology. Both components also were conducted by the same survey research firm (see endnote 10) and were administered by trained female interviewers using a computer-aided telephone interviewing system.
However, in assessing the influence of different methodologies for measuring sexual victimization, the two studies differed on one methodological issue: the wording of the screen questions and the wording of the incident-level questions used to determine the type of incident. As previously described, the main study substantially modified the NCVS format to include a range of behaviorally specific screen questions. In contrast, the comparison component used a format that was closely aligned with the survey format of NCVS. All of the screen questions used in the comparison component came directly from NCVS, as did the incident-level questions used to determine what type of violent victimization the respondent had experienced. Both components used a hierarchical algorithm to classify the type(s) of victimization that the respondent described in the incident report.
We should note, however, that the methodology used in the comparison component differs from that used in NCVS in one respect. In addition to structured responses to the survey questions, NCVS interviewers record a brief "verbatim description" of the victimization incident from those respondents who report experiencing rape or sexual assault. These verbatim responses are used to clarify what occurred in an incident and to code whether an incident should count as a sexual victimization. Thus, according to BJS staff:
In our comparison component study, the estimates were not adjusted using verbatim responses. We do not know how much this consideration affects the findings reported for the comparison component that is, again, based on NCVS methodology. None of the Criminal Victimization in the United States annual publications report how much the NCVS estimates are adjusted using verbatim responses, or whether such adjustments cause estimates to increase or decrease compared with estimates coded solely on respondents' answers to the structured screen and incident-report questions.
This definition guided the classification of incidents in the comparison study as a completed rape, an attempted rape, or a threat of rape. In the Criminal Victimization in the United States series published by BJS, estimates for attempted rape and threats of rape are reported separately. The same is true in this report so as to compare rape estimates from the two components of the study.
How do the rape estimates from these two studies compare? It should be noted that studies that use behaviorally specific screen questions generally find higher levels of sexual victimization than those reported by NCVS. Most important, this finding has occurred in recent research using a national-level sample and behaviorally specific questions. 
Looking at exhibit 4, it is clear that estimates from the comparison study for completed rape, attempted rape, and threats of rape are considerably lower than the respective estimates from the main study. The percentage of the sample that reported experiencing a completed rape in the comparison study was 11 times smaller than the percentage of victims in the main component (0.16 percent compared with 1.7 percent). The attempted rape estimate from the comparison component was six times smaller than the attempted rape estimate (0.18 percent compared with 1.1 percent) from NCWSV. A similar pattern was evident for threats of rape; the estimate based on the comparison component was four times smaller than the NCWSV estimate (0.07 percent compared with 0.3 percent).
What accounts for these differences? Given the similarities between the two studies, it would appear that the differences most likely stem from the wide range of behaviorally specific screen questions used in the NCWSV study. Compared with the NCVS screen questions employed in the comparison component, the use of graphically worded screen questions in NCWSV likely prompted more women who had experienced a sexual victimization to report this fact to the interviewer. Their responses in the incident report determined whether those answering "yes" to a rape screen question were subsequently classified as rape victims. Even so, it appears that behaviorally specific screen questions are more successful in prompting women who have in fact been sexually victimized to answer in such a way that they are then "skipped into" the incident report by interviewers.
What is unknown, however, is whether behaviorally specific screen questions produce higher estimates of victimization in general or only higher estimates of sexual victimization. It is possible that, due to the sensitive nature of sexual victimization, graphically descriptive screen questions are needed to prompt reluctant victims to report their victimization to interviewers. The other possibility, however, is that a large set of behaviorally specific questions would result in more victim reports for any type of victimization, including property crimes and other forms of violent crime (e.g., aggravated assault, robbery). Future research on NCVS methodology might profit from exploring this issue.
Do women define their victimization as a rape?
In each incident report, respondents were asked, "Do you consider this incident to be a rape?" For the 86 incidents categorized as a completed rape, 46.5 percent (n = 40) of the women answered "yes," 48.8 percent (n = 42) answered "no," and 4.7 percent (n = 4) answered "don't know." Among women who experienced other forms of sexual victimization (n = 1,318), it is noteworthy that 3.4 percent (n = 42) defined their sexual victimization as a rape and 1.1 percent (n = 14) answered "don't know."
Some scholars believe that the failure of women to define a victimization as a rape calls into question whether researchers have truly measured the crime of rape. Others suggest, however, that the true prevalence of rape is best measured by carefully worded questions on victimization surveys, such as NCWSV. Women may not define a victimization as a rape for many reasons (such as embarrassment, not clearly understanding the legal definition of the term, or not wanting to define someone they know who victimized them as a rapist) or because others blame them for their sexual assault. Which of these reasons is more or less correct cannot be definitively substantiated here because little systematic research has examined why women do or do not define as a rape an incident that has met the researcher's criteria for a rape.
How extensive are other forms of sexual victimization?
Exhibit 5 presents the extent of victimization across 10 forms of sexual victimization other than rape. Threats of sexual victimization happened less often than other forms of sexual victimization. Across the 10 types of victimization in exhibit 5, the incident rate per 1,000 female students ranged from a low of 9.5 to a high of 66.4.
Exhibit 6 presents the data in a slightly different form and contains rape incidents. This exhibit illustrates the percentages of women in the sample who had experienced at least one victimization in three separate categories: (1) physical force, (2) nonphysical force, and (3) either physical or nonphysical force or both. Because the third category includes respondents who have experienced both types of victimization, its percentage is not computed by summing the percentages in the physical and nonphysical categories. As is shown, 15.5 percent of the college women were sexually victimized during the current academic year. In the sample, 7.7 percent experienced an incident involving the use or threat of physical force, and 11.0 percent experienced a victimization that did not involve force.
How extensive is prior sexual victimization?
Respondents were also asked if they had experienced sexual victimization incidents before starting school in fall 1996. These incidents were measured only with single questions, not incident reports (that is, the two-stage process of screen questions followed by an incident report was not used). To limit bias, we attempted to use the detailed questions shown in exhibit 7. Still, the findings must be assessed in light of this methodological limitation.
As exhibit 7 shows, about 1 in 10 college women said they had experienced a rape, while the same proportion stated that they were victims of an attempted rape. Almost the same proportion also had sexual intercourse or contact in which they were subject to threats of nonphysical punishment or promises of reward. Unwanted or uninvited sexual contacts were widespread, with more than one-third of the sample reporting these incidents.
Do victims know their offenders?
Most victims knew the person who sexually victimized them. For both completed and attempted rapes, about 9 in 10 offenders were known to the victim. Most often, a boyfriend, ex-boyfriend, classmate, friend, acquaintance, or coworker sexually victimized the women. College professors were not identified as committing any rapes or sexual coercions, but they were cited as the offender in a low percentage of cases involving unwanted sexual contact. The victim-offender relationship for rape incidents is displayed in exhibit 8.
Variation in the type of sexual victimization that occurred on a date was evident. With regard to date rape, 12.8 percent of completed rapes, 35.0 percent of attempted rapes, and 22.9 percent of threatened rapes took place on a date.
When does sexual victimization occur?
The vast majority of sexual victimizations occurred in the evening (after 6 p.m.). For example, 51.8 percent of completed rapes took place after midnight, 36.5 percent occurred between 6 p.m. and midnight, and only 11.8 percent took place between 6 a.m. and 6 p.m.
Where does sexual victimization occur?
The majority of sexual victimizations, especially rapes and physically coerced sexual contact, occurred in living quarters. Almost 60 percent of the completed rapes that occurred on campus took place in the victim's residence, 31 percent occurred in other living quarters on campus, and 10.3 percent took place in a fraternity. Off-campus sexual victimizations, especially rapes, also occurred in residences. However, particularly for sexual contacts and threatened victimizations, incidents also took place in settings such as bars, dance clubs or nightclubs, and work settings.
Are women victimized on or off campus?
College women are victimized both on campus and off campus. For nearly all types of sexual victimization, however, off-campus victimization is more common (exhibit 9). This conclusion must be qualified because off-campus sexual victimizations may take place in bars and nightclubs or in student residences close to campus. Thus, even if a student is victimized off campus, she may be engaged in an activity that is connected to her life as a student at the college she attends.
Do sexual victims take protective actions during the incident?
As exhibit 10 shows, for nearly all forms of sexual victimization, the majority of female students reported attempting to take protective actions during the incident. For both completed rape and sexual coercion, victims of completed acts were less likely to take protective action than those who experienced attempted victimization. This finding suggests that the intended victim's willingness or ability to use protection might be one reason attempts to rape or coerce sex failed.
Exhibit 11 reports the most common forms of protective action taken by victims during rape incidents. Note that the most common protective action was using physical force against the assailant. Nearly 70 percent of victims of attempted rape used this response--again, a plausible reason many of these acts were not completed. Other common physical responses included removing the offender's hand, running away, and trying to avoid the offender. Verbal responses also were common, including pleading with the offender to stop, screaming, and trying to negotiate with the offender.
Are victims hurt in the victimization incidents?
Victims in the sample generally did not state that their victimization resulted in physical or emotional injuries. In about one in five rape and attempted rape incidents, victims reported being injured, most often citing the response "bruises, black-eye, cuts, scratches, swelling, or chipped teeth." The percentage injured by other types of victimization was lower, ranging from 0 percent (completed sexual contact without force) to 16.7 percent (threatened rape).
Are some women more at risk of being sexually victimized?
Multivariate logit models for each type of sexual victimization measured were estimated to predict the likelihood of having been victimized. Consistent across the models, it was found that four main factors consistently increased the risk of sexual victimization: (1) frequently drinking enough to get drunk, (2) being unmarried, (3) having been a victim of a sexual assault before the start of the current school year, and (4) living on campus (for on-campus victimization only).
Do women report victimization incidents to the police?
Few incidents of sexual victimization were reported to law enforcement officials. Thus, fewer than 5 percent of completed and attempted rapes were reported to law enforcement officials. In about two-thirds of the rape incidents, however, the victim did tell another person about the incidents. Most often this person was a friend, not a family member or college official.
Victims gave a number of reasons for not reporting their victimizations to law enforcement officials (exhibit 12). Some reasons indicated that they did not see the incidents as harmful or important enough to bring in the authorities. Thus, the common answers included that the incident was not serious enough to report and that it was not clear that a crime was committed. Other reasons, however, suggested that there were barriers to reporting. Such answers included not wanting family or other people to know about the incident, lack of proof the incident happened, fear of reprisal by the assailant, fear of being treated with hostility by the police, and anticipation that the police would not believe the incident was serious enough and/or would not want to be bothered with the incident.
How extensive is stalking?
In addition to the 12 types of sexual victimization (exhibit 2), this research assessed a form of victimization that has been infrequently studied: stalking. In general, for behavior to qualify as stalking, the attention given to someone must be repeated and it must create fear in a reasonable person. Accordingly, stalking was measured with this screen question: "Since school began in fall 1996, has anyone--from a stranger to an ex-boyfriend--repeatedly followed you, watched you, phoned, written, e-mailed, or communicated with you in other ways that seemed obsessive and made you afraid or concerned for your safety?" If a respondent answered "yes," she was then given an incident report that asked detailed questions about the stalking that occurred.
The survey indicated an incidence rate of 156.5 per 1,000 female students. Indeed, fully 13.1 percent of the female students in the sample (n = 581) had been stalked since the school year began. This figure approximates what was found in a pretest of the survey conducted on students attending one university. It also is similar to the 6-month prevalence figure reported by Mustaine and Tewksbury, which, in a survey of 861 women attending 9 postsecondary institutions, found that 10.5 percent of the female students reported that they had been stalked.
In contrast, Tjaden and Thoennes' national study of women reports much lower annual rates of stalking: 1 percent to 6 percent, depending on the definition of stalking used. Compared with the current study, the lower extent of stalking in Tjaden and Thoennes' research may be because (1) their study used a more restrictive definition of stalking; (2) their study focused on females across the life course (age 18 to 80 years or older), rather than on younger women among whom stalking is more prevalent; and (3) their study did not focus specifically on college students. It may be that the social domain of college places women in situations and in contact with a range of men that increase the chances of being stalked.
It should be noted, however, that like the study by Tjaden and Thoennes, the estimates in this study of the extent of stalking vary depending on the criteria used to define what counts as stalking victimization. Again, more than 13 percent of the women in the sample were stalked if this victimization is defined as a woman experiencing repeated, obsessive, and frightening behavior that made the victim afraid or concerned for her safety. Even so, if we were to decide that such behavior counts as a stalking victimization only if the person were actually threatened with harm--a requirement for criminal stalking in many States--the extent of stalking victims in the sample falls to 1.96 percent. These results suggest that, in the future, researchers should examine how estimates of the extent of stalking may vary widely depending on the criteria used to define what "counts" as a stalking victimization.
What is the nature of stalking incidents?
As with other sexual victimizations, four in five victims knew their stalkers. Of the stalkers who were known, they were most often a boyfriend or ex-boyfriend (42.5 percent), classmate (24.5 percent), acquaintance (10.3 percent), friend (5.6 percent), or coworker (5.6 percent). Female students were infrequently stalked by college professors or graduate assistants.
Stalking incidents lasted an average of 60 days. About 30 percent of the female students were stalked only off campus; the remaining victims were stalked either only on campus or both on and off campus. The most common forms of stalking behaviors reported by victims were being telephoned (77.7 percent), having an offender waiting outside or inside places (47.9 percent), being watched from afar (44.0 percent), being followed (42.0 percent), being sent letters (30.7 percent), and being e-mailed (24.7 percent). Almost two-thirds of the sample indicated that they were stalked at least two to six times a week.
Although some victims reported being physically injured, the most common consequence was psychological: Almost 3 in 10 women said they were "injured emotionally or psychologically" from being stalked. In 15.3 percent of incidents, victims reported that the stalker either threatened or attempted to harm them. In 10.3 percent of incidents, the victim reported that the stalker "forced or attempted sexual contact."
In nearly three-fourths of incidents, victims reported that they had taken "actions as a result of their stalking." Exhibit 13 shows actions victims took following stalking incidents. Two of the most common responses were "to avoid the stalker" (43.2 percent) or, conversely, "to confront the stalker" (16.3 percent). Beyond the data in exhibit 13, in about 17 percent of incidents, victims reported the stalker to the police. In contrast, in more than 9 in 10 incidents, victims confided in someone--such as a friend, family member, or roommate--that they were being stalked.
Are some women more at risk of being stalked?
A multivariate logit model was estimated to predict the likelihood of being stalked. The risk of being a stalking victim was increased by a number of factors: the propensity to be in places with alcohol; living alone; being in a dating relationship, especially early in the relationship, as opposed to being married or living with an intimate partner; being an undergraduate; being from an affluent family; and having experienced sexual victimization before the beginning of the current academic year. Also, among racial/ethnic groups, Asian/Pacific Islander women were significantly less likely to be stalked while American Indian/Alaska Native women were significantly more likely to be stalked compared with women in other racial/ethnic groups. Notably, American Indian/Alaska Native women had the highest likelihood of any racial/ethnic group to experience a stalking. This is consistent with Tjaden and Thoennes' research, which reported that American Indians/Alaska Natives are at greatest risk of being stalked.
What is the extent of visual and verbal sexual victimization?
Finally, this research measured the extent to which women were involuntarily exposed to visual images and verbal comments that would generally be considered sexually victimizing. Since these relatively "minor" types of victimization were plentiful, it was not possible to obtain a detailed report on each incident. Instead, results showed only whether a type of victimization was experienced and, if so, how many times it happened both on and off campus.
As exhibit 14 reveals, most respondents did not experience visual victimization. Still, about 6 percent of female students had been shown pornographic pictures, almost 5 percent had someone expose their sexual organs to them, and 2.4 percent were observed naked without their consent. Verbal victimizations, moreover, were commonplace. About half the respondents were subjected to sexist remarks and to catcalls and whistles with sexual overtones. One in five female students received an obscene telephone call and was asked intrusive questions about her sex or romantic life. One in ten students had false rumors spread about her sex life.
The sexual victimization of college students has emerged as a controversial issue, pitting feminist scholars who claim that the sexual victimization of women is a serious problem against conservative commentators who claim that such victimization is rare and mostly a fictitious creation of ideologically tainted research. The research reported here undoubtedly will not settle this debate; battle lines are solidly entrenched and how the data are interpreted will, to a degree, lie in the "eye of the beholder." However, the current study attempts to add a judicious voice to this conversation by attempting to furnish a methodologically sound assessment of the extent and nature of the sexual victimization of female students.
To summarize, the national-level survey of 4,446 college women suggests that many students will encounter sexist and harassing comments, will likely receive an obscene phone call, and will have a good chance of being stalked or of enduring some form of coerced sexual contact. During any given academic year, 2.8 percent of women will experience a completed and/or attempted rape. This figure is not based only on broadly worded, behaviorally specific screen questions because all victimization incidents reported in the screen questions were verified through subsequent questions in the incident report. Furthermore, the level of rape and other types of victimization found in the survey becomes an increasing concern when the victimization figures are projected over a full year, a full college career, and the full population of women at one college or at colleges across the Nation.
The results also hold important methodological implications. The comparison component study sponsored by the Bureau of Justice Statistics allowed the rare opportunity to conduct a quasi-experiment in how the methods used to measure sexual victimization might potentially affect estimates of victimization. Thus, two randomly selected samples of college women were surveyed using very similar methodology, with one noteworthy exception: A different way of measuring sexual victimization was used with each sample. Results showed that a methodology that uses behaviorally specific screen questions in combination with an incident report yields considerably higher estimates of completed, attempted, and threatened rape than are found using NCVS methodology.
Future research should explore the implications of this finding for NCVS. As noted, it was not determined whether using a number of behaviorally specific screen questions tends to increase estimates of only rape or whether the technique would also increase estimates of other types of victimization (i.e., the more widely in scope and the more closely in detail that possible victimizations are probed, the more victims are prompted to report their victimization to interviewers). However, assuming that the methodology used in this study is defensible, it seems likely that NCVS underestimates the true incidence of rape victimization in the United States.
We should note, however, one other possible factor that might have contributed to the differences in victimization between the main and comparison components: the "context" of two surveys. In the main component of NCWSV, the respondents were instructed in an initial contact letter and in instructions during the interview that the survey was focusing on "unwanted sexual experiences." In contrast, the comparison component was patterned after NCVS, which is a crime survey. In this part of the study, respondents were sent an initial letter that mentioned the "increasing concern about criminal victimizations that women may experience during college," and the interview itself contained questions measuring victimization by other types of criminal offenses. It is conceivable, therefore, that respondents on the main component study were sensitized to report a broad range of sexual victimization incidents they experienced, while those on the NCVS-based comparison component limited their reports to the incidents they defined as criminal. If so, this contextual difference would mean that the comparison component was measuring a much narrower domain of sexual victimization. One caution in this line of reasoning is that, as discussed previously, nearly half of the completed rape victims defined their victimization as a "rape," a clear criminal offense. Even when the count of rape victims is limited to this group, the prevalence of rape victims is several times greater in the main component than in the comparison components. Still, the impact of survey context on respondents' responses to sexual victimization questions remains an area that warrants further research.
Of course, many other methodological issues in addition to the use of behaviorally specific screen questions and survey context will have to be addressed in the quest to design surveys capable of achieving more accurate estimates of rape and other forms of sexual victimization. These would include, but not be limited to, issues such as the differential meaning that words used in questions might have to respondents, the impact of the sequencing of questions on answers, the use of more verbatim descriptions of victimization incidents in coding "what happened" in a sexual assault, and perhaps the use of computer-aided personal interviewing as a means of encouraging respondents to disclose traumatic events. In short, systematic, rigorous experimental research into the factors that affect victim responses and, in turn, victimization estimates--especially in the sensitive area of rape and sexual assault-- remains in its beginning stages.
Although exceptions exist, most sexual victimizations occur when college women are alone with a man they know, at night, and in the privacy of a residence. Most women attempt to take protective actions against their assailants but are then reluctant to report their victimization to the police. Although based on fewer cases, these same patterns were found as well in the comparison component survey, which used NCVS methodology. The analysis also revealed that some college women were more at risk of being victimized than others. Several factors appeared to increase various types of victimization: living on campus, being unmarried, getting drunk frequently, and experiencing prior sexual victimization.
Finally, in the aftermath of this study, an important challenge remains: Taking the information found and developing programs and policies that may reduce female students' risk of victimization. Minor forms of sexual victimization--sexist statements, harassing catcalls, sexually tainted whistles--appear to be commonplace. How can a more civil environment be achieved without compromising free speech? Much is known about the circumstances under which sexual victimization, including rape, most often occurs. How can this information be used in crime prevention programs, including rape awareness seminars designed for women or rape prevention seminars designed for men? Furthermore, the relatively high prevalence of stalking--a form of victimization often ignored by college officials--is cause for concern. What strategies can women use to prevent or end stalking? What programs might colleges implement to control or counsel men who stalk? More generally, how can the lives of college women--whether on, close to, or off campus--be made safer and thus free from the costs imposed by the experience of sexual victimization?
Exhibit 1: Survey Screen Questions
Women may experience a wide range of unwanted sexual experiences in college. Women do not always report unwanted sexual experiences to the police or discuss them with family and friends. The person making the advances is not always a stranger, but can be a friend, boyfriend, fellow student, professor, teaching assistant, supervisor, coworker, somebody you meet off campus, or even a family member. The experience could occur anywhere: on or off campus, in your residence, in your place of employment, or in a public place. You could be awake, or you could be asleep, unconscious, drunk, or otherwise incapacitated. Please keep this in mind as you answer the questions.
Now, I'm going to ask you about different types of unwanted sexual experiences you may have experienced since school began in fall 1996. Because of the nature of unwanted sexual experiences, the language may seem graphic to you. However, this is the only way to assess accurately whether or not the women in this study have had such experiences. You only have to answer "yes" or "no."
Exhibit 2: Descriptions of Types of Victimizations
About the National Institute of Justice
The National Institute of Justice (NIJ), a component of the Office of Justice Programs, is the research agency of the U.S. Department of Justice. Created by the Omnibus Crime Control and Safe Streets Act of 1968, as amended, NIJ is authorized to support research, evaluation, and demonstration programs, development of technology, and both national and international information dissemination. Specific mandates of the Act direct NIJ to:
In recent years, NIJ has greatly expanded its initiatives, the result of the Violent Crime Control and Law Enforcement Act of 1994 (the Crime Act), partnerships with other Federal agencies and private foundations, advances in technology, and a new international focus. Examples of these new initiatives include:
The Institute Director establishes the Institute's objectives, guided by the priorities of the Office of Justice Programs, the Department of Justice, and the needs of the criminal justice field. The Institute actively solicits the views of criminal justice professionals and researchers in the continuing search for answers that inform public policymaking in crime and justice.
To find out more about the National Institute of Justice please contact:
To obtain an electronic version of this document, access the NIJ Web site
If you have questions, call or e-mail NCJRS.
This site is a cooperative project of VAWO - the Violence Against Women Office, Office of Justice Programs, U.S. Department of Justice and MINCAVA - the Minnesota Center Against Violence & Abuse at the University of Minnesota.
Additional information about this site can be obtained by contacting: [email protected]
Copyright 1998, 1999, 2000, 2001 MINCAVA