Spatial Correlates of Gun Deaths in Harris County, Texas
| Author | David Persse,Dwayne A. Wolf,Yijiong Yang,Thomas F. Reynolds,John B. Holcomb,Ned Levine,Stacy A. Drake,Charles E. Wade |
| DOI | 10.1177/1088767920924448 |
| Date | 01 February 2021 |
| Published date | 01 February 2021 |
| Subject Matter | Articles |
research-article2020
Article
Homicide Studies
2021, Vol. 25(1) 37 –60
Spatial Correlates of Gun
© 2020 SAGE Publications
Article reuse guidelines:
Deaths in Harris County,
sagepub.com/journals-permissions
https://doi.org/10.1177/1088767920924448
DOI: 10.1177/1088767920924448
journals.sagepub.com/home/hsx
Texas
Ned Levine1 , Stacy A. Drake2,
Thomas F. Reynolds3, Yijiong Yang3,
Dwayne A. Wolf4, David Persse5,
Charles E. Wade3, and John B. Holcomb6
Abstract
Firearm homicide and suicide deaths for 2014 were examined in Harris County, Texas,
for spatial variation and socio-economic correlates. Higher firearm homicide rates
were found closer to central Houston than firearm suicide rates. Local hot spots were
identified for both types but overlapped at only two locations. A Poisson–Gamma–
Exposure regression model showed the homicide rate was associated with higher
percentages of persons who are Black, aged 15 to 29 years, and living in poverty whereas
the suicide rate was associated with a higher percentage of persons of non-Hispanic
White ethnicity. Firearm policies that reduce homicides may also reduce suicides.
Keywords
firearm, homicide, suicide, spatial analysis, spatial modeling
This study examined spatial patterns in firearm homicides and suicides and their socio-
economic correlates in Harris County, Texas. Our perspective is broader than just
homicide and includes gun deaths from suicide and accidents. Firearm death is a major
public health problem in the United States. It is the third leading cause of injury deaths
1Ned Levine & Associates, Houston, TX, USA
2Texas A&M University, Houston, USA
3The University of Texas Health Science Center at Houston, USA
4Harris County Institute of Forensic Sciences, Houston, TX, USA
5Emergency Medical Services, City of Houston, TX, USA
6The University of Alabama at Birmingham, USA
Corresponding Author:
Ned Levine, Ned Levine & Associates, 8422 Bluegate Street, Houston, TX 77025-3212, USA.
Email: Ned@nedlevine.com
38
Homicide Studies 25(1)
when suicides and homicides are combined (Centers for Disease Control and
Prevention [CDC], 2017).
In the United States, firearms are the major weapon for homicide unlike in other
developed countries, a percentage that has been relatively stable (Grinshteyn &
Hemenway, 2016; Smith & Cooper, 2013). But, they are also the major mechanism by
which suicide occurs and there are far more gun suicides than homicides. In 2017,
CDC documented 23,854 firearm-related suicides compared with 14,542 firearm-
related homicides (Kochanek et al., 2019; Table 11). Nearly three quarters of all homi-
cides and about half of all suicides were caused by gun discharges (Kochanek et al.,
2019; Table 6). The remaining gun deaths, accidents, and legal intervention, represent
very small percentages of gun deaths.
Since the early 1990s, there has been little change in the overall rate of firearm
deaths. But, the mix has changed as firearm-related homicides decreased over time
whereas firearm-related suicides increased (Kochanek et al., 2019, Table 6; Kegler
et al., 2018; Fowler et al., 2015; Wintemute, 2015; Smith & Cooper, 2013).
Demographic Characteristics
Males are disproportionately more likely to commit both gun-related homicide and
end their lives with suicide (Kegler et al., 2017; Smith & Cooper, 2013). Younger
persons have higher homicide rates (Development Services Group, 2016; Smith &
Cooper, 2013) whereas older persons have higher suicide rates than other age groups
in most countries (Suicide Prevention Resource Center [SPRC], 2019; Conejero et al.,
2018). In particular, ages 15 to 29 have the highest rates of homicide in the United
States (Land et al., 1990; Cohen & Land, 1987) whereas ages 45 to 64 have the highest
rates of suicide (SPRC, 2019).
In urban areas, minorities, especially Blacks (African Americans) and persons from
low-income backgrounds, are more likely to be victims of firearm homicide (Berthelot,
2019; Xu et al., 2018; Beard et al., 2017; Development Services Group, 2016). For
suicides, however, Whites (non-Hispanic Whites) and Native Americans have the
highest rates (Kegler et al., 2017).
Associated Social and Economic Conditions
Starting with Durkheim (1897) but following directly from Henry and Short’s formu-
lation in 1954, there have been many studies about the relationship between economic
conditions and homicide and suicide. Using time series data, Henry and Short (1954)
postulated an inverse relationship between homicide and suicide rates that correlated
with the business cycle. Some studies have found support for this hypothesis (e.g.,
Varnik et al., 2003; Lester & Yang, 1997), whereas others have found contrary or
uncertain results (e.g., Dos Santos et al., 2016; Bills & Li, 2005; Pierce, 1967). One
study examined it cross-nationally over 88 countries and found a complex curvilinear
relationship between development, homicide and suicide, basically suggesting that
Levine et al.
39
total lethal violence is unrelated to either development or inequality (Unnithan et al,
1994).
The macro-economic association aside, it is clear that the relationship to economic
conditions appears to be much stronger for homicides than for suicides. An important
study of socio-economic correlates for cities, metropolitan areas, and states found four
factors associated with homicide rates: larger population size, particularly the popula-
tion aged 15 to 29 years; economic deprivation; higher proportion of divorced males;
and the southern United States having higher rates than other regions (Land et al.,
1990). Subsequent studies have tended to confirm these associations, at least for the
larger ecological units used (McCall et al., 2010). Concentrated disadvantage (poverty,
lack of jobs, and broken families) has long been shown to correlate with homicide
rates (e.g., Thompson & Gartner, 2014; Sampson, 2012; Tcherni, 2011; Stults, 2010;
Kubrin & Herting, 2003; Wilson, 1987).
However, the relationship of economic conditions to suicide is much weaker. In the
United States between 1999 and 2013, Case and Deaton, a Nobel Prize winner in
Economics, documented an increase in deaths of despair (suicide, alcohol-related
liver diseases, drug overdoses) between 1999 and 2013 among middle-aged, non-col-
lege educated, White men and women, reversing decades of decline in mortality, and
suggested that economic insecurity underlay the increase (Case & Deaton, 2015). This
hypothesis has been tested in subsequent studies using large geographical units such
as counties or states but with mixed results (Brown & Tucker-Seeley, 2018; Ruhm,
2018; Case & Deaton, 2017; Gelman & Auerbach, 2016).
But, aside from time series variation in homicides and suicides, the two phenomena
have generally been treated separately. Furthermore, uncertainty associated with using
ecological relationships to make inferences about individual behavior is well known
and makes such conclusions suspect (Freedman, 2015; Chainey & Ratcliffe, 2005; Alt
et al., 2001).
Spatial Patterning of Firearm Homicides and Suicides
Urban–Rural Differences in Rates
The spatial pattern of homicides and suicides is well established. Typically, urban areas
have higher homicide rates whereas rural areas have higher suicide rates (Ingraham, 2018;
Kegler et al., 2017; Miller et al., 2012; Branas et al., 2004). Furthermore, these urban–rural
differences in rates have increased over time (Harne, 2018; Rossen et al., 2018; JAMA,
2017; Smith & Cooper, 2013).
Nevertheless, although the suicide rate relative to population is higher in rural
areas, the majority of suicides still occur within urban areas. For example, between
2012 and 2016, there were 2,227 suicides in Harris County, the central county in the
Houston metropolitan area (Governing the States and Localities, 2016). This was
greater than the combined number of suicides that occurred in the 12 surrounding
counties (1,377). Similar patterns can be seen in most metropolitan areas, namely that
the number of suicides is almost always higher in the central counties.
40
Homicide Studies 25(1)
Concentration Within Urban Areas
There are many studies of spatial variation of homicides within urban areas. In gen-
eral, lower-income, minority-dominated, and higher density areas within cities have
been associated with higher homicide rates (Beard et al., 2017; Sampson, 2012;
Weisburd et al., 2012).
However, there are few studies of spatial variation in suicides. Several studies
showed that suicides were clustered, but the degree of concentration was small, cer-
tainly compared with homicide (Kassem et al., 2019; Ngamini Ngui et al., 2014).
Furthermore, none of the suicide clustering studies identified underlying social factors
that could account for the clustering. Considering that Durkheim (1897) long ago used
suicide to examine social conditions, the identification of clusters needs to be tied to
environmental factors. On this note, one study of suicides in Stockholm, Sweden,
found that suicide concentrations tended to occur near rail stations embedded in areas
with high drug-related crime rates and suggested that economic disadvantage might be
an underlying factor (Ceccato & Uittenbogaard, 2016).
Goals of the Study
Our study had four goals with regard to firearm homicides and suicides in Harris
County:
1. Understand how these vary within the county;
2. Identify concentrations of deaths;
3. Identify overlap...
To continue reading
Request your trial