Vous êtes sur la page 1sur 25

Gender Inequality and Female Victimization

Analytical Paper
Ismail Nooraddini
Master of Arts in Applied Sociology
Department of Sociology and Anthropology
University of Maryland, Baltimore County
07/01/2012
Professor Marina A. Adler, First Reader
Professor Christine A. Mair, Second Reader

Abstract

ABSTRACT: Ninety-five percent of female homicides are committed by males. Previous


research have found links between general economic inequality and homicide rates, but research
on gender status and male-on-female homicides have not been as consistent. Using feminist
theory, I hypothesize gender equality, as measured by occupational status and the median income
gap, will have a negative relationship with the average amount of male-on-female homicides.
Census data for 132 US cities with populations exceeding 100,000 people in 2000 are examined
to assess the degree to which womens status, relative to men, affects average female homicides.
The null hypothesis is only rejected with occupational status, but we failed to reject the null with
the median income gap. I conclude its not financial stability thats the driving force behind
female homicides, but full time employment. I end with a discussion of the results, limitations
and suggestions for future research.

INTRODUCTION
Ninety-five percent of female homicides are committed by males (Titterington 2006;
ABA 2011). That is 1,818 female homicides in 2009, with a national rate of women killed by
men of 1.25 per 100,000 (VPC 2011). Feminist scholars have argued male-on-female violence
can be explained by gender relations. Specifically, they believe social and institutional conditions
of gender inequality can establish a social climate that allows or limits male violence toward
women (Daly & Chesney-Lind 1988). Previous literature has assessed the links between gender
status and domestic violence (Yllo & Straus 1984; Straus 1994), rape (Bailey 1999; Whaley
2001), and in severe cases, homicide (Whaley & Messner 2002; Titterington 2006). The alarming
number of female deaths at the hands of men calls for attention.
Previous literature has found links between general economic inequality and homicide
rates (For reviews, see Land, McCall & Cohen, 1990), but research on gender status and female
homicides of women have not been as consistent. Feminist theory (Daly & Chesney-Lind 1988)
believes both gender inequality and equality have a positive relationship with female
victimization, with the former increasing structural disadvantage of women (Titterington 2006;
1| P a g e

Vieraitis et al. 2008), and the latter leading to a backlash affect (Gartner et al. 1990; Whaley
2001; Whaley and Messner 2003). Empirical research has attempted to assess these theories
through relative and absolute female statuses in society, which have offered different dimensions
of gender status. Despite the various approaches, research has yet to find a common ground.
The purpose of this paper is to examine the relationship between gender status and female
victimization. Census data for 132 US cities with populations exceeding 100,000 people in 2000
are examined to assess the degree to which womens status, relative to men, affects average
female homicides. This will be done in three parts. First I offer a brief description of feminist
theory which has guided the empirical research. A review of previous theory and research
suggests two competing hypotheses about the link between womens status and their
victimization; (a) ameliorative hypothesis, which states greater inequality between men and
women make women suitable targets for male violence; or (b) backlash hypothesis, which
argues equality between men and women may threaten the system of patriarchy, and men may
respond to the threat of womens advancement with violence. The second part describes the
research methods. This includes how gender status is conceptualized, where data comes from,
and the analytical procedures involved. OLS and Poisson regression are inappropriate
approaches because of the paucity of female homicides, therefore a negative binomial regression
is used. And finally I close with a presentation of results and a conclusion.
LITERATURE REVIEW
Structural correlates with violence
There has been a plethora of research on the links between structural factors and
homicide rates. Work originally began with Durkheim, Marx, and the Chicago socioecologists
2| P a g e

asking about how homicides were influenced by structural conditions. Land, McCall, and Cohen
(1990) summarized most of this work. Their analysis identified the social-structural factors
associated with homicides in the United States as the percentage of people in poverty, region, the
percentage of the population that was African American, the population size, and density. These
findings were further reiterated in Steffensmeir and Hanies (2000a, 2000b) articles, and more
recently Tcherni (2011).
Tcherni (2011) used existing theory to provide structure to the empirical data. Three
theories relevant to this paper are the following. Economic disadvantage (Shaw & McKay 1942)
purports communities who lack needed resources rely on physical violence as a means of
resolving issues. Social disorganization (Sampson and Wilson 1995) argues poverty-stricken
neighborhoods lack social control and cohesion, which ultimately leads to the state of
disorganization in communities. And finally, the Southern culture of violence, (Wolfgang &
Feracuti 1967) accounts for the higher rates of homicide in the south due to various factors
pertaining to cultural remnants of the civil war (i.e. culture of honor) and weather.
However, a separate body of research during the 1990s showed that these structural
correlates varied when homicide rates are disaggregated for specific subgroups (i.e. race) of
persons (Phillips 2002). Since 2000, research began focusing on topics such as homicide and
relationships (Mize & Shackelford 2008), income inequality (Pridemore 2008), and most
recently gender status (Vieraitis et al. 2008, 2011). Feminist theory has since then been used to
explain empirical discrepancies.
Feminist perspectives on violence against women

3| P a g e

Feminists have multiple of theories to explain female victimization. Some of these


perspectives include socialist, Marxist, and radical feminist approaches (Daly & Chesney-Lind
1988). Though feminists generally agree that gender inequality is one of the sources of female
victimization, they differ on the source of gender inequality.
Socialist and Marxist feminists maintain that gender inequality is due to the inherent
inequality structured by capitalist patriarchy (Jaggar 1983; Schwendinger & Schwendinger 1983)
and thus gender equality may have an ameliorative effect on violence (Whaley 2001). The
ameliorative hypothesis suggests as women begin to reach occupational, economic, and
educational equality, their chances of being victimized decrease. These theorists argue men have
historically made up a majority of the work force, relegating domestic work to women. This
places them in a vulnerable position, as they become dependent on men (Schlegel 1977;
Titterington 2006). Moreover, this vulnerability is exacerbated by an extension of the division of
labor to the home (Vieraitis & Williams 2002). Lower class men may struggle with their low
occupation status, and seek to maintain control of their lives by victimizing their partner. This
victimization can translate to domestic violence (Straus 1994), rape (Whaley 2001), and in some
cases, homicide (Vieraitis & Williams 2002; Titterington 2006). Once women have equal access
to employment and education, as well as equal pay, they should be able to evade harmful
situations, thus female homicide rates should decrease.
On the other hand, radical feminists argue gender equality could actually lead to higher
female homicide rates. Literature tends to refer to this as the backlash hypothesis (Russel 1975;
Williams & Holmes 1981). Seemingly paradoxical, the logic behind this argument is that power
is controlled and maintained through patriarchy. As women gain more power in educational,
occupational, economic, and political spheres, this begins to upset the hegemonic patriarchal
4| P a g e

system, and in turn men retaliate by violent force to control womens strides (Russel 1975). It
must be noted, however, violence may be an ephemeral solution to a long term transition. As
women enter the work force and politics, policies are written to solve new problems (e.g. sexual
harassment) (Chafetz 1990; Whaley 2001). Here men use violence to ensure traditional gender
roles, and thus patriarchal power structures, are maintained.
Taken together these feminist theories have formulated two competing hypotheses (1)
Ameliorative hypothesis- as women begin to achieve economic, occupational, and educational
equality, relative to men, there will be less female victimization. However, this may not be the
case but instead a (2) backlash effect may occur. As women begin to achieve higher
socioeconomic, occupation, and education statuses, men will retaliate with violent behavior. The
following section will review the empirical evidence for the discussed theory.
Empirical studies of the competing hypothesis
Looking at previous literature, several studies have examined female homicide
victimization to assess whether gender inequality is correlated with these rates. Though structural
correlates have been settled (Land et al. 1990; Steffensmeir & Hanies 2000a,b; Tcherni2011),
and feminist theory has been relatively strong (Daly & Chesney-Lind 1988), empirical research
has not been as consistent. This section takes a look at past literature concerning gender relations
and homicide rates.
Three articles have found empirical evidence either to support the backlash or ameliorate
hypothesis. Gartner, Baker, and Pampel (1990) performed a multiyear study of 18 Western
nations and found statistically significant, positive effects of gender inequality on female
homicides. Three variables, relevant to the status of women, were used; female share of labor
5| P a g e

force, occupational segregation of females, and the female share of college enrollments. The
results indicated that in nations with higher female participation in labor force and where
occupations are less segregated by gender, the proportion of female homicide victims were
higher. However, nations with higher levels of women in college had no statistically significant
relationship between homicide rates and the two other gender statuses. These results supported
the backlash hypothesis.
In another study, Whaley and Messner (2002) found partial support for the backlash
hypothesis. Using 1990 census data and five years of SHR data (1990-1994) for 193 US cities,
the authors explored the effects of gender stratification on the gendering of homicides. They
found gender equality to be related to rates of male killings of females (and males), but only in
the Southern regions of the US. Consistent with previous research (Greenberg 1997; WyattBrown 1986), this is reflective of the code of honor among men. Conversely, in regions other
than South, gender equality exhibited a significant, negative effect on rates of male killings of
females. Southern regions reflected support for the backlash hypothesis, while other regions were
found to support the ameliorate hypothesis. These findings were later reiterated in DeWees and
Parkers (2003) article.
Five articles were found to have mixed results. Stout (1992) yielded contradictory results
when she looked at socioeconomic and political gender inequalities, and their relationships with
intimate femicide. The analysis used 1980 Census data and a 3 year average of SHR data (19801982) from 160 US cities, and found that in states where women were disproportionally
unemployed, relative to men, intimate female homicide rates were higher. In addition, though
only two political variables were significant, they were both in the negative direction. Both
findings lent support to the ameliorative hypothesis. However, there was also a positive
6| P a g e

correlation between the number of women in high skilled management positions, and female
homicide rates. This finding supported the backlash hypothesis.
Baily and Peterson (1995) also found sporadic results for the ameliorative hypothesis.
Using a similar data year for 170 US cities, they examined how the status of women influenced
their overall, as well as type-specific (offender to victim relationship) homicide victimization.
Their results initially found that measures of gender inequality were not associated with rates
across US cities, with a few exceptions. Specifically, the greater the male-female gaps in college
attainment and female-male unemployment, the higher were the rates of wife killings. In
addition, the greater the male-female income gap, the more argument and acquaintance killings
there were.
A few years later Vieraitis and Williams (2002) assessed the absolute and relative status
of gender inequality and their effect on female homicide rates for both White and African
American women. They used 1990 Census data, with a 3 year average of SHR data (1989-1991)
from 156 US cities. Initially, with regards to the absolute and relative status of women, they
found evidence for the backlash hypothesis. However, when variables were racially
disaggregated, results were found contrary to theoretical expectations. For White women, the
resource deprivation index was significant in the positive direction. For African American
women, none of the gender equality variables were statistically significant. It was concluded that
White women were more vulnerable to victimization. This was later reiterated in Haynie and
Armstrongs (2006) article, where they argued the racial difference in the importance of gender
equality may represent the lower level of gender inequalities between Black men and women
compared to that of White men and women [pg. 23].

7| P a g e

In one of the most recent articles, Vieraitis et al. (2008) found weak support for the
ameliorative hypothesis, and attributed this finding to chance. Using 2000 census data and the
three year average of SHR data (1999-2001) from 206 US cities, the authors assessed the impact
of womens absolute status and relative gender inequality and their risk of homicide
victimization by relationship. They found only womens absolute status index to be statistically
significant with the number of intimate partner homicides. Though significant, the correlation
was weak. Female relative status was found statistically insignificant, thus held no relationship
with rates of female homicide.
And finally, a few articles found no significant relationship between gender inequality
and female victimization. Brewer and Smith (1995) found that gender differences in income,
education, employment, and poverty were not associated with female homicide rates when they
controlled for Land et al.s (1990) structural correlates (i.e. region, divorce rate, population size,
etc). These findings were later reiterated in Marvel and Moodys (1999), and more recently, Lee
and Stevensons (2006) article. Marvel and Moody tested 65 years of homicide victimization
data, and found once structural and demographic variables were controlled for, gender median
income were not statistically correlated with female victimization rates. Lee and Stevenson
analyzed gender specific rates in 1,678 rural counties using 1990 Census data, and found
measures of gender inequality had no association with male-on-female homicide rates.
In the most recent article Vieraitis, Britto, and Morris (2011) incorporated a time-series
analysis, and again drew mixed results. The authors examined gender equality and womens
absolute status on female homicide victimization using three decade points of census data (1980,
1990, 2000) from 165 US cities, and found mixed results depending on relationships and time
frames. Overtime, as gender equality rose, they found total, and friend, female homicide rates
8| P a g e

decreased. Moreover, as absolute female status increased, only intimate female homicides
decreased. Neither gender equality nor womens absolute status appeared to influence trends in
stranger or family homicides.
Empirical research has been inconsistent in their assessment of gender inequality and
female victimization rates. This has been because of different time frames, data sets, and
methods of operationalization. Several articles over the past few decades have either found
gender inequality to have a negative or positive (and in some cases null) relationship with female
homicide rates. Given the available literature, I rely on the ameliorative hypothesis
Hypothesis- The higher gender equality is, as measured by occupational and economic factors,
the lower female victimization rates will be, after controlling for structural factors.
METHODOLOGY
Data
The data for the analysis comes from the US Census and the Supplementary Homicide
Reports (SHR 1999-2001). Independent and control variables used for this analysis were
computed using 2000 Census data from 132 US cities with populations of 100,000 or more (US
Bureau of Census 2000). Data on homicide rates came from the SHRs, collected by the Federal
Bureau of Investigation as part of the United States Crime Reporting Program. Due to the
relative infrequency of male-on-female homicides, I averaged data across 3 years, from 1999 to
2001. The cities are located in states across the country: about 19% in the North Eastern region
(excluding Vermont and Maine), 29% in the south (excluding West Virginia), 28% in the West
(excluding Montana and Wyoming), and 24% in the Midwest (excluding North Dakota) (Census
2012). Some states were omitted due to having a population of less than 100,000 people.
9| P a g e

Dependent variable
The dependent variable is the three year average of male on female homicides, defined as
homicides in which the victim is female and the perpetrator male. Female homicide data is
abstracted from Supplementary Homicide Report (SHRs) for 1999-2001. To control for
variations in female homicide, previous literature has averaged death data for three years (Baily
and Peterson 1995, Brewer and Smith 1995, Parker and DeWees 2003, Vieraitis et al 2008). Data
usage is limited to cases with single male offenders and their female victims. If a case with a
female victim had a female or unknown offender, it was not taken into account. The years 19992001 are used to correspond to data for the independent variables derived from the 2000 national
census.
Gender equality measures
To assess womens status relative to men, two indicators are used; employment status and
income. In calculating gender inequality in employment, I divide the percent of males aged 16
and older who are employed full-time by females aged 16 and older employed full-time. The
percentage of full time work is based on the number of people working 35 hours or more per
week for 50 or more weeks out of the year. And finally, in determining a gender inequality in
income, I divide male median income by female median income.

Table1. Description of Variables


Variable labels
Independent variables
Gender equality
Occupational
10| P a g e

Variable description

Ratio of male over female working full time

Income
Dependent variable
Average Female homicide
Control variables
Social structural
Social disorganization

Economic deprivation

Regional variance

Ratio of male over female median income for full-time


employees

3 yr average of male on female homicide, 1999-2001

Logged population
% of population ages 18-24
% African-American
% below Poverty line
(South = 1) (non-South = 0)

Control Variables
Drawing on the extensive homicide literature, I control for several structural correlates (Land et
al. 1990; Steffensmeir and Hanie 2000a,b; Tcherni 2011). Relevant theoretical concepts are
economic disadvantage and social disorganization (Shaw & McKay 1942; Sampson and Wilson
1995), and regional variance in homicides (Wolfgang & Feracuti 1967). Economic disadvantage
is operationalized as the percent of the population that is African American and the percent of the
population that is below poverty threshold. An examination of the Pearsons Correlation table
(Appendix A) reveals the measures of percent of poverty and percent of population thats African
American share a pearsons correlation of .6526. This is not surprising as several articles (Phillips
2002; Tcherni 2011) have mentioned areas with a high rate of poverty are typically concentrated
with a large African American population. To account for the issue I collapsed the 2 variables
into 1 measurement labeled Economic Disadvantage. Social disorganization is the relationship
between regional crime and the neighborhoods demographic characteristics. This is
operationalized as the citys population and percent of the population between the ages 18-24.
11| P a g e

Population size is measured as the total number of people in each city, and this measure has
undergone natural logarithm transformation. And finally, I address the regional variance in
homicide rates by controlling for southern regions. As per prior literature, southern homicide
rates tend to be higher, due to a culture of honor (Greenberg 1997; Wyatt-Brown 1986).
Southern regions were dummy-coded, with non-southern cities are coded as zero.
Strategy
This study uses multivariate analysis to examine the effect of gender status on the
average of male on female homicides, when controlling for structural factors. To address the
studys statistical errors, I utilized several techniques. First, the use of Ordinary Least Square
(OLS) is not an appropriate method of analysis due to the skewed nature of the dependent
variable. Assumptions are likely to be violated, resulting in biased estimates. Instead I follow
previous literature and use a negative binomial regression (Osgood 2000). According to the
literature, because male-on-female homicides are relatively scarce, negative binomial regression
is used to analyze the theoretical models (Osgood 2000; Haynie and Armstrong 2006). Stata
package 12 was used to analyze the data.
The examinations of bivariate correlations lead to identification of potential
multicollinearity. This usually occurs when two independent or control variables are highly
correlated. A look at the Pearsons Correlation table (Appendix A) reveals two issues. First, the
independent variables relaying income and employment status shared a Pearsons Correlation of .
4824. To adjust the issue I broke the analysis into two tables- one for employment status and the
other for median income. Moreover, I applied a three step process for each table. Step one, I
examined the relationship between the independent and dependent variables, without structural

12| P a g e

controls. In step two, I added four of the five controls; percent of population between 18 and 24,
percent of the population that are African American, percent of the population that are poor, and
the region. In the final step, I control for the logged population. The third step was the result I
used to assess the hypothesis because it had all the controls. The next issue was the
aforementioned measures denoting percent of poverty and percent of population that is African
American, which was solved by collapsing the two variables into the one variable, Economic
Disadvantage. Both of these techniques addressed the multicollinearity problem.
RESULTS
Table 2 shows the means and standard distributions for independent and dependent variables, as
well as structural controls. The homicide rate for the sample 132 cities is 1.38 per 100,000,
which is .13 above the 2009 national rate of 1.25. Tables 3 and 4 present analyses examining
whether the average amount of male on female homicides can be explained by two gender
equality measures. As shown, the findings yielded mixed results. In my first analysis, shown in
Table 3, I found that gender equality, as measured by employment status, has a statistically
significant and negative relationship with the average amount of female homicides. In my second
analysis, shown in Table 4, I found that the significant effect gender equality, as measured by
median income, disappears when population size is added to the model.
Table 2. Descriptive Statistics of Study Variables (N=132)

Men: women full time employment


Men full time median income (Unlogged)
Women full time median income
(Unlogged)
Men: women median income
% of 18-24 in population
13| P a g e

Mean
1.25
33,884
.50
26,481
.50
1.28
0.12

SD
0.09
5201.
43
4146.
73
0.10
0.03

% African American
% below poverty line
Region (South = 1)
Economic disadvantage
(Combined % below poverty line and %
African American)
Logged population
Male on female homicide average
Male on female homicide rate

0.22
0.16
0.29

0.20
0.05
0.45

0.38
12.47
5.95
1.38

0.24
0.82
10.77
0.90

Table 3 shows results from the negative binomial model of the impact of gender equality,
as measured by employment status, on the average number of male on female homicides. The
results indicate that in cities with a higher female full time employment ratio, their average
amount of male on female homicides are lower. The first step showed gender employment ratio
held a significant and negative relationship with female homicides at the p<.01 level. Controls
were added in the second step, which did not affect the employment variable dramatically. The
only significant control was economic disadvantage, with a positive coefficient. Once I
controlled for the population size, the effect of the employment variable is reduced to the p<.05
level. Population and economic disadvantage were both significant with positive coefficients.
The amount of variance explained by the final model is .328. This table finds support for the
ameliorative hypothesis, which predicts that gender equality reduces male on female violence.

Table 3. Negative Binomial Regression for Impact of Gender Equality, as


Measured by
Employment Status, on Average Female Homicides When Controlling for
Structural Variants
Variables
14| P a g e

Male: female full time


employment

-8.33**
(1.21)
-

-5.88**
(1.34)
1.44**
(.44)

-1.70*
(.72)
1.33**
(.19)

Region (south = 1 or non-south


= 0)

% pop 18-24 years old

Population (logged)

0.04
(.20)
-8.44
(3.51)
-

11.93**
(1.49)
.06

9.28**
(1.68)
.08

0.04
(.09)
-3.71
(2.05)
0.91*
(.04)
-8.05**
(1.23)
.33

Economic disadvantage

Constant
Pseudo R2
(N= 132) *p<.05**p<.01

Table 4. Negative Binomial Regression for Impact of Gender Equality, as


Measured by
Median Income, on Average Female Homicides When Controlling for
Structural Variants
Variables
Male: female median income

1
-5.93**
(.91)
-

2
-5.47**
(.84)
1.89**
(.36)

3
.00
(.57)
1.54**
(.17)

Region (South = 1 or nonsouth= 0)

% pop 18-24 years old

Population (logged)

.08
(.19)
-12.94**
(3.23)
-

9.20**
(1.15)
.05

9.27**
(1.16)
.11

.011
(0.10)
-4.22*
(2.11)
.94**
(1.05)
-10.59**
(1.31)
.32

Economic disadvantage

Constant
Pseudo R2
(N= 132) *p<.05 **p<.01

15| P a g e

Table 4 shows results from the negative binomial model of the impact of gender equality,
as measured by median income, on female homicide. The results indicate that in the first and
second model the gender income ratio has a significant negative effect. However, once I control
for population size (step 3), gender differences in economic status is not a significant indicator of
male on female homicides. Initially the income was significant at the p<.01 level, with a negative
coefficient. This finding remained once we controlled for economic disadvantage, region, and
teens. Once we controlled for the logged population, the gender income gap was no longer a
factor in female homicides. One interesting finding did occur. Specifically, an increased
percentage of teens between the ages of 18 and 24 years old, was associated with fewer female.
The amount of variance explained by the model is .321. This table does not find support for any
hypothesis, and raises the point that females relative economic status my not be indicative of
female homicides.
The structural controls were consistent in predicting average amount of female
homicides. As depicted in Table 3 and 4, I found economic disadvantage and logged population
to be significant, each holding a positive relationship with the average amount of female
homicides. Though the percentage of teens between 18 and 24 were found to hold a positive
relationship with the constant in both tables, it was only significant in Table 4. Region was found
insignificant in both tables, with a positive relationship.
Taken together Tables 3 and 4 offer similarities and differences in findings. Table 3 found
cities where females worked full time year round, relative to men, had a lower amount of female
homicides. However, this was not reflected in Table 4, where I found once you controlled for
population size, female relative economic status was not correlated with female homicides. Both
tables found economic disadvantage and population size to be significant predictors of increased
16| P a g e

female homicides. That is, as cities have an increased percentage of African Americans and
poverty, as well as larger populations, the amount of female homicides is likely to increase. One
interesting finding was associated with teens. Teens, though not a statistically significant
predictor in Table 3, were significantly associated with a lower female homicide rate in Table 4.
Discussion and Conclusion
The purpose of this research was to assess the relationship between relative gender status
and female homicides. This was done by examining 2000 Census data from 132 US cities.
Previous research, stemming from feminist theory, has yielded inconsistent results. Feminists
have used socialist, liberal, Marxist, and radical theory to explain reasons behind male-on-female
homicide victimization. These theories have led to two competing theories; (1) the ameliorative
hypothesis, which purports gender equality should lead to an environment less conducive to male
on female violence, or (2) the backlash hypothesis, which argues that gender equality would
threaten the patriarchal status quo, ultimately leading to a backlash effect where womens strides
are controlled through violence. The findings of this analysis provide partial support for the
ameliorative hypothesis.
The findings are displayed in two tables. First, with regard to Table 3, employment status,
the analysis found partial support for the ameliorative hypothesis. Once structural variants were
controlled for, cities where more women were employed full time relative to men had a lower
average of female homicides. These findings were consistent with the previous work of Vieraitis
et al. (2011), Titterington (2006), Haynie and Armstrong (2006), Stout (1995), and Bailey and
Peterson (1995). Fulltime employment can often empower the victim to take charge and leave
dangerous situations. Routine activities theory (Cohen & Felson 1979) argues female work

17| P a g e

routine may increase chances of harm. This is because an increase of outside time denotes an
increase of susceptibility to harm. However, Haynie and Armstrong (2006) and Vieraitis et al.
(2011) found most female homicides to occur between friends and intimate partners, which is
within the domestic domain.
Second, with regard to the gender income gap, my analysis found no support for the
ameliorative or backlash hypothesis. In other words, a womans relative income is not a predictor
of male on female homicides when controlling all variables in the model. These findings are
consistent with the previous work of Vieraitis et al (2008), Lee and Stevenson (2006), Marvel
and Moody (1999), and Brewer and Smith (1995). However, why is the citys ratio of men to
women employed full time indicative of female homicides, but not their economic status?
Marxist and feminist theory argue its social and economic capital which facilitates greater access
to protective resources (housing, transportation, or services to flee and stay away from dangerous
situations). If working full time typically leads to increased pay, then what does full time
employment offer that income does not?
I offer one possibility, which breaks down into two explanations. Women may not go into
work for monetary gains, but for personal connections and emotional work (i.e. pink collared
jobs). To understand this, let us reassess the conceptual measurements. The employment measure
captures the proportion of women employed outside of the home. The income measure only
captures the median income of the women employed outside the home. So the driving force
behind these findings may not be financial resources, but employment.
This may be the case for two reasons. First, employment is empowering. It instills
purpose, gives rise to one's self esteem, and thus motivation. Females may use these

18| P a g e

psychological factors to avoid dangerous partners. Where financial security may provide a means
of "escaping" dangerous situations, it may not account for the activation of a dangerous
node. Second, employment can be seen as a form of cultural capital. It provides the networks
necessary (coworkers and help hotlines) to navigate otherwise violent terrain. Though income
provides economic capital, it may not address the reasons for entering, and staying in, a violent
relationship.
The controls were found to be rather consistent, with one exception. The analysis found
support for the economic disadvantage, and partial support for social disorganization on male on
female violence. Cities with a larger percentage of African Americans and poverty had more
female homicides. Larger populations also had more male on female homicides. However,
expectantly, the percentage of 18 to 24 year olds in a city decreases female homicides. Previous
literature has found teens to be the primary suspects in many homicides. This finding was
consistent with Vieraitis et al. (2011), and may suggest that among the younger population,
same-gender homicides predominate. The region was not found to play a role in male on female
homicide.
These findings do not go without several limitations. First, the sample I worked with is
too small due to data limitations. Previous articles have used between 138 and 206 cities.
Second, I did not disaggregate for relationships. If feminist theory frequently discusses intimate
violence, and a majority of research has been on domestic violence, then it would only make
sense to disaggregate murder by victim-offender relationships. However, given the vast amount
of data and limited time frame, I was unable to disaggregate for relationships. Third, SHRs had
several unknown offenders I had to ignore, omitting a large number of cases. An unknown
offender means unknown gender. Because I limited data scope to male on female homicides, the
19| P a g e

unknown gender prevented me from taking these into account. After a certain point, researchers
began imploring mathematical techniques such as allocation method, imputation algorithm
(Messner et al. 2002), which I am unfamiliar with, to deal with unknown cases. And fourth, this
design was cross-sectional. Thus, I am unable to determine the consequences of long-term term
equality.
Previous research has found inconsistent findings on gender status and female homicides.
Despite the inconsistent findings, social scientists have continued to seek out relationships
between gender status and male on female violence, instead of understanding why findings are
inconsistent. More specifically, if some aspects of female relative status are not correlated with
female violence, then what is? In addition to reconceptualizing and operationalizing gender
status, future scientists should try to also understand why some dimensions of female status may
not be indicative of female homicides.

Notes
1. Since this study utilizes Supplementary Homicide Reports (SHRs), as one of the primary
data sources, all the limitations associated with SHRs apply. Specifically, incomplete and
inconsistent reporting of SHRs is a concern.
2. No cities within the states Wyoming, Montana, Maine, Vermont, West Virginia, or North
Dakota were tested because their populations were under 100,000 people.
3. Hawaii, Florida, Kentucky, and Washington, DC were dropped from the sample due to
missing death data.
4. Cities with >100,000 population were exhausted from the Eastern region.
5. Indianapolis, IN and Nashville, TN census data had been labeled Remainder- the portion
not within any place.
6. Death data in Illinois was only available for Chicago, dropping- Naperville, Joliet,
Rockford, and Aurora.
7. Several cities had missing death data for 1999- Hunstville, AL; Topeka, KS; Chattanooga,
TN; Austin, TX.

20| P a g e

21| P a g e

REFERENCES
Bailey, William C. The Socioeconomic Status of Women and Patterns of Forcible Rape for
Major U.S. Cities. Sociological Focus 32 (1999): 43-63. Web.
Bailey, W.C., & Peterson, R.D. Gender Inequality and Violence Against Women: The Case of
Murder. In J. Hagan & R.D. Peterson (Eds.), Crime and Inequality (1995):174-205.
Standford, CA: Standford University Press.
Brewer, V. E., and M. D. Smith. "Gender Inequality and Rates of Female Homicide Victimization
Across U.S. Cities." Journal of Research in Crime and Delinquency 32.2 (1995): 175-90.
Web.
Chafetz, Janet Saltzman. Gender Equity: An Integrated Theory of Stability and Change. (1990).
Newbury Park, CA: Sage.
Daly, Kathleen, and Meda Chesney-Lind. "Feminism and Criminology." Justice Quarterly 5.4
(1988): 497-538. Web.
DeWees, Mari A., and Karen F. Parker. "Women, Region, and Types of Homicide: Are There
Regional Differences in the Structural Status of Women and Homicide Offending?"
Homicide Studies 7.4 (2003): 368-93. Web.
"Domestic Violence Statistics ." Commission on Domestic and Sexual Violence . American Bar
Association, (2011). Web. 10 Feb. 2012.
<http://www.americanbar.org/groups/domestic_violence/resources/statistics.html>
Felson, Richard B., and Paul-Philippe Pare. "Gun Cultures or Honor Cultures? Explaining
Regional and Race Differences in Weapon Carrying." Social Forces 88.3 (2010): 1357378. Web.
Gartner, Rosemary, Kathryn Baker, and Fred C. Pampel. "Gender Stratification and the Gender
Gap in Homicide Victimization." Social Problems 37.4 (1990): 593-612. Web.
Greenberg, K.S. Honor and Slavery. (1997). Princeton University Press.
Haynie, Dana L., and David P. Armstrong. "Race and Gender-Disaggregated Homicide
Offending Rates: Differences and Similarities by Victim-Offender Relations Across
Cities." Homicide Studies 10.1 (2006): 3-32. Web.
Land, Kenneth C., Patricia L. McCall, and Lawrence E. Cohen. "Structural Covariates of
Homicide Rates: Are There Any Invariances Across Time and Social Space?" American
Journal of Sociology 95.4 (1990): 922-63. Web.

22| P a g e

Lee, Matthew R., and Ginger D. Stevenson. "Gender-Specific Homicide Offending in Rural
Areas." Homicide Studies 10.1 (2006): 55-73. Web. 12 Mar. 2012.
Like-Haislip, Toya Z., and Patricia Y. Warren. "Routine Inequality: Violent Victimization at the
Intersection of Race and Ethnicity Among Females." Violence and Victims 26.1 (2011):
88-102. Web.
Marvell, Thomas B., and Carlisle E. Moody. "Female and Male Homicide Victimization Rates:
Comparing Trends And Regressors*." Criminology 37.4 (1999): 879-902. Web.
Messner, S. F., Deane, G., & Beaulieu, M. A Log-Multiplicative Association Model for
Allocating Homicides with Unknown Victim-Offender Relationships. Criminology.
(2002). 40, 457-479. Web.
Mize, Krystal D., and Todd K. Shackelford. "Intimate Partner Homicide Methods in
Heterosexual, Gay, and Lesbian Relationships." Violence and Victims 23.1 (2008): 98114. Web.
Osgood, D.W. Poisson-based Regression Analysis of Aggregate Crime Rates. Journal of
Quantitative Criminology, 16. (2000). 21-24. Web.
Phillips, Julie A. "White, Black, and Latino Homicide Rates: Why the Difference?" Social
Problems 49.3 (2002): 349-73. Web.
Pridemore, William Alex. "A Methodological Addition To The Cross-National Empirical
Literature On Social Structure And Homicide: A First Test Of The Poverty-Homicide
Thesis." Criminology 46.1 (2008): 133-54. Web.
Russell, Diana E. H. The Politics of Rape: The Victim's Perspective. New York: Stein and Day,
(1975). Print.
Steffensmeier, Darrell, and Dana Haynie. "Gender, Structural Disadvantage, And Urban Crime:
Do Macrosocial Variables Also Explain Female Offending Rates?*." Criminology 38.2
(2000a): 403-38. Web.
Steffensmeier, Darrell, and Dana L. Haynie. "The Structural Sources of Urban Female Violence
in the United States: A Macrosocial Gender-Disaggregated Analysis of Adult and Juvenile
Homicide Offending Rates." Homicide Studies 4.2 (2000b): 107-34. Web.
Stout, Karen. "Intimate Femicide: An Ecological Analysis." Journal of Sociology and Social
Welfare 19 (1992): 29-50. Web.
Tcherni, Maria. "Structural Determinants of Homicide: The Big Three." J Quant Criminol 27
(2011): 475-96. Web.

23| P a g e

Titterington, Victoria. "A Retrospective Investigation Of Gender Inequality And Female


Homicide Victimization." Sociological Spectrum 26.2 (2006): 205-36. Web.
U.S. Bureau of the Census. Social, Economic, and Housing Characteristics of the Population
(Summary File No. 3). (2000). Washington, DC: Web.
U.S. Department of Justice. Supplementary Homicide Reports. Federal Bureau of
Investigation. (1999-2001). Washington, DC: Web.
Yllo, Kersti, and Murray Straus. The Impact of Structural Inequality and Sexist Family Norms
of Rates of Wife Beating. Journal of International and Comparative Social Welfare 1
(1984): 16-29. Web.
Vieraitis, Lynne M., Sarah Britto, and Robert G. Morris. "Assessing the Impact of Changes in
Gender Equality on Female Homicide Victimization: 1980 - 2000." Crime &
Deliquency XX.X (2011): 1-26. Web. 15 Mar. 2012.
Vieraitis, Lynne M., and Marian R. Williams. "Assessing the Impact of Gender Inequality on
Female Homicide Victimization Across U.S. Cities: A Racially Disaggregated Analysis."
Violence Against Women 8.1 (2002): 35-63. Web.
Vieraitis, L. M., T. V. Kovandzic, and S. Britto. "Women's Status and Risk of Homicide
Victimization: An Analysis With Data Disaggregated by Victim-Offender
Relationship."Homicide Studies 12.2 (2008): 163-76. Web. 12 Mar. 2012.
Whaley, Rachel B., and Steven F. Messner. "Gender Equality and Gendered Homicides."
Homicide Studies 6.3 (2002): 188-210. Web.
Whaley, Rachel B. "The Paradoxical Relationship Between Gender Inequality and Rape."
Gender & Society 15.4 (2001): 531-55. Web.
"When Men Murder Women: An Analysis of 2009 Homicide Data." Violence Policy Center. The
Violence Policy Center, Sept. (2011). Web. 29 Mar. 2012.
Williams, Joyce E., and Karen A. Holmes. The Second Assault: Rape and Public Attitudes.
Westport, CT: Greenwood, (1981). Print.
Wyatt-Brown, Bertram. Honor and Violence in the Old South. (1986). Oxford University Press.

24| P a g e