1. Introduction
Reporting crime to the police is well established as a critical issue for the administration of justice. Most crimes become known to the police through victims, witnesses, or other citizens. In traditional practice, the police react to criminal events rather than proactively prevent them. Police rely heavily on both public knowledge and their assistance to become aware that a crime even occurred. Thus, without public cooperation, the police may be unable to maintain public safety (Reference SkoganSkogan 1984; Reference Tyler and FaganTyler and Fagan 2008; Reference WhitakerWhitaker 1980). Police notification can also benefit victims by providing them with information about services that can help address socioemotional problems and other needs arising from victimization (Reference ZaykowskiZaykowski 2014). Disadvantaged groups face greater risk of victimization than more affluent and structurally advantaged groups, and consequently may benefit the most from police protection and access to victim services (Reference Desmond, Papachristos and KirkDesmond et al. 2016; Reference HarrellHarrell 2007; Reference SeredSered 2014). However, multiple studies reveal that many crimes are never reported to the police. Recent estimates indicate that fewer than half of all nonfatal violent crimes are reported (Reference Morgan and TrumanMorgan and Truman 2018).
Many explorations of the gap in reported and unreported crimes focus on crime severity (i.e., injury, use of weapons, greater perceived harm and more expensive property loss) as a key motivator for mobilizing the police (Reference Baumer and LauritsenBaumer and Lauritsen 2010; Reference Conaway and LohrConaway and Lohr 1994; Reference Goudriaan, Wittebrood and NieuwbeertaGoudriaan et al. 2006; Reference KaukinenKaukinen 2004; Reference McCart, Smith and SawyerMcCart et al. 2010; Reference SkoganSkogan 1984). At the same time, this body of literature has predominantly ignored explaining why certain groups may be less likely to report similar crimes, which is an arguably more important concern (for a notable exception, see e.g., Reference Xie and LauritsenXie and Lauritsen 2012). Other research finds that marginalized groups such as racial and ethnic minorities and lower-class individuals hold more cynical attitudes toward the police as a result of negative personal and vicarious encounters (Reference AndersonAnderson 1999; Reference Berg, Stewart, Intravia, Warren and SimonsBerg et al. 2016; Reference Brown and Reed BenedictBrown and Reed Benedict 2002; Reference Carr, Napolitano and KeatingCarr et al. 2007; Reference Nadal, Davidoff, Allicock, Serpe and ErazoNadal et al. 2017; Reference Sampson and BartuschSampson and Bartusch 1998), so it would not be surprising that these groups would also be unlikely to call the police for assistance when victimized (Reference Desmond, Papachristos and KirkDesmond et al. 2016).
Nonetheless, despite evidence pointing to strained relationships between marginalized groups and the police, research suggests that social marginality is not related to reporting or may be associated with greater police mobilization in some circumstances. Findings on the relationship between race and class and reporting have been mixed, though the majority of research finds that females are more likely than males to report (Reference Baumer EricBaumer Eric 2002; Reference Berg, Slocum and LoeberBerg et al. 2013; Reference Felson, Messner and HoskinFelson et al. 1999; Reference Tarling and MorrisTarling and Morris 2010; Reference Wolitzky-Taylor, Resnick, McCauley, Amstadter, Kilpatrick and RuggieroWolitzky-Taylor et al. 2011; Reference ZavalaZavala 2010). Recent studies have highlighted possible reasons for reporting among marginalized groups, including lack of other alternatives (Reference BellBell 2016; Reference Xie and BaumerXie and Baumer 2019; Reference Xie and LauritsenXie and Lauritsen 2012). This research has also found that negative attitudes toward the police can be superseded by instrumental concerns such as immediate need for help (Reference Hagan, McCarthy, Herda and Cann ChandrasekherHagan et al. 2018; Reference SlocumSlocum 2018). This literature points to a potential paradox that despite holding cynical attitudes toward the police, some marginalized groups may actually be more likely to seek their help than people who are indifferent or have favorable views of the police.
Although there has been growing evidence of this paradox, research to date has not adequately addressed how reporting to the police varies across social groups within the context of criminal victimization. Questions still remain due to conflicting findings in prior research, particularly with regard to the relationship between race and class and reporting. This paper seeks to demonstrate that using an intersectional lens can resolve mixed findings regarding race and class by addressing the differing structural positions of women and men, including their relationships to economic and social control institutions (Reference AndersonAnderson 1999; Reference BellBell 2016; Reference GoffmanGoffman 2009; Reference RiosRios 2011). Intersectionality holds that social identities cannot be understood as independent, but rather are intersecting and constituted through structural arrangements of power and privilege (e.g., Reference Burgess-ProctorBurgess-Proctor 2006; Reference Collins and BilgeCollins and Bilge 2016; Reference CrenshawCrenshaw 1991). Failure to address how race, class, and sex/genderFootnote 1 together shape victimization experiences results in an incomplete understanding of crime reporting behavior.
In this study, we examine whether the combination of multiple marginalities (i.e., across sex, race and ethnicity, and social class) increases the likelihood that crime victims notify the police. We use a nationally representative sample of victims and control for incident characteristics. We find that the intersectional approach shows us a more nuanced understanding of crime reporting behavior. Not only are females much more likely to report than males, the effects of race and class function differently for males and females.
2. Social Status and Reporting to the Police
Extensive research suggests that social status is linked to different experiences with the legal system, which in turn may influence the likelihood that law is mobilized (e.g., Reference AndersonAnderson 1999; Reference BlackBlack 1976; Reference Carr, Napolitano and KeatingCarr et al. 2007; Reference MerryMerry 1990; Reference NielsenNielsen 2000). For example, Reference BlackBlack (1976) argued that the distribution of resources across society can influence the mobilization of law. In particular, he argued that poor people have “less law,” so they will be less likely to mobilize the police and have their problems addressed through the legal system. Scholars extended this hypothesis by arguing that in the U.S. victims of lower social status also include racial minorities (Reference Xie and LauritsenXie and Lauritsen 2012; see also Reference Avakame, Fyfe and McCoyAvakame et al. 1999).
Attitudes toward the law, which vary by social status, are also argued to play a large role in reporting to the police (Reference Desmond, Papachristos and KirkDesmond et al. 2016). For example, scholars have observed that disadvantaged populations such as racial and ethnic minorities and low-income persons have greater legal cynicism than white and more affluent populations (Reference AndersonAnderson 1999; Reference Brown and Reed BenedictBrown and Reed Benedict 2002; Reference Carr, Napolitano and KeatingCarr et al. 2007; Reference Hagan, Shedd and PayneHagan et al. 2005; Reference Hagan, McCarthy, Herda and Cann ChandrasekherHagan et al. 2018; Reference Kirk and PapachristosKirk and Papachristos 2011; Reference Nadal, Davidoff, Allicock, Serpe and ErazoNadal et al. 2017; Reference NielsenNielsen 2000; Reference PeckPeck 2015; Reference Sampson and BartuschSampson and Bartusch 1998; Reference Sun and WuSun and Wu 2018; Reference WeitzerWeitzer 2014; Reference Weitzer and TuchWeitzer and Tuch 2006; Reference Wu, Lake and CaoWu et al. 2015). Although populations with high levels of legal cynicism believe in the “substance of law,” they find that its institutions and agents are “illegitimate, unresponsive, and ill-equipped to ensure public safety” (Reference Kirk and PapachristosKirk and Papachristos 2011: 1191). The development of legal cynicism is a consequence of both “over-policing” tactics such as stop and frisk, excessive use of force and police brutality, and “under-policing,” in which law enforcement officers are slow to respond to calls for service assistance and are unsympathetic to the needs of victims (Reference AndersonAnderson 1999; Reference Carr, Napolitano and KeatingCarr et al. 2007; Reference Desmond, Papachristos and KirkDesmond et al. 2016; Reference Kirk and MatsudaKirk and Matsuda 2011; Reference Kirk and PapachristosKirk and Papachristos 2011; Reference RiosRios 2011). Legal cynicism, in turn, is related to reduced confidence in the police and attempts to resolve problems through self-help, including the use of violence to resolve disputes (Reference Kirk and PapachristosKirk and Papachristos 2011).
Some empirical research provides evidence to support theoretical arguments that racial and ethnic minorities and lower social-class victims may be less likely to report to the police (Reference AndersonAnderson 1999; Reference Baumer EricBaumer Eric 2002; Reference Carr, Napolitano and KeatingCarr et al. 2007; Reference MillerMiller 2008; Reference Slocum, Taylor, Brick and EsbensenSlocum et al. 2010; Reference Sokoloff and DupontSokoloff and Dupont 2005). For example, in his ethnography of a Philadelphia community, Reference AndersonAnderson (1999) argues that economically disadvantaged African Americans only call the police in a dire emergency, although he did not specify what this would include. To do otherwise would be counter to the cultural orientation that police cannot be trusted and would put one at risk for victimization from others in the neighborhood (Reference AndersonAnderson 1999). In a study of lower-income African-American women in one of the poorest sections of St. Louis, participants asserted that calling the police would not help them if they were victimized, police would be slow to respond or not show up at all, and in general, calling for help could just make the situation worse. Neighborhood norms emphasized minding one's own business and to avoid snitching to the police (Reference MillerMiller 2008). Other research has suggested that women of color, particularly African-American women, may be reluctant to report to the police because they do not want their intimate partners to be mistreated by the police and because doing so would contribute to stereotypes about race and crime (Reference CollinsCollins 1998; Reference Sokoloff and DupontSokoloff and Dupont 2005).
Despite some research suggesting racial and ethnic minorities would be opposed to calling the police, the majority of extant research finds that that race and ethnicity have either no effect (Reference Baumer EricBaumer Eric 2002; Reference Berg, Slocum and LoeberBerg et al. 2013; Reference Felson, Messner and HoskinFelson et al. 1999; Reference Tarling and MorrisTarling and Morris 2010; Reference Wolitzky-Taylor, Resnick, McCauley, Amstadter, Kilpatrick and RuggieroWolitzky-Taylor et al. 2011; Reference ZavalaZavala 2010) or a significantly higher likelihood of reporting incidents involving black victims compared to white victims (Reference Akers and KaukinenAkers and Kaukinen 2009; Reference BachmanBachman 1998; Reference Bachman and CokerBachman and Coker 1995; Reference Felson, Messner, Hoskin and DeaneFelson et al. 2002; Reference Kaylen and PridemoreKaylen and Pridemore 2015; Reference Rennison, Gover, Bosick and DodgeRennison et al. 2011; Reference Webb and MarshallWebb and Marshall 1989; Reference Xie and LauritsenXie and Lauritsen 2012).Footnote 2 For example, contrary to their hypothesis, when considering the race of the victim and offender, Reference Xie and LauritsenXie and Lauritsen (2012) found that black victims of a black offender had the greatest probability of reporting. Research examining reporting among Hispanic and Latino victims has yielded mixed findings depending on the type of crime examined (Reference Baumer and LauritsenBaumer and Lauritsen 2010; Reference RennisonRennison 2007). For example, Reference RennisonRennison (2007) found that, Hispanic victims were less likely to report sexual assault and completed robbery than non-Hispanic white victims, but more likely to report simple assault.
Research exploring indicators of economic disadvantage such as income, poverty status, education level, and home ownership has also revealed mixed findings in how or even whether they influence reporting behavior. While some research has found that economic disadvantage is also associated with a higher likelihood of reporting (Reference Kaylen and PridemoreKaylen and Pridemore 2015), other research has found no effect (Reference Akers and KaukinenAkers and Kaukinen 2009; Reference Bachman and CokerBachman and Coker 1995; Reference Gottfredson and HindelangGottfredson and Hindelang 1979; Reference ZaykowskiZaykowski 2013). A few studies have found that the effect varies by type of crime, measure of socioeconomic status, and victim characteristics (Reference Baumer EricBaumer Eric 2002; Reference Baumer and LauritsenBaumer and Lauritsen 2010; Reference Berg, Slocum and LoeberBerg et al. 2013; Reference RennisonRennison 2007). For example, Reference Baumer and LauritsenBaumer and Lauritsen (2010) examined household income, educational attainment, and home ownership. Home ownership was associated with a higher likelihood of reporting robbery and aggravated assault, but had no effect on simple assault or rape/sexual assault reporting. Higher educational attainment was associated with a greater likelihood of reporting robbery, but lower likelihood of reporting simple assault. Higher household income was not associated with reporting, except for simple assault. However, this study was limited in not distinguishing whether the crime was reported by the victim or by a third party (i.e., someone other than the victim): the effects of socioeconomic status could be obscured as third parties may be unaware of the victim's economic status and consequently the impact of the victim's decision to report remains unclear. No established consensus emerges from past work about the relationship between socioeconomic disadvantage, including race and ethnicity and social class and the likelihood of reporting to the police.
3. Theoretical Framework
These mixed and seemingly contradictory findings have led scholars to reconsider Reference BlackBlack's (1976) theory of law, recognizing a “paradox” that individuals who hold marginal social positions may actually be more likely to call the police (Reference BellBell 2016; Reference Hagan, McCarthy, Herda and Cann ChandrasekherHagan et al. 2018; Reference Xie and LauritsenXie and Lauritsen 2012). Reference BlackBlack (1976), like other scholars of his time and since, considered each form of disadvantage as a distinct factor that could be measured while holding all other factors equal. By contrast, we argue that the relationship between social location and reporting is more complex than previously articulated. To help illustrate this complexity, we turn to intersectionality as a theoretical framework. Intersectionality provides a lens through which we can consider how the effects of race and class may intersect with or be moderated by gender. Men and women have different structural positions that shape their risk for victimization, experiences with the police, and social constructions of victimhood (i.e., who is more likely to be treated as a “victim”) (e.g., Reference AndersonAnderson 1999; Reference CrenshawCrenshaw 1991; Reference MadrizMadriz 1997).
In its simplest form, intersectionality theory emphasizes how gender, race, and class combine with many other factors, to produce unique experiences of power and oppression that vary substantively across individuals and groups (Reference CollinsCollins 2000; Reference CrenshawCrenshaw 1991; Reference PotterPotter 2013). Using intersectionality as a lens to understand reporting moves beyond viewing the effect of multiple marginalities as “additive” (i.e., having two disadvantages) and rejects the idea that studying each factor separately can uncover the complete experience (Reference Burgess-ProctorBurgess-Proctor 2006; Reference DalyDaly 1993). For example, viewing race, class, gender as distinct analytical categories is not the same as examining the experience of labor market experiences through racialized constructions of gender (e.g., stereotypes distinctly associated with black women and white women) (Reference Browne and MisraBrowne and Misra 2003). Instead, social location should be viewed as multiplicative, an interaction between different marginal statuses (Reference Burgess-ProctorBurgess-Proctor 2006; Reference DalyDaly 1993).
While early intersectional studies focused on the experiences of poor black women (e.g., Reference CrenshawCrenshaw 1991), more recent studies have taken an increasingly complex view of different social locations and factors that simultaneously privilege and disadvantage individuals and groups within social structures of power and inequality. This work argues that the meaning of structural categories is not static and in fact changes over time and across social settings, and has been somewhat critical toward traditional perspectives in intersectionality which remain reductive in the nature of social categories and their meanings (Reference Collins and BilgeCollins and Bilge 2016; Reference Desmond and EmirbayerDesmond and Emirbayer 2009; Reference McCallMcCall 2005). Reference McCallMcCall (2005), for example, found that the extent of race and class wage inequalities differed by gender, but that this effect varied depending on which city was examined. Without viewing these relationships through an intersectional lens, the contextual effects could not be observed.
The value of adding an intersectional perspective lies in understanding that the approach to quantitative analysis in which demographic characteristics function solely as model controls does not take into consideration the relationships between multiple social locations. An intersectional perspective allows investigation of the nature and extent of such differences in reporting crime to the police, moving beyond standard indicators, and adding a level of complexity. It is with this perspective that we approach reporting criminal victimization.
3.1 Sex Differences in Reporting
It is well known that criminal victimization is strongly patterned by gender. Although research suggests that males are more likely to be victims of crime, particularly with regard to most forms of serious violence, females are more likely to be victims of sexual assault. Females are also more likely than men to be victims at the hands of intimate perpetrators than strangers and acquaintances, to be victimized in private places, to experience injuries as a result of being a victim, and to report socioemotional problems resulting from victimization than males (Reference Langton and TrumanLangton and Truman 2014; Reference Melton and BelknapMelton and Belknap 2003; Reference Tjaden and ThoennesTjaden and Thoennes 2000; Reference ZaykowskiZaykowski 2014). Women are also more likely to be viewed as sympathetic victims than men (Reference ChristieChristie 1986; Reference MadrizMadriz 1997). The feminization of victimhood as weak and dependent places women in an arguably better position to “claim” the position of victim (Reference Holstein and MillerHolstein and Miller 1990).
Based on this research, we would anticipate that overall, female victims would be more likely to report to the police than male victims (H1). This hypothesis would also be consistent with prior research on crime reporting (Reference BachmanBachman 1998; Reference Bachman and CokerBachman and Coker 1995; Reference MillerMiller 2008; Reference PotterPotter 2008; Reference Stanko and HobdellStanko and Hobdell 1993; Reference WeissWeiss 2010).
3.2 Female Victimization and Reporting by Race, Ethnicity, and Social Class
Victimization experiences are not the same for all females. Some scholars have argued that society considers affluent white women to be “ideal victims” because they engender the greatest public concern and political support (Reference ChristieChristie 1986; Reference Kulig and CullenKulig and Cullen 2017; Reference MadrizMadriz 1997; Reference NeelyNeely 2015; Reference RitchieRitchie 2017). This is despite the reality that Hispanic women have similar victimization rates compared to white women (Reference RennisonRennison 2002; Reference Rennison and WelchansRennison and Welchans 2000) and black women, particularly those who are economically disadvantaged, have significantly higher rates of victimization (Reference CrenshawCrenshaw 1991; Reference HarrellHarrell 2007; Reference MillerMiller 2008; Reference PotterPotter 2008). Research also suggests that women of color themselves imagine a “typical” victim as a female who is white and middle class (Reference MadrizMadriz 1997) and are less likely to label themselves as “victims” (Reference CollinsCollins 1998; Reference PotterPotter 2008).
The criminal justice system also tends to favor white women as victims over women of color, particularly when compared to black women. Police exert more effort responding to white and female victims including faster response time and follow-up after the crime (Reference HowertonHowerton 2006). White women are less likely to be abused and harassed by the police, and do not experience racism as do women of color in their interactions with the police (Reference RitchieRitchie 2017). Criminal justice proceedings involving white female victims are also more likely to result in harsher punishment for offenders (Reference Holcomb, Williams and DemuthHolcomb et al. 2004; Reference Williams, Demuth and HolcombWilliams et al. 2007). Women of color are overrepresented among females as suspects including police stops, arrests, and experiences with police misconduct (Reference RitchieRitchie 2017; Reference Weitzer and TuchWeitzer and Tuch 2002).
Social beliefs about victimhood, who is perceived to be (and treated as) a more “ideal victim,” and experiences with the police suggest that women of color, particularly black women, would be less likely to call the police than white women. Evidence from the broader help-seeking literature would also support this position. White women are more frequent users of social and mental health services and may be more likely to seek help from friends and family for health issues than women of color as a result of victimization (Reference KaukinenKaukinen 2004; Reference SnowdenSnowden 1998). Additional burdens such as child care responsibilities, poverty, and unemployment resulting from race and class oppression limit access to options that are more often available to affluent and middle-class white women when reporting victimization or seeking help (Reference CrenshawCrenshaw 1991). If we consider reporting to police as similar behavior to other forms of help seeking and who is considered an “ideal victim,” from these literatures we would hypothesize that white females from more affluent backgrounds would be more likely to call the police than poor females of color, particularly black females (H2a).
Other research, however, suggests that intersecting forms of oppression for minority women may not prohibit reporting, but in fact be used as a tool to face challenges associated with multiple marginalities that cannot be addressed via other social systems. Economic disadvantages and lack of other cultural resources may leave the police to be the only option for economically disadvantaged women. In a study of poor African-American mothers, Reference BellBell (2016) found that participants relied on the police to deal with certain types of problems—particularly problems inside or near their homes. Participants viewed the police as ineffective for street crime but very useful for handling domestic problems and threats. Participants also expressed “situational trust” by using the police to control children and to mobilize social welfare institutions. Mothers believed that calling the police regarding their own children reduced the chances that their children would be removed by protective services, a fate that was more likely if the child was reported to the police by someone else (Reference BellBell 2016).
Additional research has found that the police are mobilized by disadvantaged individuals to control intimate partners (Reference GoffmanGoffman 2009) and children (Reference RiosRios 2011). Work by Reference Messing, Becerra, Ward-Lasher and AndroffMessing et al. (2015) found that although Latinas in their study feared deportation, the ability to develop trust with law enforcement enhanced intentions to report to the police. In other words, disadvantaged women of color may reach out to the police to address pressing needs—such as the need for protection and help managing other problems—despite concerns about legal consequences. While women of color are more likely than white women to have negative experiences with police and the law more generally, their economic and other social disadvantages supersede their concerns about the police, with whom they seek an uneasy alliance due to limited alternatives. Based on these studies, we present an alternative hypothesis to H2a that females of color, particularly disadvantaged black females, would be more likely to call the police than white affluent females (H2b).
3.3 Male Victimization and Reporting by Race, Ethnicity, and Social Class
Help seeking by men runs counter to dominant gender expectations embodied by hegemonic masculinity (Reference ConnellConnell 2000), which emphasizes strength and fearlessness (Reference AndersonAnderson 1999; Reference Burcar and ÅkerströmBurcar and Åkerström 2009; Reference GoodeyGoodey 1997; Reference Stanko and HobdellStanko and Hobdell 1993; Reference WeissWeiss 2010). Rather than seek help, men are more likely to take matters into their own hands, particularly in areas with high levels of legal cynicism (Reference BlackBlack 1983; Reference Kirk and PapachristosKirk and Papachristos 2011; Reference Piquero, Intravia, Stewart, Piquero, Gertz and BrattonPiquero et al. 2012). Research suggests that males who cannot conform to traditional gender roles such as being able to financially provide for one's family are more likely to adopt street codes, including the use of violence and avoiding the police to handle disputes (Reference AndersonAnderson 1999; Reference MesserschmidtMesserschmidt 1993).
Black men have the least confidence in the police, and are most likely to experience racial discrimination, police use of force, and racial profiling relative to other groups, particularly when compared to white men (Reference Brown and Reed BenedictBrown and Reed Benedict 2002; Reference Nadal, Davidoff, Allicock, Serpe and ErazoNadal et al. 2017). Poor men of color are also likely to have direct negative encounters with police both in the community and for domestic violence calls (Reference BellBell 2016; Reference Felson, Messner, Hoskin and DeaneFelson et al. 2002; Reference GoffmanGoffman 2009). It is not surprising then that white men have more favorable attitudes toward the police, and also tend to support more law-oriented perspectives (Reference Brown and Reed BenedictBrown and Reed Benedict 2002; Reference Nadal, Davidoff, Allicock, Serpe and ErazoNadal et al. 2017; Reference NielsenNielsen 2000). Furthermore, research shows that black men and boys are not perceived as “ideal victims” (Reference ChristieChristie 1986; Reference MadrizMadriz 1997; Reference Williams, Demuth and HolcombWilliams et al. 2007). Although the criminal justice system disproportionately punishes black male offenders, cases in which the victim is a black male result in more lenient punishment for perpetrators (Reference Williams, Demuth and HolcombWilliams et al. 2007).
Hispanic and Latino men tend to have more positive experiences and impressions of the police relative to black males, but less confidence in the police than white men (Reference Brown and Reed BenedictBrown and Reed Benedict 2002; Reference Hagan, Shedd and PayneHagan et al. 2005; Reference Nadal, Davidoff, Allicock, Serpe and ErazoNadal et al. 2017; Reference PeckPeck 2015; Reference Sun and WuSun and Wu 2018; Reference WeitzerWeitzer 2014; Reference Weitzer and TuchWeitzer and Tuch 2006). Reference Hagan, McCarthy, Herda and Cann ChandrasekherHagan et al. (2018) argues that these differences reflect the longer history with racism and discrimination against African Americans and less alienation overall among Hispanics and Latinos. However this is not always the case (Reference RiosRios 2011), and the increasing growth and visibility of immigration from Mexico and Central America may change these trends for Hispanic and Latino victims, regardless of whether the individual is in fact an immigrant (Reference Landale, Oropesa and NoahLandale et al. 2017). Therefore, while economically disadvantaged minority women may have “situational trust” with the police, the same is not likely to be true of their male counterparts (Reference AndersonAnderson 1999; Reference BellBell 2016; Reference RiosRios 2011). Following this research, we would expect that disadvantaged males, particularly black males, and to a lesser extent Hispanic males, would be less likely to report to the police than all other intersectional groups (H3).
4. Current Study
Despite growing interest in intersectional approaches and empirical qualitative literature examining reporting to the police, few quantitative studies have considered the effects of race, class, and gender/sex as anything beyond control variables. Our primary aim is to explore whether the intersectionality perspective can help explain victims' reporting behavior beyond the “additive” model, in which race/ethnicity, sex, and social class are considered distinguishable analytic categories. In other words, we wanted to see if social location was associated with reporting through the intersection of multiple identities. To accomplish this, we modeled victim reporting first using only main effects, then adding two- and three-way interactions. We further contextualize our findings by looking at the reasons victims gave for reporting or not reporting to the police.
We also recognize that the types of victimization experienced by different groups vary across race, ethnicity, sex, and class. Therefore, we examine demographic intersections when “all else is held equal” (Reference BlackBlack 1976, Reference Black1979), controlling for key characteristics of the incident. In particular, an extensive literature documents the influence of crime severity on reporting (Reference Baumer and LauritsenBaumer and Lauritsen 2010; Reference Conaway and LohrConaway and Lohr 1994; Reference Goudriaan, Wittebrood and NieuwbeertaGoudriaan et al. 2006; Reference KaukinenKaukinen 2004; Reference McCart, Smith and SawyerMcCart et al. 2010; Reference SkoganSkogan 1984). This allows us to examine whether differences in reporting behavior are in fact a function of social location extending beyond the differences in the types of crimes experienced by demographic groups.
5. Data and Methods
5.1 Data
Data for this study come from the National Crime Victimization Survey (NCVS) (United States Department of Justice 2016, 2017). The NCVS uses a multistage cluster sample of households in the United States. All eligible household members aged 12 or older who agree to participate are interviewed seven times over the course of 3 years using a rotating panel design. The first interview is conducted in person, while subsequent interviews are conducted either in person or over the phone. Each eligible participant in the household is first screened to see if he or she experienced one of several nonfatal forms of victimization, including property crime, violent physical crime, and sexual offenses. If a participant meets these criteria, he or she is asked follow-up questions about the incident, perpetrator, and help-seeking behavior.
Sample sizes and response rates vary from year to year. In 2016, response rates for households (N = 134,690) and persons within households (N = 224,520) were 78 and 84 percent, respectively. Weights are provided by the NCVS to account for differences in characteristics of the sample and population, survey nonresponse, and the multistage cluster sample design. Further information about the NCVS can be found on the Bureau of Justice Statistics web site (Bureau of Justice Statistics 2019). The current study focused on personal violent victimizations (robbery, simple and aggravated physical assault, and rape and sexual assault) occurring in the United States from interviews conducted between 1994 and 2016 (N = 48,008).
5.2 Measures
The dependent variable, victim reported to police, measured whether the victim reported or not (1 = reported by the victim). The key independent variables included demographic characteristics of the victim—race/ethnicity (black non-Hispanic, Hispanic, other race non-Hispanic, and white non-Hispanic [reference]) and sex (female = 1).Footnote 3 Socioeconomic status was measured with four different variables: renter, unemployed, less than high school education, and poverty status.Footnote 4
Renters included those without homeownership, while homeowners include those who own their home or have a home mortgage. Employment status and education importantly vary with age. In the NCVS, employment status is only asked of participants age 16 and older. However, a 16-year-old or a college student who is unemployed may not necessarily be considered “economically disadvantaged” if they are not employed. To account for this, participants who were attending school (grade school or college) were excluded from the unemployed group. Education is also contingent upon age. Not completing high school for someone who is 15 is fundamentally different than an older adult. For participants under 20 years old, the “principal person” in the household's educational attainment was used. The principal person is usually considered the “household head” in the survey interview process (Reference Berzofsky, Smiley-McDonald, Moore and KrebsBerzofsky et al. 2014).
The NCVS does not include poverty status, but it can be approximated with the household income, year, and household size variables. Household income unfortunately contains a high level of missing data (15.2 percent), in part because income is not asked at each interview wave. Research has found that participants' prior interview wave income is a good predictor of future income (Reference Berzofsky, Smiley-McDonald, Moore and KrebsBerzofsky et al. 2014). In the analyses here, income data available from previous waves were used to replace missing data in later waves (Reference Berzofsky, Smiley-McDonald, Moore and KrebsBerzofsky et al. 2014; Reference SlocumSlocum 2018). This procedure reduced the missing data to 7.2 percent of unweighted cases. Income was then assigned a poverty status (below poverty or not) using federal poverty guidelines for the contiguous United States (c.f. Reference SlocumSlocum 2018).
The models also control for other victim characteristics: marital status (never married, married [reference], other marital status), Metropolitan Statistical Area (MSA) status (center city MSA = 1 [reference]; not center city MSA = 2; outside of MSA = 3), age (continuous), years lived at current address (continuous), and number of children in the household under 12 (continuous). Type of crime was measured as rape/sexual assault, physical assault [reference], and robbery. Other incident characteristics included the victim relationship to the offender when there was a single offender (intimate partner, nonintimate known offender, stranger [reference], and unknown relationship). When there was more than one offender, the relationship was coded as multiple offenders. Models also controlled for victim injury (injured = 1), presence of a weapon (any weapon used by offender = 1), and whether there was a bystander present (bystander = 1). The location of the incident included the victim's home or “other” location. Offender characteristics included the offender's sex (male, female, other/unknown), age (child, adult, or unknown), and race (white, black, or other/unknown). For multiple offenders, sex and race are coded to reflect the status of the majority of offenders. For example, if the victim reported “mostly male” offenders, offender sex was coded as male. For instances where no group was predominant among offenders, offender composition was coded as other/unknown. This is similar to prior treatments of these variables (c.f. Reference Baumer and LauritsenBaumer and Lauritsen 2010). Analyses also controlled for whether the incident was a series offense (i.e., a victimization with several similar indistinguishable events), whether the incident occurred in the same city/town where the participant lived, and the interview year.
5.3 Analytic Approach
Sample weights were used in the analyses to account for the NCVS's complex sampling design. Although research using the NCVS often uses unweighted cases, these analyses may produce biased results due to the survey design. The survey data were adjusted using the guidance provided in the NCVS technical documentation (Appendix F) (Reference Shook-Sa, Lance Couzens and BerzofskyShook-Sa et al. 2014). A comparison of the analyses using unweighted data (see Appendix B) suggested differences in the effects of key variables, which further justified the survey adjusted weights. Models were estimated with Stata 16.0 using the svy command and subpopulation of violent crime victims. Stata's svy command includes robust errors as the default with this method and adjusts for the nonindependence of observations (Reference StataCorpStataCorp 2013). Data from 1992 and 1993 did not include sample weight information, and therefore were not included in the analyses. In addition, victimizations occurring outside of the United States were excluded (Reference Shook-Sa, Lance Couzens and BerzofskyShook-Sa et al. 2014). Logistic regression was used to predict reporting to the police. Models were first estimated using the main effects and then analyses explored two- and three-way interactions between race/ethnicity, sex, and indicators of economic disadvantage. Listwise deletion was used for the missing cases. Supplemental analyses descriptively examined the most important reasons for reporting and not reporting among the groups with notable findings using the weighted survey data.
6. Results
6.1 Sample Characteristics
Table 1 presents the unweighted sample characteristics for violent crime victimizations.Footnote 5 Approximately one quarter of victimizations were reported to the police by the victim (26.2 percent), or 44,008 instances of victimization between 1994 and 2016, which become the basis for our analysis. In terms of victim characteristics, just under half of victims were female, with a majority white non-Hispanic (71.1 percent), followed by victims identifying as Hispanic (12.2 percent), black non-Hispanic (12.1 percent), and 4.6 percent identified as non-Hispanic other or mixed race. Just under half of victims were renters (48.2 percent), one-third had less than a high school education, nearly one in five were unemployed (17.5 percent), and about one quarter were from families in poverty (25.7 percent). Just over one quarter were married (27.0 percent) with 52.0 percent never married and 21.0 percent reporting another marital status (divorced, separated, or widowed). The sample included 38.5 percent of cases in which victims lived in a central city, 47.3 percent in a MSA, not in the center city, and 14.1 percent did not reside in an MSA. Participants' age ranged from 12 to 90 (M = 31.7; s.d. = 15.0). On average victims lived at their address for 5.7 years (range 0–83; s.d. = 7.7). The number of children under 12 in victim households ranged from 0 to 8 (M = 0.6; s.d. = 1.0).
Table 1. Sample Characteristics of Violent Crime Victims
Unweighted cases presented (N = 44,008).
Offender characteristics included the offender's sex, age, and race (according to the victim's perception). Most offenders (or offender groups) were male (75.7 percent); 16.5 percent of offenders were female. Offenders for whom sex was unknown or offending groups with about equal numbers of male and female offenders in an incident involving multiple perpetrators composed 7.8 percent of cases. Most offenders for whom the age was known were adults (69.7 percent) and 20.6 percent were under 18. The remaining offenders' ages were unknown or the victim was unsure (9.7 percent). A majority of offenders were described as white (55.1 percent), followed by black (24.9 percent), with the remainder of another or unknown race (20.0 percent).
The most common relationship between the victim and the offender with a single offender was known nonintimate (33.8 percent), followed by stranger (31.2 percent), and intimate partner (11.3 percent). In about one in five incidents (19.1 percent), the event involved multiple offenders (of any relationship) and in 4.5 percent of incidents, the relationship or the number of offenders was unknown. Most of the incidents were physical assaults (84.8 percent), followed by robberies (10.8 percent), and rape or sexual assault (4.4 percent). About one quarter of violent crimes resulted in an injury (25.6 percent) and involved a weapon (23.7 percent). Thirty-two percent of victimizations occurred at the victim's home and 71.8 percent occurred in the same city as the victim's residence. Almost two-third of incidents had a bystander present (62.2 percent) and 4.9 percent were considered a “series offense.”
6.2 Logistic Regression Analyses
The purpose of this study was to examine whether differences across social groups in reporting behavior can be explained by the multiplicative effects of multiple identities, as a way of exploring whether intersectionality can improve our understanding of reporting beyond the “additive” model. First, we estimated the main effects model (Table 2, Model A).Footnote 6 Controlling for other demographic and incident characteristics, the results indicate that females were more likely to notify the police compared to males (Exp[B] = 1.49; p<.001). There were no significant effects of race/ethnicity on reporting. Participants who were black, Hispanic, or other/mixed race were no more likely than white participants to report to the police. Similarly, indicators of economic disadvantage—poverty, unemployment, less than a high school education, and being a renter—were not significantly related to reporting.
Table 2. Logistic Regression Predicting Victim Reported to Police
Model A: N = 44,008 unweighted cases (175,832,621 weighted cases); F(31, 608) = 56.36. Prob > F = 0.000. Model B: N = 44,008 unweighted cases (175,832,621 weighted cases). F(41, 598) = 46.5. Prob > F = 0.000.
*p<.05; **p<.01; ***p<.001.
To assess whether demographic patterns in reporting could be better understood through intersectionality's “multiplicative marginality” perspective, we explored two- and three-way interactions. This enabled us to model the multiplicative effects of race/ethnicity, class, and sex. Table 2 (Model B) presents these results. Several two-way interactions were significant. The results indicate that when incorporating the two-way interactions, the main effect for sex remained significant (Exp[B] = 1.23; p<.001), indicating that females of any racial or class background were more likely than males to report to the police. Indicators of economic disadvantage and race/ethnicity remained nonsignificant in terms of their main effects, with the exception that poverty becomes significant (Exp[B] = 0.86; p<.05]. There were significant two-way interactions between race/ethnicity and sex, sex and economic disadvantage (poverty and education), and education and race/ethnicity.
The two-way interactions indicated that black females (Exp[B] = 1.39; p<.01), females with less than a high school degree (Exp[B] = 1.23; p<.05), and females living in families below the poverty line (Exp[B] = 1.37; p<.001) had an increased likelihood of reporting. In terms of interactions between race/ethnicity and social class, Hispanics with less than a high school education had a decreased likelihood of reporting (Exp[B] = 0.73, p<.01). Overall, the interactions (Model B) showed that the effects of poverty and education were moderated by sex and race/ethnicity.
These findings are also presented in Figures 1 and 2, which show the predicted probabilities of reporting (excluding the other/mixed race groups for parsimony) by sex, race/ethnicity, and economic disadvantage variables with significant interactions. Figure 1 shows the predicted probabilities of reporting by sex, race/ethnicity, and poverty status. As shown in Figure 1, the main effect of sex was evident in the large gap between the probability that females and males report to the police. In addition, the effect of poverty was different for females and males. Females in poverty were more likely to report to the police than females above poverty, whereas the effect for males was the reverse. Black females in poverty had the highest probability of reporting (.45) while black males in poverty with the lowest predicted probability (.24).
Figure 1. Predicted probabilities of victim reporting to the police for violent victimization by race/ethnicity, sex, and poverty status.
Figure 2. Predicted probabilities of victim reporting to the police for violent victimization by race/ethnicity, sex, and education.
Figure 2 shows the predicted probabilities of reporting by sex, race/ethnicity and education. A similar pattern in the relationship between economic disadvantage, sex, and race/ethnicity emerges with education. The graph displays the significant main effect of sex on the probability of reporting. Females with less education had the highest probability of reporting, with less-educated black females again with the highest probability of reporting (.46). Among the racial and ethnic groups, white and black males had similar predicted probabilities regardless of education. However, having a high school degree revealed different predicted probabilities for Hispanic males and females—when combined with higher education, they had increased probabilities of reporting. In contrast, Hispanic males without a high school degree had the lowest likelihood of reporting (.24).
6.3 Reasons for Reporting and Not Reporting to the Police
The NCVS data do not allow us to explore cultural orientations directly or determine whether particular attitudes shaped reporting decisions. However, some insights may be gained from participants' responses regarding the most important reasons why they reported (among those who did call the police) or why they did not report (among those who did not call the police) (Reference Felson, Messner, Hoskin and DeaneFelson et al. 2002; Slocum 2018). Supplemental analyses explored these reasons as they pertained to the hypotheses and regression results. Regression results supported H1 that females were more likely than males to report to the police. When examining the reasons for reporting, a greater proportion of females than males stated that the reason to report was to “stop the incident” (26.9 percent compared to 19.7 percent) and “to prevent more crimes” (22.8 percent compared to 13.2 percent). However, males were more likely than females to state that they reported “because it's a crime” (29.1 percent compared to 20.5 percent).
Although both poverty and education were found to be influential factors for reporting in the intersectional analysis, the current section focuses on poverty for the purpose of exploring reasons for reporting or not reporting.Footnote 7 Table 3 presents the reasons for reporting and not reporting by sex, race/ethnicity, and poverty status. Two alternative hypotheses were presented for H2. The results supported H2b that disadvantaged black females were more likely to report than affluent white females. The reasons for reporting to the police were similar for disadvantaged black females and affluent white females. However, when examining the reasons for not reporting, disadvantaged black females were more likely to state that they did not report due to “fear of reprisal” (15.0 percent compared to 6.5 percent), whereas affluent white females did not report because they “reported to another official” (21.5 percent compared to 10.4 percent).
Table 3. Most Important Reason for Reporting and not Reporting by Poverty Status (percent)
1 Other reason for reporting included: Need help after, collect insurance, improve surveillance, and “other reason.”
2 Other reason for not reporting included: kid stuff, not clear crime, no insurance, found out too late, could not recover property, could not identify the offender, offender a police officer, do not want offender to get in trouble, advised not to report, too inconvenient or “other reason.”
Reported includes 11,862 unweighted and 45,076,807 weighted cases. Not reported includes 23,820 unweighted and 97,368,210 weighted cases. Weighted data presented. “P” indicates poverty “A” indicates above poverty.
Abbreviations: BF, black female; BM, black male; HF, Hispanic female; HM, Hispanic male; WF, white female; WM, white male.
Support in the regression results was also found for H3 that disadvantaged males of color were less likely to report than other groups. Although the literature suggests that victims, particularly disadvantaged black and Hispanic males, may not report due to legal cynicism (Reference Desmond, Papachristos and KirkDesmond et al. 2016), few noted that the reason for not reporting was due “lack of police effectiveness” or “police bias.”Footnote 8 However, approximately 10 percent of black males and 7percent of Hispanic males in poverty stated they did not report due to believing the police would not find their victimization important. For black males in poverty, the most important reason not to report was because the incident was a personal/private matter (21.8 percent). For Hispanic males in poverty, the most important reason not to report was because they considered it was a minor crime (19.8 percent). It is possible that disadvantaged males did not express concerns about the police due to suspicions about the interviewing procedures and responding to sensitive questions (Reference Shedlin, Decena, Mangadu and MartinezShedlin et al. 2011; Reference VaiseyVaisey 2009), or it could be that how one feels about the police is not a strong predictor of reporting with regard to victimization (Reference Desmond, Papachristos and KirkDesmond et al. 2016).
7. Discussion
Although extensive studies have examined factors associated with victim reporting behaviors, research to date has not adequately examined differences in reporting across social groups. Although the existing body of literature finds that females are more likely to report, evidence has been mixed with regard to race/ethnicity and class. The current study sought to address this discrepancy by adopting an intersectional framework, exploring whether the effects of race/ethnicity and social class on reporting can be better understood through a more complex identity and the social structural positions of victims.
The results supported our first hypothesis (H1) and prior literature that females were more likely to report than males, controlling for other characteristics of the victim and the crime incident (Reference Baumer and LauritsenBaumer and Lauritsen 2010; Reference Hart and RennisonHart and Rennison 2003; Reference Langton, Berzofsky, Krebs and Smiley-McDonaldLangton et al. 2012). The main effects of sex remained significant even when interactions were added to the model. Supplemental analyses exploring the reasons for reporting and not reporting suggest notable differences between males and females. The most important reasons given by females as to why they reported included “to stop the incident” and “prevent more crimes,” while males reported “because it was a crime.” These findings suggest that police are an important resource for females for protection, but may not be viewed the same by males (Reference BellBell 2016).
We also found support that incorporating interactions for race/ethnicity, sex, and social class provided a more complete picture of demographic differences in reporting holding other factors such as the severity of the incident, type of crime, and victim–offender relationship constant. We presented two alternative hypotheses as to which female groups were more likely to report to the police. On the one hand, many scholars have suggested that white affluent females are “ideal victims,” who are given the most attention and public sympathy in the media, politics, criminal justice outcomes, and public opinion (Reference ChristieChristie 1986; Reference Holcomb, Williams and DemuthHolcomb et al. 2004; Reference HowertonHowerton 2006; Reference Kulig and CullenKulig and Cullen 2017; Reference MadrizMadriz 1997; Reference NeelyNeely 2015; Reference RitchieRitchie 2017; Reference Williams, Demuth and HolcombWilliams et al. 2007), which may encourage them to be more likely to report when they are victims of crime (H2a). On the other hand, we also considered that poor females of color may be more likely to report than white and more affluent females because they lack other options and access to social institutions (H2b) (Reference Avakame, Fyfe and McCoyAvakame et al. 1999; Reference KaukinenKaukinen 2004; Reference SnowdenSnowden 1998). In other words, although minority women may not be confident that the police will be able to help them, their concerns are superseded by an immediate need for help (Reference BellBell 2016; Reference SlocumSlocum 2018).
We found support for the latter of these hypotheses. Significant interactions were observed for sex and economic disadvantage (poverty status and education), race/ethnicity and sex, and race/ethnicity and education. Black females were more likely to report than white females. Black females in poverty and black females with less than a high school education had the highest predicted probabilities of reporting overall. Although reasons for reporting were similar between disadvantaged black females and affluent white females, white affluent females were more likely to state that they did not report because they contacted another official. These findings suggest it is possible that white affluent females have greater access to alternative resources than disadvantaged black females, who must rely on the police over other methods of help seeking (Reference BellBell 2016). These findings also support the need for more research that simultaneously considers multiple forms of help seeking and not solely focus on crime reporting (Reference Xie and BaumerXie and Baumer 2019).
The third hypothesis predicted that disadvantaged males of color would be the least likely group to report to the police. Following prior research, we expected the lowest reporting probability to be observed among disadvantaged black males, followed by disadvantaged Hispanic males Reference Brown and Reed Benedict(Brown and Reed Benedict 2002; Reference Hagan, Shedd and PayneHagan et al. 2005; Reference Nadal, Davidoff, Allicock, Serpe and ErazoNadal et al. 2017; Reference Sun and WuSun and Wu 2018; Reference WeitzerWeitzer 2014; Reference Weitzer and TuchWeitzer and Tuch 2006). While this pattern was evident when measuring disadvantage by poverty, the lowest reporting probability when measuring disadvantage b education was observed in Hispanic males. While these findings supported our hypotheses, when exploring the reasons for reporting, we did not find strong support that legal cynicism could explain why these groups did not report.
Police bias and police effectiveness were not considered the most important reasons when deciding not to report being a victim of crime. These findings raise questions about the relationship between legal cynicism, race, and reporting to the police—at least for victims of crime (see also Reference Xie and BaumerXie and Baumer 2019). The results did indicate that although it was not the most prominent reason for reporting, a small but notable proportion of disadvantaged male victims stated they did not report because they thought the police would not think the incident was important. Our findings do not conclusively explain why legal cynicism was not more prominent among disadvantaged male groups. Victims in this study were asked about why they reported in retrospect and we cannot state for certain if the reasons given by victims actually motivated their decision to report or not report, if victims justified or rationalized their behavior afterward, or if, in general, the interview methodology may be influencing the participants reasons (Reference VaiseyVaisey 2009). Additional research is needed to understand the reporting behavior of disadvantaged men of color and how it relates to cultural orientations such as legal cynicism.
It is unclear why Hispanic males were the least likely group to report to the police when considering the effect of education. Although the literature on Hispanics and the police is relatively (and surprisingly) small (Reference RennisonRennison 2007; Reference WeitzerWeitzer 2014), in one notable study, higher educational attainment increased the likelihood of reporting among Hispanic populations, but had no effect on reporting for other racial and ethnic groups (Reference RennisonRennison 2007). Although the reasons behind this finding were unclear, Reference RennisonRennison (2007) speculated that less education may coincide with language barriers. In addition to limited English proficiency, scholars have presented other possibilities including concerns about immigration status (undocumented and legal) and cultural beliefs that the police (as an outsider) should not be involved with personal matters or challenge the head of the household as possible reasons for nonreporting (Reference CarterCarter 1983; Reference Gutierrez and KirkGutierrez and Kirk 2017; Reference Herbst and WalkerHerbst and Walker 2001; Reference RennisonRennison 2007; Reference WeitzerWeitzer 2014). However, if these reasons were the case, they still would not explain the divergence between Hispanic males with Hispanic females (see Reference Messing, Becerra, Ward-Lasher and AndroffMessing et al. 2015 for accounts of Latinas' reasons for not reporting). Future research is needed to better explain the reporting behavior of Hispanics, and different Hispanic subgroups, and economically disadvantaged Hispanic males, in particular. Footnote 9
There are several limitations with this study. First, the data were treated as cross-sectional, combining multiple data years over a long period of time. Although a variable was included to control for interview year, it is important to better understand these results over time, particularly in light of more recent events that have sparked concern over relationships between minority communities and the police. Nonetheless, the combination of multiple years may help to reduce the effects high profile police misconduct incidents, which have been shown to influence reporting behavior (Reference Desmond, Papachristos and KirkDesmond et al. 2016). Second, it would also be important to examine victims' and their families' prior experiences with police. Evidence indicates that victims who did not report prior victimizations are less likely to report future victimizations (Reference Xie, Pogarsky, Lynch and McDowallXie et al. 2006), although this did not seem to be the case for participants in Reference BellBell's (2016) study. Other work has also found that involuntary contacts with police (e.g., police stops) may have a greater influence on reporting than voluntary contacts (Reference SlocumSlocum 2018). Third, the current study was not able to capture differences across neighborhoods, which would be important to examine cultural orientations such as legal cynicism (Reference Berg, Stewart, Intravia, Warren and SimonsBerg et al. 2016; Reference Carr, Napolitano and KeatingCarr et al. 2007; Reference Sampson and BartuschSampson and Bartusch 1998).
Despite these limitations, this study provides important insights regarding reporting to police. Negative attitudes toward the police may not preclude victims from reporting, even if participants say they would not report. There may be a difference between reporting intentions (whether one would report) and actual reporting behavior (did the individual report when were victimized) (Reference Desmond, Papachristos and KirkDesmond et al. 2016; Reference Xie and BaumerXie and Baumer 2019). It is also important to reiterate that the current study focused on the reporting behavior of victims with regard to their own victimization. The cultural frame that the police should not be called or are unreliable may or may not remain an issue for “third party” reporting (i.e., people other than the victim) or for victimless crimes (i.e., crimes of the public order). Future research should examine reporting behavior comparing victim reporting with the reporting of other neighborhood problems and from the perspectives of third parties or witnesses to crime incidents (Reference MillerMiller 2008; Reference WarnerWarner 2007).
Although findings from this study do not provide a definitive answer to patterns in reporting behavior, the intersectional perspective allowed us to more closely understand the complexities in reporting differences across social groups. This study also helped to provide new insights to the theoretical lens of intersectionality. In particular, we build on a growing literature that examines not just one disadvantaged group, but develops a comparative approach that considered advantage and disadvantage simultaneously across structures of power and inequality (e.g., Reference McCallMcCall 2005). We also show how intersectionality can help resolve the seemingly contradictory paradox that some disadvantaged victims are more likely to report to the police (Reference BellBell 2016; Reference Hagan, McCarthy, Herda and Cann ChandrasekherHagan et al. 2018; Reference Xie and LauritsenXie and Lauritsen 2012), while others are less likely to report (Reference AndersonAnderson 1999). In the context of economic disadvantage for women of color, reporting to the police likely reflects a lack of alternative support systems and resources (Reference BellBell 2016). By contrast, both advantaged and disadvantaged males, who may take a more self-help oriented approach to resolving issues, avoiding both police and alternative sources of help (Reference BlackBlack 1983). Without considering how victimization is gendered and how gender shapes how race and class affect reporting, it would be difficult to explain this paradox in victim reporting.
In addition to theoretical contributions, this study has implications for crime victim policy and victim services. The study does confirm that overall, males are less likely to report to the police than females. Although this may be a reflection of hegemonic masculinity and cultural orientations of who is considered an “ideal victim” (Reference ChristieChristie 1986; Reference MadrizMadriz 1997), it also indicates that a large group of victims is not being connected to potential services and support. Future policy provisions should focus on how law enforcement can more efficiently reach more low-income, disadvantaged males who have been victimized. This may mean developing more accessible and nontraditional avenues for reporting for non-white, disadvantaged males (Reference SeredSered 2014; Zaykowski 2019), which includes revisiting the social construction of term “victim” and identity of service provides as “victim” services (Reference SeredSered 2017). A recent report by the Vera Institute of Justice has identified several examples of such services that can be integrated into existing community organizations providing a wide range of assistance such as support navigating the criminal justice process and applying for victim compensation, services for help with transportation, food security, housing and education, violence intervention and prevention training, and counseling through a trauma informed approach (Reference SeredSered 2017). In addition, if poor black women are seeking help from the police because they lack access to other forms of help seeking, this may indicate that significant improvements should be made in terms of linking such victims to services.
Appendix A: Unweighted Sample Characteristics without Listwise Deletion
Appendix B: Comparison Between Weighted and Unweighted Results
Bolded results indicate where the variable significance differs between the weighted and unweighted models (significant vs. not significant). Significance level differences between levels (e.g., between p<.01 and p<.05) are not highlighted.
*p<.05; **p<.01; ***p<.001.
Appendix C: Most Important Reason for Reporting and not Reporting by Education (percent)
1 Other reason for reporting included: Need help after, collect insurance, improve surveillance, and “other reason.”
2 Other reason for not reporting included: kid stuff, not clear crime, no insurance, found out too late, could not recover property, could not ID offender, offender a police officer, don not want offender to get in trouble, advised not to report, too inconvenient or “other reason.”
