Intimate Partner Homicide
| Published date | 01 February 2009 |
| Date | 01 February 2009 |
| Author | Darryl W. Roberts |
| DOI | 10.1177/1043986208329771 |
| Subject Matter | Articles |
Criminal Justice
Volume 25 Number 1
February 2009 67-88
© 2009 Sage Publications
Intimate Partner Homicide
10.1177/1043986208329771
http://ccj.sagepub.com
hosted at
Relationships to Alcohol and Firearms
http://online.sagepub.com
Darryl W. Roberts
University of Maryland School of Nursing
An overwhelming proportion of intimate partner (IP) homicide perpetrators are under
the influence of substances when the crime occurs, and alcohol consumption is a strong
predictor of intimate terrorism of women. In IP homicide, female victims are twice as
likely to die from a gunshot wound as from stabbing, strangling, or other methods; and
firearm ownership is shown to increase the likelihood of IP homicide by a factor of
5.38. Compiled from publicly available data sources, the present study analyzes a
database of all lethal events occurring in the U.S. from 1985 to 2004. Using a panel of
counties and negative binomial regression, the influences of alcohol and firearms,
controlling for other variables, on IP homicide and IP homicide by firearm are
estimated. Alcohol consumption and firearm ownership increase both the incidence
rates of IP homicide and IP homicide by firearm. However, highly restrictive firearms
carry laws also increase the incidence of IP homicide. IP homicide is strongly
influenced by alcohol and firearms availability, but some types of firearms carry laws
might be counterproductive in decreasing the incidence of this crime.
Keywords:
intimate partner homicide; alcohol; firearms; panel studies
Homicide is a crime of men (Batton, 2004; Fox & Zawitz, 2006; Silverman &
Kennedy, 1987; Wolfgang, 1967; Zahn & Sagi, 1987). Men perpetrate seven of eight
homicides and are the victims in three of four homicides. The homicide victimiza-
tion rate was more than 3 times higher for men (8.7 per 100,000) than for women
(2.4 per 100,000) in 2004 (Fox & Zawitz, 2006). More than 65% of all male-perpe-
trated homicides also had a male victim. The smallest homicide category is that of
female-on-female homicides, which account for 2.4% of all homicides in the United
States (Fox & Zawitz, 2006; Glass, Koziol-McLain, Campbell, & Block, 2004).
However, more than 30% of homicides committed by women have an intimate partner
(IP) as victim (Jensen, 2001). In fact, women are the perpetrators of 34.7% of IP
homicides1 but only 11.3% of the remaining categories of homicide (Fox & Zawitz,
2006). IP homicide is the only type of lethal violence in which the principal victims
are women. Since 1976, the distribution of IP homicide has changed from a near 50-50
Author’s Note: The author could not have completed this research without the aid of Dr. Dave Marcotte,
University of Maryland Baltimore County; and Drs. Carolyn Yocom and Judy Ozbolt, University of
Maryland School of Nursing. Additionally, the author thanks Dr. James Alan Fox of Northwestern
University for providing access to his exceptional data files.
67
68
Journal of Contemporary Criminal Justice
split between the sexes to one in which three quarters of the victims are female. In
2004, 1,301 women were killed by their male partners, which accounts for one third
of all women’s homicides. In that same year, there were considerably fewer male IP
homicide victims (n = 515) (Fox & Zawitz, 2006).
The Federal Bureau of Investigation (FBI) and National Center for Health
Statistics (NCHS) currently classify IP homicide as a form of criminal homicide that
is counted along with nonintimate (NI) homicide in reporting of a jurisdiction’s
homicide incidence. This classification has a noteworthy impact on law enforcement
methods (Brannan, 2006; Buzawa & Buzawa, 2002; National Institute of Justice,
1995; Task Force on Local Criminal Justice Response to Domestic Violence, 2005;
Wisconsin Department of Justice, 2002) that goes beyond the scope of the current
article. In a recent 20-year national panel study, I found significant differences
between NI homicide and IP homicide, suggesting that a separate classification for
IP homicide might be appropriate (Roberts, 2009). The current article reports on
important findings from that study relating to the relationships between alcohol and
firearms accessibility and the county-level incidence of IP homicide. The aim for the
current article is to answer the following questions:
1. Do IP homicide incidence rates vary with the county’s quantity of alcohol purchases
(a proxy for alcohol consumption)?
2. Do firearm restrictions increase, decrease, or have no effect on IP homicide?
3. Does IP homicide incidence vary with the prevalence of firearm ownership?
Background and Significance
The criminology literature correctly refers to homicide as a rare event. According
to Fox and Zawitz (2006), the U.S. homicide rate in 2006 was 6.1 per 100,000 persons.
Thus, IP homicide, which accounts for 7% of all homicides, is very rare. However, this
level of rarity does not negate the importance of understanding as much as possible
about IP homicide to develop and apply policy alternatives focused on eliminating it
altogether. Perhaps it is due to its rarity that social scientists and criminologists should
focus additional attention on IP homicide, because as Campbell and colleagues
(Campbell, 1994, 1995; Campbell et al., 2003) and Websdale and colleagues
(Websdale, 2003; Websdale, Sheeran, & Johnson, 2004) have noted, IP homicide is the
most preventable form of lethal violence.
Additionally, the rarity of IP homicide does not diminish its impact. Consider that in
2004 alone more Americans lost their lives to violence at the hands of intimates (1,816
deaths) than lost their lives to violence at the hands of enemy forces in Iraq during the
first 30 months of the war (1,798 deaths) (see the U.S. Department of Defense Web site,
http://www.defenselink.mil). Furthermore, male IP homicide perpetrators are signifi-
cantly more likely to commit suicide after the crime (Koziol-McLain et al., 2006;
Roberts / Intimate Partner Homicide
69
Milroy, 1995; Starzomski & Nussbaum, 2000). This highly replicated finding speaks to
the significant effect IP homicide has on the perpetrator and the community.
Campbell, her mentees, and her colleagues have published several papers identi-
fying IP homicide risk factors (Campbell, 1994, 1995; Campbell & Soeken, 1999;
Campbell et al., 2003). The most well-known risk factor is a prior history of IP vio-
lence. Additional factors that increase IP homicide risk include estrangement, male
partner’s unemployment, firearm ownership, and presence of stepchildren (Bullock
& Cubert, 2002). Stalking is present in 70% to 90% of IP homicides (McFarlane,
Parker, Soeken, Silva, & Reed, 1999). Walsh and Hemenway (2005) concluded that
illicit drug and alcohol use do not significantly correlate with IP homicide but that
nearly 70% of perpetrators were under the influence of drugs or alcohol when the
crime took place.
Additionally, many researchers report a strong relationship between pregnancy
and IP homicide risk (Campbell et al., 2003; Chang, Berg, Saltzman, & Herndon,
2005; Decker, Martin, & Moracco, 2005; Frye, 2001), particularly among teens
(Krulewitch, Roberts, & Thompson, 2003). Forced sex during pregnancy also
increases IP homicide risk (Bullock & Cubert, 2002). Furthermore, mental illness
among both victims and perpetrators increases IP homicide risk (Bourget, Gagne, &
Moamai, 2000; Campbell & Soeken, 1999; Jurik & Winn, 1990; Rosenbaum, 1990).
Whereas the aforementioned change in sex distribution of IP homicide victims
from half men and half women to three women for every man presents a disturbing
trend, both men and women have benefited from a significant decrease in IP homi-
cide incidence over the past three decades (Fox & Zawitz, 2006). In 1976, there were
2,957 IP homicide deaths in the United States. By 2004, that number decreased to
1,816 deaths. The decrease in IP homicide deaths is consistent with the overall
decrease in homicide rates since they peaked in 1980 at 10.4 per 100,000 persons. A
breakdown of this decrease by sex shows that in 1980 male homicides peaked at 16.6
per 100,000, dropping to 9.6 per 100,000 in 2005;2 whereas female homicides
peaked at 4.4 per 100,000 and dropped to 2.5 per 100,000 in that year (Fox &
Zawitz, 2006).
The reason for the decrease is somewhat elusive; however, Blumstein, Rivara, and
Rosenfeld (2000) present a convincing argument that suggests that the 1980 homi-
cide spike was a function of the demographic peak caused by baby boomers who
reached the highest crime age (late teens to early 20s) in the late 1970s and early
1980s. By the late 1980s, this cohort matured sufficiently that violent crime rates
decreased. These consistent decreases in violent crime rates overall and in homicide
in particular seem to support Blumstein et al.’s position. The recent plateau in homi-
cide rates is consistent with U.S. Census Bureau (2006) reports indicating that the
growth rate in the U.S. population has also hit a plateau.
Whereas Blumstein et al.’s (2000) position provides a strong argument to explain
the decrease in homicide overall, it does not sufficiently explain the decrease in IP
homicide. In general, homicide perpetration peaks among persons who are in their
70
Journal of Contemporary Criminal Justice
20s; however, IP homicide perpetration peaks among persons who are somewhat
older, usually older than 35 years (Paulozzi, Saltzman, Thompson, & Holmgreen,
2001). Therefore, Blumstein et al.’s hypothesis suggests that IP homicide would
have peaked in about 1995, but that spike in IP homicide...
To continue reading
Request your trial