INTRODUCTION CONSIDERING MACRO-LEVEL THEORIES

Considering macro-level theories of crime and victimization is critical for the

reduction and forecasting of crime. In the 1990s, crime declined unexpectedly in the

United States of America. Arrests for homicide and robbery peaked in the 1980s and

early 1990s, but by 1998 were lower than they had been since the late 1960s according to

the Uniform Crime Report (UCR) which reports offenses that are known to the police or

that result in arrest (Blumstein & Wallman, 2000; Levitt, 2004; Zimring, 2007).

Aggravated assault arrests peaked in the early 1990s and then declined sharply as well

(Blumstein & Wallman, 2000). Notably, victimization surveys which do not require crime

incidents to be reported to the police also recorded a decline in rates of victimization. For

instance, “Violent Victimization and Race,” a report produced by the Bureau of Justice

Statistics (BJS) using data from the National Crime Victimization

Survey (NCVS), found that from 1993 to 1998, violent crime victimization decreased by

29% for white people and 38% for Black people (Rennison, 2001b). This decline

continued so that in “Crime Victimization, 2000,” Rennison (2001a) reported that in 2000

crime victimization was recorded at the lowest rate since the first report in 1974. Seeing

the decline in official data and victimization survey data suggests that the decline reflects

a behavioral change rather than policy or practice changes.

During the crime decline, violence against women was included amongst the

violence categories that experienced marked declines. Violence against women often

refers to intimate partner violence, rape, and sexual assault with a girl or woman victim.

According to the “Violence by Intimates report” which combined official data from the

UCR and victimization survey data from the NCVS, fatal intimate partner violence

against Black women decreased from a rate of 14 per 100,000 in 1976 to under 4 per

100,000 in 1996. For white women rates decreased from 1.3 per 100,000 in 1976 to .85

per 100,000 in 1996 (Greenfield et al., 1998). Additionally, sexual victimization peaked

in 1995 and then declined 64% over the following decade (Planty, Langton, Krebs,

Berzofsky, & Smiley-McDonald, 2013).

The crime decline of the 1990s was substantial, but moreover it was unexpected.

The unexpected nature of the crime decline highlighted insufficient crime forecasting

abilities and a dearth in robust macro-level longitudinal research and data (Rosenfeld &

Goldberger, 2008; Blumstein & Rosenfeld, 2008). In response, scholars explored various

hypotheses including longstanding assertions that trends of crime are related to the

economy or police tactics, to newer suggestions that crime may be related to the

availability of legal abortions or changes in the crack cocaine drug market (Levitt, 2004;

Blumstein & Rosenfeld, 2008). Notably, although violence against women also declined

during this period, it is rarely addressed in theoretical discussions or empirical

investigations (for exceptions see Heimer & Lauritsen, 2008; Lauritsen 2011). This is

likely because violence against women is typically left out of general violence

discussions due to its distinctive context which suggests unique indicators. Nevertheless,

investigations and theorizing are warranted, since a similarity in trends could indicate

similar underlying factors. Blumstein and Rosenfeld (2008) explained this possibility in

regards to robbery and murder when noting that “both reach their peaks and their troughs

within a year of each other. This may suggest that similar factors are affecting both

trends” (p. 14). They go on to say that there are other possible explanations including that

one of the trends may be influencing the other one. Yet, the similarity between general

violence and violence against women trends has not prompted considerable quantitative,

macro-level, and longitudinal research or theories of violence against women.

Another element that remains out of central discussions regarding the crime

decline is the fact that the details of the decline such as the onset and magnitude differ by

race (for exceptions see Heimer & Lauritsen, 2008; Parker, 2008). This is true for

violence against women as well as other forms of violence. These lapses drive the design

of the present study. For instance, this study uses an intersectional perspective, which

Black Feminist scholars introduced and have applied in many disciplines (Crenshaw,

1989; Crenshaw, 1991; Potter, 2015). Specifically, I use a framework that considers

multiple historical and current societal structures and their influence on individual

interactions, interactions with institutions, and personal responses to investigate trends of

violence against women. Through this perspective, I introduce an additional way of

operationalizing a key concept to an existing body of literature.

This study approaches violence against women from, arguably, the most

prominent macro-level explanation for violence against women, which posits that

changes in violence against women are related to changes in gender inequality. Yet, while

this posited relationship serves as the basis for organizational and global initiatives,

Whaley, Messner, and Veysey (2013) suggest that the evidence regarding the relationship

has been “inconsistent and inconclusive” (p. 733). Examples of the influence of this

explanation include the United Nations’ creation of the United Nations Entity for Gender

Equality and the Empowerment of Women (commonly referred to as UN Women) in

2010 (UN Women, n.d.). This initiative links the concepts of violence against women and

gender inequality in their purpose statement suggesting that increases in gender equality

will decrease violence against women. Additionally, the United Nations Population Fund

discussed the influence that gender inequality has on gendered discrimination including

violence on their website. This group also posits that decreasing gender inequality will

decrease discrimination. While there is substantial theoretical support for these initiatives,

the empirical support has been limited. The commonly assumed positive relationship

suggesting that decreasing gender inequality will decrease violence against women has

motivated a considerable number of empirical investigations in the last few decades, but

with mixed findings (for examples see: Baily, 1999; Martin, Vieraitis, & Britto, 2006;

Whaley, Messner, & Veysey, 2013; Yodanis, 2004). Ultimately, within the quantitative

empirical literature, support exists for multiple relationships including negative, temporal,

and curvilinear (Whaley et al., 2013).

In general, scholars have conducted cross-sectional analyses which may

contribute to the inconclusiveness of the literature. These analyses do not investigate the

temporal element of the relationship between changes in gender inequality and changes in

violence against women. The longitudinal studies that have been conducted have yielded

important findings; however, further longitudinal work is necessary to gain a

comprehensive understanding of the relationship between gender inequality and violence

against women (Bailey, 1999; Whaley, 2001; Xie, Heimer, & Lauritsen, 2012).

Also, detracting from the conclusiveness and applicability of the gender inequality

literature is the lack of a thorough investigation of the role race plays in the relationship

between gender inequality and violence against women. Notably, there have been studies

which have disaggregated gender inequality and rates of violence against women by race

(Vieraitis & Williams, 2002; Eschholz & Vieraitis, 2004; Pridemore & Freilich, 2005).

However, I will argue that these studies lacked a comprehensive theoretical framework to

lead a conclusive analysis of the role of race. Therefore, although these studies advance

the literature by making salient the racial differences in gender inequality and violence

against women, they fall short of providing a comprehensive theoretical foundation for

the relationship between race, gender inequality, and violence against women. The

present study advances the literature by including a longitudinal methodology and

intersectional framework.

Chapter two provides a broader theory and literature review. I frame this

discussion by explaining what violence against women is and why it is treated differently

and often separately from general violence. This discussion includes information on racial

disparities in violence against women. I also outline extant hypotheses regarding the

relationship between gender inequality and violence against women and their empirical

support. Lastly, I introduce an intersectional approach, as introduced by Black

Feminist scholars, for understanding violence against women (Crenshaw, 1989;

Crenshaw, 1991; Hill-Collins, 2013; Potter, 2015). This perspective will guide the

proposed study including its framework and design.

Chapter three outlines the current study, the research questions, and how it

contributes to extant literature by applying an intersectional-lens to the gender inequality

perspective. From this perspective, I outline several hypotheses. The primary contribution

of an intersectional approach is the broadening of gender inequality to assess not just

intra-racial inequality but also inter-racial inequality for Black and Hispanic women.

Chapter four provides detailed information about the data, measures, and

proposed analytical techniques to show how the research questions will be answered. This

study contributes to the literature by using an intersectional approach and by

incorporating key elements from past theoretical and empirical literature. Specifically,

this study investigates the relationship between gender inequality and violence against

women on the national- and state-levels. The national-level investigation uses homicide

data from the Supplementary Homicide Reports (SHR), non-fatal violence data from the

National Crime Survey which was revamped into the National Crime Victimization

Survey (NCS/NCVS), and demographic data from a variety of sources. The homicide and

non-fatal violence data were used to create rates of violence disaggregated by the victim’s

race/ethnicity, sex, and relationship to their offender. The demographic data was gathered

to reflect gender inequality, intersectional inequality (a measure used for Black and

Hispanic women to reflect an intersectional perspective), absolute status, and general

economic condition. Together these data were used to conduct a series of bivariate

analyses to assess whether trends of violence appear to be significantly related with the

gender inequality, intersectional inequality, absolute status, and general economic

indicators. The state-level investigation uses homicide data from the SHR. These data

were used to create rates by state and disaggregated by victim’s race, sex, and relationship

to their offender. Due to data availability, the analysis is conducted using a time series

panel regression with fixed effects at 1980, 1990, 2000, and 2010 for the 50 states and

Washington, D.C. allowing for a change analysis. The same indicators of gender

inequality, intersectional inequality, and absolute status that were used in the national-

level investigation are used in the state-level investigations.

Chapter five presents the findings from the national-level investigation while

Chapter six presents the findings from the state-level investigation. In Chapter seven, I

discuss the implications of the national- and state-level findings and the potential for

future research and data collections.

CHAPTER TWO: THEORETICAL AND LITERATURE REVIEW

The purpose of this chapter is to provide the theoretical and empirical framework

for this study. To do this, violence against women must be defined. It is important to

explain why violence against women is often analyzed separately from general violence

especially since data illustrating the crime decline of the 1990s show that trends of

violence against women declined along with general violence. While these trends are not

identical they are similar and correlated according to a report prepared by Lauritsen in

2011 for the Panel on Measuring Rape and Sexual Assault in Bureau of Justice Statistic

Surveys. While this similarity may indicate similar underlying explanations, the study of

violence against women has distinct theories. Hence second, I review the hypotheses

associated with the most popular macro-level explanation for violence against women

which posits a relationship between gender inequality and violence against women. In

addition, I outline some findings related to the various hypotheses. Lastly, I reframe the

discussion of gender inequality from an intersectionality perspective as introduced by

Black Feminists. This chapter sets the foundations for Chapter three which presents my

hypotheses.

DEFINING VIOLENCE AGAINST WOMEN

The concept of violence against women literally applies to all forms of violence

for which the victim is a woman or girl; however, rape, sexual assault, and intimate

partner violence (non-fatal and fatal) tend to be the focus (Kruttschnitt, McLaughlin, &

Petrie, 2004). Research on violence against women tends to occupy its own space in

criminological research and is studied separate from other forms of violence (Kruttschnitt

et al., 2004). Kruttschnitt et al. (2004) suggested that “this intellectual separation of

research on violence against women stems from the premise that distinctive features of

the social and political context of violence against women, particularly the context of

intimate relationships, sets it apart from other forms of violence” (p. 2). They follow this

explanation by declaring that they “urge an end to the almost total separation that has

characterized the field” (p. 2). Indeed, the social and political context of violence against

women characterizes society in the United States as a whole and thereby could have an

impact on other forms of violence. The results of this separation is that although trends in

violence against women declined with other forms of violence in the 1990s, violence

against women has been largely omitted from the great crime decline literature. The vice

versa is also true, the discussion of the great crime decline has largely been omitted from

investigations of violence against women. The separation of these literatures may be

limiting our understanding of crime trends, their explanations, and how they relate to one

another.

Advocating for more crossover between the literatures does not negate that

violence against women differs contextually from general forms of violence. These

differences exist in who the victims are, how the violence was enacted, by whom the

violence was enacted, and what response followed the violence. For example, rape is the

only violent index crime for which women and girls make up the majority of the victims

(Kruttschnitt et al., 2004). Additionally, a greater proportion of the victimization that

women and girls experience is committed by intimate partners than is true for men and

boys (Kruttschnitt et al., 2004). Moreover, women and girls are more likely to be the

victim in intimate partner violence. In fact, in 2001, women and girls were the victim in

85% of nonfatal intimate partner victimizations (Kruttschnitt et al., 2004). The different

context of violence against women and violence against men suggests that gender

matters. Gender could matter in multiple ways at multiple levels. For instance, gender

may affect crime and victimization through its influence on the societal power structure,

which defines the power that individuals and groups hold in relation to one another and

with institutions and, therefore, influences access to institutions and interpersonal

interactions (Allen, 2011; Brownmiller, 1975; Miller & Mullins, 2006).

Also, the changes in the way that rape has been defined reflect its distinctive

context. For example, the Federal Bureau of Investigation’s (FBI) official definition for

rape has evolved over time. Prior to a change in 2013, the FBI defined rape as “the carnal

knowledge of a female forcibly and against her will” (Rape, n.d.). In 2013, the definition

changed to “penetration, no matter how slight, of the vagina or anus with any body part

or object, or oral penetration by a sex organ of another person, without the consent of the

victim” (Rape, n.d.). Moreover, laws have expanded to include rape of married females,

rape by husbands, and rape against males (Brownmiller, 1975). These expansions of the

definition do not reflect a change in behavior but rather a shifting understanding of

sexuality and the relationship between men and women. Writing specifically about

changes in statutory rape, Cocca (2004) argued that the legislation of “statutory rape has

served as a site for multiple constructions of gender and sexuality” (p. 3). Similar

arguments have been presented regarding rape more broadly (Toner, 1982). Therefore,

laws about rape influence how gender and gendered relations are constructed while also

responding to the extant cultural values. The legal and cultural construction of rape (as

well as other forms of violence against women) has real consequences that further

distinguish it from other forms of violence.

One consequence of the legal and cultural construction of rape and other forms of

violence against women is that it has influenced enforcement practices and the treatment

of victims. For instance, rape and intimate partner violence have often not been a priority

for law enforcement. This has resulted in the refusal to officially handle rape and intimate

partner violence allegations. However, when these allegations have been handled

officially, victims of rape have often been “revictimized” by the criminal justice process

through suspicion, apathy, and accusations (Belknap, 2010; Russell, 1984). Arguably, the

extent of this animosity has decreased as rape has been recognized as an abhorrent crime

through policy and legislative changes (Baumer, Felson, & Messner, 2003; Belknap,

2010). Regardless of the official changes, improvements, and increases in reporting, rape

continues to be one of the most underreported crimes (Baumer et al., 2003; Britton,

2011). Continuing negative relations between police officers and intimate partner

violence victims have been noted in the literature as well (Xie et al., 2012b). As a result,

the history of rape, sexual assault, and intimate partner violence is marred by waves of

disbelief, silence, and silencing which suggest that it may be distinct from other forms of

violence.

It is also important to address that women experience violence according to other

identities such as race, immigration status, class, sexuality, etc. (Crenshaw, 1991; Potter,

2015). For instance according to the “Female Victims of Violence” report by BJS, Black

women have historically experienced higher rates of intimate partner violence than white

women (Catalano, Smith, Snyder, & Rand, 2009). This was true of rape and sexual

assault as well (Catalano et al., 2009). The heightened level of risk that Black women

experience has largely been neglected within the gender inequ

ality literature.