Abstract
By now, age and marital status are well-established correlates of criminal victimization, including adult women’s sexual victimization. National crime statistics, as well as a large body of scholarly literature, have specified that younger women and unmarried women are at comparatively higher risk of sexual victimization than older women and married women. Few studies, however, have examined the relationship between age, marital status, and risk of victimization across diverse situational contexts of sexual victimization. The current study used data from the National Crime Victimization Survey (NCVS) to examine whether the relationship between age, marital status, and risk of sexual victimization varied across victimization experiences involving three victim–offender relationships: stranger, acquaintance, and intimate partner. Results indicate both similarities and differences in the relationship between age, marital status, and risk of victimization across these three situational contexts of victimization. As expected, age was a significant predictor of victimization in all models; however, younger women’s increased risk of victimization was far more pronounced for acquaintance and intimate partner victimization experiences as compared with stranger experiences. Also, consistent with prior research, unmarried women were at higher risk of victimization in all models; however, within unmarried status categories, separated women were at highest risk of both intimate partner and acquaintance victimization experiences as compared with never married or divorced women.
By now, age and marital status are well-established correlates of criminal victimization, including sexual victimization. National crime statistics, as well as a large body of scholarly literature, have specified that age and marital status are among the most consistent predictors of sexual victimization for adult women with younger women and unmarried women at higher risk of sexual victimization as compared with older women and married women (Bureau of Justice Statistics, 2013). In the criminological literature, these demographic differences in risk have often been explained using a lifestyles and routine activities theoretical approach that combines lifestyle-exposure theory (Hindelang, Gottfredson, & Garofalo, 1978) and routine activity theory (Cohen & Felson, 1979) and proposes that the lifestyles or daily routines of some women might increase their relative vulnerability or risk of victimization.
Lifestyle-exposure theory (Hindelang et al., 1978) proposes that demographic characteristics, such as age and marital status, are often associated with structural constraints and behavioral expectations that shape personal lifestyles and daily activities, and, in turn, moderate exposure to “risky” situations across demographic groups. For example, middle aged married women’s age and marital status may result in structural constraints, such as household and family responsibilities, or behavioral expectations, such as expectations to spend more of their leisure time with their families at home. Younger unmarried women, on the other hand, may be less likely to experience the same types of constraints and expectations, and, therefore, may be more likely to spend more of their leisure time alone or with friends and acquaintances outside of the home. According to this perspective, status characteristics, such as age and marital status, moderate differences in exposure to potentially “risky” situations and, in turn, result in variations in vulnerability and victimization rates among women of different demographic groups.
Routine activity theory (Cohen & Felson, 1979) shares much in common with lifestyle-exposure theory, though it focuses more on the situational contexts of victimization rather than victim characteristics alone. According to the theory, crime can only occur when three elements converge in space and time: (a) a motivated offender, (b) a suitable target, and (c) the absence of a capable guardian. For example, a motivated offender may be equally motivated to assault any woman within access; however, he may choose to assault a woman who he sees to be alone or intoxicated because he perceives her to be a more suitable target who is less likely to be able to resist. According to this perspective, an individual’s daily activities or routines can affect the degree to which she is exposed to motivated offenders, is considered a suitable target for victimization, and is perceived by potential offenders to have capable guardianship to protect against victimization.
The combined lifestyles/routine activities theoretical approach proposes that demographic characteristics, such as age and marital status, carry with them specific structural constraints and behavioral expectations that shape personal lifestyles and daily activities, and, in turn, moderate exposure to “risky” situations by regulating the convergence of suitable targets with motivated offenders in the absence of capable guardianship. Though both lifestyle-exposure theory and routine activities theory were proposed as more general theories of crime, they have both been applied to the study of sexual violence and victimization and used in criminological studies of micro-level predictors of sexual victimization.
It is important to note that feminist theories of sexual violence generally propose a near constant supply of motivated offenders and suggest that notions of “riskiness” have less to do with women’s behavior and more to do with the social and cultural factors that make it seem legitimate to engage in violence against women, especially in particular situations or contexts, such as when a woman is dressed a certain way or is intoxicated (Abbey, Ross, McDuffie, & McAuslan, 1996). Feminist scholars propose that within contemporary society there exists a “rape culture” that legitimates and sustains high levels of violence against women (Brownmiller, 1975; Buchwald, Fletcher, & Roth, 1993; Cahill, 2001; Griffin, 1971; Koss, 2003; Parrot & Bechhofer, 1991; Russell & Bolen, 2000). Deeply embedded patriarchy within mainstream culture reinforces the permissibility of rape and sexual assault in certain situations and contexts.
The assessment of differential levels of vulnerability within this cultural context is worthy of scholarly consideration to identify and specify the relationship between macro-level rape culture and micro-level victimization of women. If crime is situational in nature, it is important to understand the situational factors that lead to the convergence of motivated offenders with suitable targets. Although such research is intended to inform whether rape culture is more predatory on certain groups of women, it is not intended to prescribe that women deemed “at-risk” should, necessarily, limit or alter their activities. Rather, by understanding why certain groups of women are more likely to be targeted and under what circumstances they are more likely to be targeted, more effective efforts to change the cultural norms that legitimize violence and victim-blaming can begin to emerge.
Prior research on sexual victimization has indicated that age and marital status are strong predictors of women’s victimization. These studies have emphasized that younger women and unmarried women are at comparatively higher risk of victimization, noting specific vulnerabilities such as romantic and sexual inexperience, inability to distinguish danger cues, frequent and excessive alcohol consumption, frequent socialization with unfamiliar men, and frequent casual sexual experiences, all of which would make them more attractive targets for victimization than older and married women (Abbey et al., 1996; Belknap, 1987; Combs-Lane & Smith, 2002; Harney & Muehlenhard, 1991; Himelein, 1995; Koss & Dinero, 1989; Laumann, Gagnon, Michael, & Michaels, 1994; Schwartz & Pitts, 1995). 1 Few studies, however, have explored whether the relationship between age and marital status remains consistent across different types and contexts of sexual victimization.
Sexual violence is often considered a unitary construct when it is, in fact, a heterogeneous phenomenon that includes a range of experiences from threats of sexual assault to violent rape and a range of victim–offender relationships from stranger to intimate partner. Research is only beginning to examine whether there are differences in women’s vulnerabilities for different types and situational contexts of sexual victimization with some studies exploring differences in risk across different types of victimization experiences or involving different victim–offender relationships (Franklin, 2010, 2011; Messman-Moore, Coates, Gaffey, & Johnson, 2008; Messman-Moore, Ward, & DeNardi, 2013; Mohler-Kuo, Dowdall, Koss, & Wechsler, 2004; Testa & Dermen, 1999; Testa, Livingston, VanZile-Tamsen, & Frone, 2003; Testa, VanZile-Tamsen, & Livingston, 2007; Ullman, Filipas, Townsend, & Starzynski, 2006; Walker, Messman-Moore, & Ward, 2011). Few studies, however, have specifically investigated the relationship between demographic variables, such as age and marital status, and victimization within subcategories or situational contexts of sexual victimization.
Feminist scholars have long urged criminologists to examine diverse manifestations of violence against women more specifically to disentangle divergent situational contexts of violence and avoid unintentionally making some phenomena “invisible” (Koss, 2003). Meier and Miethe (1993) wrote that “the failure to examine whether variables have different effects across different contexts is a type of model misspecification that may dramatically alter substantive conclusions about the predictive validity of current theories” (p. 489). Examining the situational contexts of sexual victimization experiences involving different victim–offender relationships is, therefore, an important step for theoretical validation and development of theories of sexual violence that combine theoretical arguments from both criminological and feminist approaches to the study of rape and sexual assault.
The Current Study
The purpose of the current study was to examine whether the relationship between age, marital status, and risk of sexual victimization varied across three victim–offender relationships (stranger, acquaintance, and intimate partner). Age and marital status were expected to be stronger predictors of victimization risk for acquaintance and intimate partner assaults, which may be more likely to occur in the context of certain types of lifestyles and socialization activities more prone to rape culture, as compared with stranger assaults, which are more likely to be crimes of random opportunity. For example, younger, unmarried women who participate in culturally normative socialization activities, such as dating, drinking, flirting, or hooking up, may be more vulnerable to motivated offenders who see women exercising their independence and sexual freedom as suitable targets for victimization. Similarly, young, unmarried women who are more likely to experience some level of relationship instability may also be more vulnerable to situations in which an intimate partner uses violence to try to reassert power and control in a relationship or exact revenge for perceived mistreatment or humiliation. These are examples in which rape culture would permit would-be offenders to rationalize their actions.
Among unmarried status categories, separated women were expected to be at greatest risk of acquaintance and intimate partner violence due to the instability associated with separation. Whereas most prior research has examined marital status as dichotomous (married/unmarried) or has grouped divorced and separated women together, the current study examined divorced and separated women independently with the expectation that separated women may be at greater risk of sexual victimization than divorced or never married women. On the one hand, separated women may be likely to experience higher risk of victimization from the intimate partners who they are trying to separate from and who may be trying to reassert power and control in the relationship and, on the other hand, they may be likely to experience higher risk of victimization due to their renewed participation in socialization and courtship behaviors, such as dating and hooking up.
The two main hypotheses for the study are summarized below:
Data and Methods
Data
The data for the current study were drawn from the National Crime Victimization Survey (NCVS). The NCVS is the only nationally representative source of data specifically designed to measure victimization in the United States. The NCVS uses a stratified, multistage cluster sampling technique to select households for participation in the survey. Households are rotated into the sample for 3 years. After 3 years, households are rotated out and new households are rotated in. All members of selected households aged 12 or older are interviewed every 6 months about their victimization experiences in face-to-face or telephone interviews. Respondents are first asked screener questions about whether they have been attacked or threatened in any way (including rape, attempted rape, or sexual assault) or if they have been forced or coerced into unwanted sexual activity. If they indicate any victimization experiences in the previous 6 months, then follow-up questions are asked for each victimization incident.
The NCVS is a particularly rich data source due to its large sample size and high response rates. Nevertheless, sexual victimization reports are still quite rare in NCVS data. To ensure adequate number of cases for statistical analysis, multiple years of data (1992-2005) were pooled together. The pooled data set, however, was so large that logistic regression modeling on the full sample would result in skewed coefficients and significance tests. In other words, there were exponentially more respondents who were not a victim of a sexual assault than were a victim. In such instances, a case control method is often the only feasible method for data analysis (Mann, 2003).
Case control studies are frequently used in epidemiology and public health to compare outcome frequencies between a group of participants that experience a particular outcome and a comparable group that does not from a source population. For example, studies have compared patients diagnosed with lung cancer with a control group of participants not diagnosed with lung cancer to determine the effects of smoking (and other risk factors) on the risk of developing lung cancer. Case control studies are less conclusive than randomized trials; however, ethical concerns about putting participants at risk of physically or psychologically harmful experiences preclude the use of randomized trials when studying topics such as cancer (and sexual victimization). Case control studies can be used to investigate the relationship between a rare outcome and potential risk factors by calculating odds ratios that approximate relative risk. Although case control studies are well-established and frequently used in other fields, they have rarely been used in criminology.
In the current study, a case control sampling strategy was used to create a data set for statistical analysis. Data for all adult female victims of sexual assault as well as any other violent crime between the years 1992 and 2005 were selected. Individual-level data were merged with household- and incident-level data for each selected respondent. A large random sample of adult female non-victims of sexual assault was added to the data set with individual-, household-, and incident-level data. Random sampling of controls is considered the simplest strategy for control selection and most efficient when there is low risk of substantial differences in confounders between cases and controls (Wacholder, Silverman, McLaughlin, & Mandel, 1992). The total sample size for this data set was 34,876 adult females of which 1,018 were victims of a sexual assault. To maintain the rarity of sexual assault cases in NCVS data, the case control sampling strategy maintained a much larger proportion of non-victims than victims in the final data set that was used for the analyses. The large sample size was also maintained to sustain the statistical power of the study and minimize bias that would likely occur from a smaller sample size. Prior research on case control studies has indicated that large sample sizes can eliminate bias that overestimates odds ratios in small to moderate sample studies (Nemes, Jonasson, Genell, & Steineck, 2009).
Measures
Dependent variables
The dependent variables in the study constituted three dichotomous variables indicating whether or not the respondent had experienced at least one sexual victimization experience perpetrated by a stranger, an acquaintance, or an intimate partner. Eight sexual victimization experiences are included in the NCVS: completed rape, attempted rape, sexual attack with serious assault, sexual attack with minor assault, sexual assault without injury, unwanted sexual contact without force, verbal threat of rape, and verbal threat of sexual assault. Respondents who experienced any of these victimization experiences were classified as sexual assault victims.
Independent variables
The primary independent variables included in the study were age and marital status. Age was recoded into five age groups: 18 to 29, 30 to 39, 40 to 49, 50 to 59, and 60 and above. Although these age groups are somewhat arbitrary, they approximately corresponded to the major decades in women’s lives from early adulthood through retirement. These age groups are often associated with particular milestones, such as graduating from college, starting a career, getting married, having a baby, being promoted at work, becoming an empty-nester, and so on. Certainly, popular culture also makes use of these age group classifications by describing women in their 20s, 30s, 40s, and 50s as distinct and marketing very different products, such as cosmetics, perfume, and clothing, to women in these different age groups.
Marital status was recoded into five dummy variables: never married, married, divorced, separated, and widowed. The never married category included women who may have been living together with a partner but were not in a common law union as well as women who may have been married at some time but had had their marriages annulled. The married category included women living with partners in common law unions and women not living with their partners due to reasons other than marital discord. The divorced category included women who were legally divorced from their former spouses. The separated category included women who were legally separated or not living with their partners due to marital discord. The widowed category included women whose spouses had died. It is important to note that most prior research grouped divorced and separated women together; however, the current study examined divorced and separated women independently.
Other demographic variables were also included in the models. These included cohabitation status, household income, and urban residency. Cohabitation status was included as a control for never married women who may have been living together with a partner to further assess whether living independently might have increased women’s vulnerability. Cohabitation status was recoded into a dummy variable: living alone. Respondents who lived alone were coded as 1. Income and urban residency were also included due to the fact that they are well-established correlates of criminal victimization with poorer women in urban areas experiencing higher risk of victimization in general. Household income was recoded into seven income groups: incomes less than 10k, incomes between 10k and 20k, incomes between 20k and 30k, incomes between 30k and 40k, incomes between 40k and 50k, incomes between 50k and 75k, and incomes of more than 75k (coded 1-7, respectively). Urban residency was recoded into a dummy variable: urban. Respondents who lived in urban areas were coded as 1.
Two other control variables were included in the analysis: household crime and year of survey. The household crime variable was a measure of the amount of criminal victimization in the respondent’s household and was included as a measure of the respondent’s home environment with the expectation that women living in violent home environments would be more likely to experience sexual violence as well. Values ranged from one to four with four representing four or more crime incidents reported in the household. Year of survey represented the year that the respondent reported the victimization and is included as a control due to the fact that multiple years of data were pooled together in the data set.
Analytical Strategy
For case control studies, logistic regression is the preferred analytical tool. Logistic regression models can be used to describe the relationship between a dichotomous outcome variable, such as sexual victimization, and a set of predictor variables, such as demographic characteristics. These models are often used when the relationship between the dependent variable and independent variables is non-linear. For case control studies, these models are particularly useful in predicting the strength and direction of associations between independent variables and a dependent variable, as well as the relative importance of each of the independent variables. Three logistic regression models were estimated to predict relative risk of sexual victimization for adult women based on situational contexts involving three victim–offender relationships: stranger, acquaintance, and intimate partner.
Results
Table 1 presents descriptive statistics for the sample. The total sample size for this data set was 34,876 adult females of which 1,018 were victims of a sexual assault. Noteworthy is the fact that acquaintance victimization was the most prevalent victimization experience and its frequency was more than twice that of intimate partner victimizations and more than three times that of stranger victimizations. These statistics are consistent with prior research which indicates that the vast majority of sexual assaults are perpetrated by non-strangers (Fisher, Cullen, & Turner, 2000; Kilpatrick, Edmunds, & Seymour, 1992; Tjaden & Thoennes, 2000). Descriptive statistics of the demographic distribution of the sample indicated that approximately a third of the sample was between the ages of 18 and 29 (30.6%) and roughly half of the sample was married (53.4%). Also, less than a fifth of the sample lived alone with no other household members (15.8%) and more than a third lived in urban areas (35.2%). The average household income of respondents ranged between U.S. $20,000 and $40,000 per year and the vast majority of respondents did not experience personal or household victimization.
Descriptive Statistics (N = 34,876).
Prior to logistic regression modeling, multicollinearity diagnostics were conducted. None of the Pearson correlation values were higher than .65 and the only correlation values above .6 were marital status variables with one another. Variance inflation factors (VIFs) were also calculated for each of the independent variables and all VIF values were below 3. As a result, none of the variables were excluded due to multicollinearity concerns. Table 2 presents the results of three comparative multivariate models of victimization risk by victim–offender relationship (Model 1 = stranger, Model 2 = acquaintance, Model 3 = intimate partner). Each model used the following independent variables to predict victimization: age, marital status, cohabitation status, urban residency, household income, household crime, and year of survey. All three models were statistically significant at the p < .001 level, and, based on the pseudo r2 values, all three models were moderate in strength, though the models for acquaintance and intimate partner victimizations were more robust (Nagelkerke r2 values of .125, .205, and .215, respectively).
Victimization Risk by Victim–Offender Relationship.
Significant at the .05 level. **Significant at the .01 level. *** Significant at the .001 level.
The analyses in Table 2 indicate that risk of sexual victimization was significantly associated with age, marital status, cohabitation status, and household crime across all three models, though the associations varied in strength across the three models. Although age was a significant predictor across all models, the effect of age on victimization risk was not consistent. For stranger victimization, there was no significant difference in risk for women between the ages of 18 and 59, though women aged 60 and above were at significantly reduced risk. These results could be interpreted to suggest that stranger assault is a crime of opportunity and motivated offenders might assault any accessible woman irrespective of their age-related socialization activities. To the extent that elderly women are more likely to be retired, their reduced risk may simply be a function of their reduced accessibility to motivated offenders.
On the other hand, for acquaintance victimization, there was no significant difference in risk for women between the ages of 18 and 39, but for every decade thereafter, women’s risk of acquaintance victimization was significantly reduced. More specifically, women in their 20s and 30s were about one-and-a half times more likely to be a victim of acquaintance assault than women in their 40s, about two-and-a-half times more likely than women in their 50s, and eight-and-a-half times more likely than women in their 60s. These results suggest that the age–victimization relationship that has been found in prior research is likely to be most representative of acquaintance assaults. The results seem to be consistent with the lifestyles and routine activities approach which suggests that younger women’s socialization behaviors might increase their exposure to motivated offenders who consider them suitable targets for assault. Disturbingly, these results might also indicate that young women who enjoy more sexual freedom are also more vulnerable to being targeted precisely for that reason. In a rape culture, when women assert their sexual freedom, motivated offenders may consider it a license to violate and assault them.
For intimate partner victimization, there was no significant difference in risk for women between the ages of 18 and 49, but for every decade thereafter, women’s risk of intimate partner victimization was significantly reduced. Although there may be alternative explanations, these results could indicate that relationship violence tends to be more common in younger women and, as women get older and more experienced in dealing with relationship problems, they may be better able to identify and protect themselves from violent or unhealthy relationships.
Similar to the results for age, marital status was also a significant predictor of victimization risk across all three models. In results consistent with prior research, never married, separated, and divorced women were at higher risk of sexual victimization than married women across all models. Nevertheless, some noteworthy findings emerged in the comparative analyses. For stranger victimization, women of all unmarried status categories, including widowed women, experienced significantly higher risk of victimization than married women. In fact, widowed women and never married women were at highest risk and were three to four times more likely than married women to experience a stranger victimization. Once again, these results seem to indicate that strangers rely on random opportunities to find targets for victimization. To the extent that unmarried women are more likely to participate in daily routines unaccompanied by other household members, they may be perceived more frequently by motivated offenders to be suitable targets without adequate guardianship.
Widowed women’s risk of victimization, however, did not differ significantly from married women’s for acquaintance and intimate partner victimizations. For both acquaintance and intimate partner victimizations, never married women, separated women, and divorced women, however, experienced significantly higher risk of victimization than married women. Most notable was the finding that separated women were at much higher relative risk, especially for intimate partner victimization. Separated women were almost six times as likely as married women and almost twice as likely as divorced women to be sexually assaulted by an intimate partner. These findings seem to indicate that unmarried women are particularly vulnerable to being victimized by non-strangers. If this association is conceptualized as a function of their lifestyles and routine activities, the results suggest that separation may be a period of particular concern for women because it may make them particularly vulnerable to motivated offenders.
Living alone and household crime were also significant predictors of risk of victimization across all three models. Results indicated that women living alone and women living in households that experienced higher rates of crime were about twice as likely to experience sexual victimization in all three models. Again, results seem to indicate that living alone and living in a violent home environment likely increase women’s vulnerability to be seen as a suitable target for victimization.
Urban residency, household income, and year of survey were significant predictors only in some models. Urban residency was a significant predictor only for Model 3, which indicated that living in an urban area was associated with a reduced risk of intimate partner victimization. This may be related to the density of housing units in urban areas and the fact that violent assaults are more likely to be heard by neighbors in an apartment building in a densely populated urban area rather than in a single-family home in suburbia. Meanwhile, household income, though not a significant predictor for stranger assaults, was a significant predictor for acquaintance and intimate partner victimizations. Respondents with higher household incomes were slightly less likely to be victimized than lower income respondents. To the extent that household income translates to having the resources to afford enhancements to one’s own guardianship, such as, for example, affording to take a taxi home after a night out rather than relying on public transportation or a ride from an acquaintance, income may influence lifestyles and routine activities as well as relative risk.
Finally, year of survey was a significant predictor in Model 2 indicating that there may be a time factor that might help explain women’s risk of acquaintance victimization. Sexual victimization rates underwent a major reduction between 1992 and 2005 and the factors that contributed to that decline might be particularly relevant for acquaintance assaults. It seems probable that the bulk of the reduction in sexual victimization rates was likely due to a reduction in reported acquaintance assaults; however, whether that is indicative of a real decline in sexual violence is a matter of debate. To date there have not been any published studies on this “rape decline,” other than to note that the decline in sexual assault corresponded with an overall decline in crime in the same time period. It is possible that this decline in reported acquaintance victimization experiences may be partly due to cultural changes in dating and courtship behavior, such as the emergence of a “hook-up culture” in the 1990s; however, further research is required to clarify and explain time trends in sexual victimization.
Discussion
The main purpose of the current study was to examine whether the relationship between age, marital status, and risk of sexual victimization varied across victimization experiences involving three victim–offender relationships: stranger, acquaintance, and intimate partner. The main findings from the study are consistent with prior research and indicate that age and marital status are significant predictors of sexual victimization across all three victim–offender relationships; however, the findings also indicate that there are important similarities and differences in the effects of age and marital status on risk of victimization across the three models. The findings underscore the value of comparative analyses based on victim–offender relationships, and, more broadly, the situational contexts of victimization experiences. Whereas prior research has suggested that younger women and unmarried women are at highest risk of sexual victimization, the current study indicates that the association between age, marital status, and risk of victimization is not absolute or uniform and, in fact, varies according to the specific situational contexts of victimization experiences, including the relationships between victims and their offenders. A secondary purpose of this study was to assess whether the associations between age, marital status, and victimization risk might correspond to a lifestyles or routine activities approach to the study of victimization that is informed by feminist theories of sexual violence. The findings of the study do seem to be consistent with this combined theoretical approach.
For stranger victimization, there was little effect of age on victimization risk; however, marital status was a strong predictor of victimization. There was no significant difference in risk of victimization for women between the ages of 18 and 59; however, women aged 60 and above were at reduced risk of victimization. Similarly, unmarried women, whether never married, separated, divorced, or widowed, were significantly more likely to be victimized by a stranger than married women. These findings can be conceptualized from a lifestyles or routine activities theoretical approach and seem to suggest that motivated strangers do not seem to discriminate based on age or physical attractiveness and choose targets based primarily on accessibility and opportunity. As women age into retirement, they may be less likely to be out in the public sphere and, therefore, less accessible in locations that would make them vulnerable to motivated strangers. Also, because married women are likely to spend a large amount of leisure time with spouses or children, they may be less likely to be perceived as a suitable target by a motivated stranger. Overall, these findings suggest that stranger victimization may be more likely to be a crime of random opportunity or circumstances in which a motivated offender stumbles upon a suitable target.
For acquaintance victimization, both age and marital status were strong predictors of victimization. The effect of age was most pronounced for acquaintance victimization with women in their 20s and 30s at highest risk of victimization and risk declining progressively with age. Never married, separated, and divorced women were also at significantly higher risk of victimization than married women. These findings are consistent with prior research that suggests that young, unmarried women’s lifestyles often involve active participation in socialization activities with unfamiliar males, frequent and excessive alcohol consumption, and willingness to engage in casual sexual experiences, all of which may increase risk of victimization by regularly putting them in contact with motivated offenders. Once again, it is important to note that these lifestyles factors can only be considered “risk factors” in the context of a rape culture that legitimizes violence against women when women choose to exercise their autonomy and sexuality in particular ways. Therefore, the findings of this study highlight the need for further public education about sexual freedom and the meaning of consent.
For intimate partner victimization, age had some effect on risk; however, marital status was the strongest predictor of victimization. Although there was no difference in risk of victimization for women between the ages of 18 and 49, women in their 50s and older were at reduced risk of intimate partner victimization, which could indicate that with age and experience, women may be better able to identify and leave abusive or unhealthy relationships and be more capable guardians of themselves. Although never married, separated, and divorced women were all significantly more likely to experience intimate partner victimization than married women, one of the most striking findings of the study was that separated women were the most at risk for intimate partner victimization. In fact, separated women were six times as likely to be a victim of an intimate partner assault, as compared with married women. Separated women were also at high risk of experiencing acquaintance victimization. In fact, they were four times as likely as married women to have experienced an acquaintance assault.
Two possible explanations might help theorize these findings. First, separation could be conceptualized as a period of transition and instability that might involve repeated violence between spouses in the process of separating. Although there has been limited research on this topic, prior research suggests that immediately after leaving a relationship, women may experience increased risk of violent victimization (for reviews of this literature see Brownridge, 2006; DeKeseredy, Rogness, & Schwartz, 2004; Fleury, Sullivan, & Bybee, 2000). Spouses from whom women are trying to separate may use sexual violence as a tool to reassert power and control or to punish their partners for perceived mistreatment or humiliation. Second, separation might be conceptualized as a period of enhanced vulnerability resulting in “rebound” relationships and lifestyles that could raise risk levels for victimization in newly formed relationships or between acquaintances. Again, these findings should not be construed to suggest that women should not be free to socialize as they see fit; rather, the findings suggest that, within a rape culture, they may be targeted by motivated offenders due to their perceived vulnerability and powerlessness. Follow-up research is needed to better understand why separated women’s relative risk levels are so much higher than married women and substantially higher than never married or divorced women as well.
Another interesting finding of the current study was the relationship between cohabitation and risk of sexual victimization. Women living alone were significantly more likely to experience sexual assault victimization than women living with other people. This may be due to the fact that women who live alone are also likely to engage in routine activities, such as grocery shopping or running errands, alone and without the added “guardianship” of cohabitants. Women living alone may also be more likely to participate in socialization activities, such as dating or “hooking up” with unfamiliar males that could potentially make them more vulnerable to being specifically targeted by a motivated offender. Future research should continue to explore this relationship and consider whether cohabitation with romantic partners differs from cohabitation with family or unrelated roommates.
Finally, high levels of household crime were also significantly associated with higher risk of sexual victimization. This finding suggests that if a woman’s home environment is a high-crime or violent environment, she may be at higher risk of experiencing sexual violence as one of many types of violent victimization that she may be exposed to at home. Considering that the effect of household crime was most pronounced for intimate partner victimization experiences, one way to interpret the results would be that within violent households, rape and sexual assault are simply part of a broader mix of violent and controlling behaviors used by one partner to manipulate and dominate the other partner.
Although this study makes a number of contributions to the literature on sexual victimization, some limitations of the study should be considered. The use of the NCVS as the primary data source for the study may have limited the analysis in some ways. The very small numbers of sexual victimizations reported in the NCVS each year might be attributed to the methods used by the NCVS to elicit responses about sexual victimization. Nevertheless, due to the small numbers reported each year, it is virtually impossible to conduct meaningful quantitative analysis on risk factors for victimization using annual statistics alone. To address this issue, the current study used a case control sampling strategy that combined data from several years of the NCVS. This case control sampling strategy was very useful for quantitative modeling and analysis; however, this statistical technique poses some challenging statistical problems in error structures, the effects of which are difficult to estimate and control. Therefore, the results must be considered with some caution and, although confidence in the directions of the relationships might remain strong, assessment of the magnitude of some of the relationships in the model need further replication in future studies. This study also utilized a very large sample and it should be noted that large samples can potentially skew results toward finding significant relationships for all variables even when no such relationship exists; however, in the models run in this study, not all variables were significant and, furthermore, confidence in relative relationships should not be compromised due to the sample size. Finally, as awareness about rape culture continues to grow, we might consider ways in which we can move away from characterization of women’s vulnerabilities in ways that might be interpreted as victim-blaming. Part of that may come from using our knowledge of women’s vulnerabilities to shape discourse on rape culture and how it affects women’s daily life experiences.
Conclusion
The results of the current study suggest that the effects of age and marital status on risk of sexual victimization differ based on the situational contexts of victimization experiences, specifically victim–offender relationships. These findings both support and challenge assumptions about the relationship between demographic characteristics and risk of sexual victimization and suggest that further model specification is required. Future analyses should replicate the findings of this study and consider differences in risk factors across different crime types. Future research should also assess whether other key demographic variables, such as race and ethnicity, vary in their relative impact on victimization risk across different contexts of sexual violence. Furthermore, ideally, future research should use data sets with more precise lifestyles and routine activities measures to further parse the impact of demographic variables on victimization risk. These studies will likely aid in the development of theories of sexual violence that take into account both criminological and feminist theoretical orientations.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.