Introdution These systems together
These systems together may be what places Black women at an increased risk for
violence. For instance, the men Black women are in closest proximity to due to historical
racial segregation and isolation are men who are more likely to experience, as Bailey
(1999) put it, higher levels of frustration and aggressions. Blau and Blau (1982) argued
that experiencing inequality, especially according to ascriptive characteristics, can
provoke violence and unrest. Anderson (1999) similarly suggested that lack of access to
the legal opportunity status structure has pushed the creation of an alternative status
attainment system for Black boys and men in urban neighborhoods that rely on the “code
of the streets.” The code of streets emphasizes violence and respect as integral to status
attainment. It is important to point out that most Black boys and men do not resort to
violence but this does not negate the possibility that structural disadvantage, inequality,
and oppression may be factors that promote the use of violence amongst Black boys and
men. In the end, this means that systems of racism x patriarchy may put Black women in
vulnerable positions for experiencing violent victimization.
This study posits that measuring Black women’s status in relation to the status of
white men will better portray, Black women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to inclu
This study posits that measuring Black women’s status in relation to the status of
white men will better portray, Black women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to incluThis study posits that measuring
Black women’s status in relation to the status of white men will better portray, Black
women’s status in general society. The explanations
for the relationship between Black women’s status relative to that of white men’s is likely
contingent on the victim-offender relationship. In terms of intimate partner violence and
violence perpetuated by people known to the victims, reduced inequality may elevate
Black women’s status in the home and family. If Black women’s status relative to white
men’s increases, Black women may gain access to resources and power they previously
were denied. This however, may be a hopeful proposition. Potter (2008) found that the
Black women in her study did not cite financial concerns as a primary reason for
remaining in abusive relationships. Rather, they discussed love, guilt linked to an
understanding of the racial oppression that their partners lived with, and wanting to make
relationships work. Yet thinking about the exposure and independence explanations
offered by Bailey (1999) and Dugan et al. (2003), considerable increases in status may
allow Black women to obtain positions of power within the Black women’s communities.
My hypothesis predicts this amelioration effect for violence perpetrated by
strangers as well. I do not anticipate that Black women will be more able to leave
violence-prone residential areas or more exposed to potentially violent strangers through
increases in their socioeconomic status. Systems of racism x patriarchy serve to maintain
racial segregation and isolation. Shihadeh and Flynn (1996) explained that although racial
discrimination in housing was outlawed, integration was not the result. Instead, racial
segregation has continued informally through various mechanisms. This continuing racial
segregation reduces the likelihood that Black women will move to other communities
even as Black women’s absolute status, status relative to Black men, or status relative to
white men increases. Additionally, Black women’s increased status will not increase
exposure to potentially violent offenders. Black women’s most likely offenders are Black
men and increases in status will not increase exposure to unknown Black men enough to
form a significant relationship. However, it is possible that considerable reductions in the
inequality between Black women and white men could promote Black women into more
positions of power or increase their value within Black women’s communities. Hence, I
hypothesize that the data will reveal an amelioration relationship between inequality and
violence against Black women by strangers.
In terms of absolute status, I hypothesize that increases in Black women’s absolute
status will be associated with decreases in violence by intimate partners and nonintimate
persons known to the victim. I hypothesize no relationship between increases in Black
women’s absolute status and stranger violence.
As noted in the introduction to this chapter, it is possible that it is increases in
women’s absolute status rather than or beyond relative status that influences women’s risk
for violence. It is possible that just having greater socioeconomic status can act as a
protective measure. However, it is also possible that greater socioeconomic status means
greater exposure to victimization. I predict that just increases in Black women’s absolute
status may allow access to resources that can help Black women, avoid, leave, and be less
dependent on abusive intimate partners, family members, and acquaintances. This effect
may diffuse throughout the community as women support one another, see examples of
other women leaving dangerous intimate and familial relationships, or as men recognize
the potential of being left. I predict no relationship between absolute status increases and
violence committed by strangers. I do not predict an amelioration relationship although
some scholars have suggested that increases in absolute status may indicate a greater
ability to obtain positions of power or reflect a decrease in economic instability (Martin,
et al., 2005; Xie et al., 2012a). Rather, I maintain that violence against Black women is
related to a vulnerable position in society which is best reflected through relative terms.
Moreover, I do not anticipate that increases in Black women’s socioeconomic status will
increase Black women’s exposure to potentially violent offenders. Considering Black
women’s extensive presence in the labor force with men, I do not expect that further
increases in socioeconomic status will expose enough additional Black women to produce
a significant relationship (see Figure 3.1).
Hypothesis for Research Question 4
Research Question: Are national- and state-level trends in gender inequality related
to violence against Hispanic women?
Table depicting the direction of relationship between the noted operationalization of
gender inequality and violence against Hispanic women by the victim-offender
relationship
Third, I hypothesize no relationship between violence against women of Hispanic
origin and gender inequality when operationalized as relative to men of Hispanic origin.
Like Black women, I suggest that, women of Hispanic origin are at risk for
violence according to their vulnerability and status in general society. I again argue that
considering gender inequality in reference to another oppressed group is less telling than
when gender inequality is calculated relative to white men.
I hypothesize that as gender inequality increases between women of Hispanic
origin and white men, violence against women of Hispanic origin will decrease (following
amelioration) regardless of the victim-offender relationship.
Similar to violence against Black women and violence against white women, I
predict amelioration when considering violence against women of Hispanic origin
committed by intimates and nonintimate people known to the victim. Drawing again from
the literature, decreases in gender inequality between women of Hispanic origin and
white men, may translate into Hispanic women holding a higher value in their homes and
interpersonal relationships due to the increased status of Hispanic women as a whole.
Also, like for violence against Black women, I predict an amelioration relationship for
violence committed by strangers. It is possible that as gender inequality decreases for
women of Hispanic origin, they will be more able to leave violence prone environments
(including residential areas). Residential segregation works to separate populations of
Hispanic origin from other populations as it does with non-Hispanic Black Americans.
However, South, Crowder, and Chavez (2005a) found that higher status people of
Hispanic origin lived in areas with larger white populations. In addition, this was
amplified when the person was a woman. Additionally, Alba, Logan, and (1994) found
that people of Hispanic origin lived in areas with lower crime than did non-Hispanic
Black Americans. This literature does not speak directly to the issue at hand. However, it
does provide some support for the idea that women of Hispanic origin may have greater
mobility with increasing status than non-Hispanic Black Americans.
In terms of absolute status, I hypothesize that increases in women of Hispanic
origin’s absolute status will be associated with decreases in violence against women of
Hispanic origin by intimates, nonintimate persons known to the victim, and strangers.
Similar to Black women and white women, I suggest that increases in absolute
status for women of Hispanic origin may provide greater resources to women of Hispanic
origin. These greater resources may not only allow for the women to avoid, leave, and be
less dependent on abusive intimates, family members, and acquaintances, research has
also suggested that women of Hispanic origin may be able to buy-into lower poverty and
lower crime neighborhoods (Alba et al., 1994; South et al., 2005a; South, Crowder, &
Chavez, 2005b).
Again, it is possible that changes in absolute status and relative status are related
to violence against women. It is also possible that both or either are influenced by
race/racism. These hypotheses were developed by considering some of the most
commonly mentioned underlying mechanisms for amelioration and backlash (or routine
activities for absolute status) hypotheses. Moreover, these hypotheses were created to
suggest which relationships can be expected theoretically. This is important since absolute
status and relative status can be related to one another. For example, when women’s
absolute status increases and men’s status does not increase or increases less, women’s
relative status will increase as well. Additionally, as mentioned, it is possible that
financial limitations are not the primary obstacle for Black women in abusive
relationships (Potter, 2008). However, these hypotheses provide a framework from which
to consider the existing gender inequality literature from an intersectional perspective.
Notably, the hypotheses do not differ considerably by subgroup. Indeed the hypotheses
presented for Black women and Hispanic women are the same except for the relationship
between absolute status and stranger violence. The similarities come from the proposition
that intersectional status is operational for violence against Black and Hispanic women.
Hence, null relationships are not hypothesized between relative status and violence
against Black and Hispanic women. Importantly it is possible that trends of relative status
are similar to trends of intersectional status. The differences were influenced by studies
that found that residential mobility was related differently to socioeconomic status for
Hispanic women than for Black women. The hypotheses presented for violence against
white women were influenced by the largely color-blind prior research and the Pridemore
and Freilich, 2005 study which suggested that white men may feel particularly threatened
by increases in white women’s status. Being led by prior research did not provoke
drastically different hypotheses across subgroup, rather it contributed to differing
explanations which cannot be directly studied in this investigation but warranted
consideration.
Considering the literature from an intersectional perspective allows this study to
contribute to a growing body of criminological research and potentially add clarity to
understanding the role of race. More scholarship is beginning to recognize that other
elements of women’s identity influence the relationship between changes in inequality
and violence against women (Sokoloff & Dupont, 2005). These elements include
sexuality, religion, and citizenship status, among others. This study encompasses elements
of race, class (through socioeconomic indicators though they will be aggregated), and
gender. Rooting this analysis in Black Feminist scholarship and ideology allows for a
limited, but thorough intersectional analysis. This sort of analysis advances literature on
gender inequality, violence against women, and victimization trends.
CHAPTER FOUR: DATA AND METHODS
To answer the questions posed in Chapter Three, this study investigates the
relationship between changes in women’s status and changes in violence against women.
These investigations are conducted at the national-level and at the state-level. They
expand upon previous studies by taking advantage of the annual data available through
the SHR and NCS/NCVS. After more than three decades of data collection, these datasets
can now support disaggregated longitudinal studies at the national-level. At the statelevel
it is possible to conduct decade change analyses using homicide data from the SHR.
Moreover, this study assesses multiple findings from extant literature. For example,
violence against women is disaggregated by victim-offender relationship for Black,
white, and Hispanic women. While this disaggregation allows for a deeper investigation
of the relationship between gender inequality and violence against women, it is notable
that this study attempts to parallel some of the extant literature for comparability. For
instance, this study investigates rates that are not disaggregated by race or victim-
offender relationship and operationalizes gender inequality in traditional ways. This is
particularly important since extant literature has produced mixed findings and this study
introduces multiple elements not previously studied together. The purpose of this chapter
is to outline the data and analytical techniques used in this study.
NATIONAL-LEVEL
The national-level analyses use correlation analyses to determine whether there is
a relationship between annual trends in various indicators of women’s relative and
absolute status and violence against women from 1980 to 2012.
Variables and Data
Violence against Women
In the literature, violence against women is often operationalized as one of the
following; homicide, beating, rape, all non-fatal violence, etc. using official arrest,
official reports to the police, and victimization data. At the national-level, this study
operationalizes violence against women in two ways: homicides and non-fatal violence.
Homicide data is available from the Supplementary Homicide Report, an official source
of data that is part of the Federal Bureau of Investigation’s (FBI) Uniform Crime
Reporting Program (UCR). These data are compiled by the FBI from reports submitted
by police departments. Police departments comply with the program voluntarily and most
departments submit data to the UCR regularly (Regoeczi & Banks, 2014). The SHR
collects detailed information about all homicides that are known to the police. The data
include information on up to 11 victims and 11 offenders associated with a homicide
incident (this can vary by year of collection). For each victim, information about their
race, age, sex, and ethnicity are included. In terms of race, the victim is able to be coded
as Black, white, Asian or Pacific Islander, or American Indian or Alaskan Native. Age
tended to be coded continuously although infants and those over the age of 99 were
sometimes grouped together. Sex is recorded as a binary variable, where the victim is
male or female (likely coded by anatomy). Ethnicity designates whether the victim was
identified as Hispanic or not. The same information was collected on the offenders when
available. In addition, the relationship between the victim and offender (when known),
type of weapon used, and circumstances of the offense were included when available.
General information about the offense such as the state and month are also in the data.
For this project, national-level and state-level trends were produced for fatal
violence against women. These trends were constructed to reflect homicide rates from
1980 to 2012. The annual data are publicly available through the Inter-Consortium on
Political and Social Research (ICPSR). To construct the longitudinal dataset, each year
had to be downloaded individually, recoded for consistency across years, and merged
together.
These data were chosen because they are longitudinal, can be disaggregated by
geographic unit and can be disaggregated by race x gender. However, they are not without
limitations. The limitations include the missingness in Hispanic classification, offender
information, and for Washington D.C. and Florida.
Data are available from 1975 to 2015; however, ethnicity was not included until
the 1980s. To not confound race and ethnicity, data from 1980 to 2012 are used for this
project. Importantly, although the SHR began to inclu
