You are on page 1of 20

Article

Police Quarterly
2017, Vol. 20(1) 6180
What Conditions Affect ! The Author(s) 2016
Reprints and permissions:
Police Response Time? sagepub.com/journalsPermissions.nav
DOI: 10.1177/1098611116657327

Examining Situational journals.sagepub.com/home/pqx

and Neighborhood
Factors
Jae-Seung Lee1, Jonathan Lee2,
and Larry T. Hoover3

Abstract
Police response time to calls for service is a crucial factor in evaluation of police
performance. While domestic violence is now considered serious interpersonal
violence, factors associated with response time to domestic violence incidents
are underexplored. Using hierarchical linear modeling, over 10,000 cases of calls
for service for domestic violence across 438 census tracts in Houston, Texas,
were examined. The result of multilevel analysis revealed that complainants race,
weapon involvement, and day and time of incidents were associated with response
time at the situational level. At the neighborhood level, concentrated disadvantage,
immigration concentration, and residential stability were significantly associated with
response time.

Keywords
police, response time, calls for service, domestic violence, hierarchical linear
model, social disorganization

1
Department of Political Science, Criminal Justice & Organizational Leadership, Northern Kentucky
University, Highland Heights, KY, USA
2
Penn State Harrisburg, School of Public Affairs, Middletown, PA, USA
3
College of Criminal Justice, Sam Houston State University, Huntsville, TX, USA
Corresponding Author:
Jonathan Lee, Penn State Harrisburg, School of Public Affairs, 777 West Harrisburg Pike, Middletown,
PA 17057, USA.
Email: JLee@psu.edu
62 Police Quarterly 20(1)

Introduction
While domestic violence is now considered serious interpersonal violence, it had
not been treated as such until the 1970s in the United States (Erez, 1986; Fagan,
1996; Johnson, 2007). The social movement for womens rights during the late
1960s and 1970s reframed the societal perception of violence against women.
Subsequently, the criminal justice system has enforced Draconian measures to
promote specic and general deterrence of domestic violence (Melton, 1999;
Lee, Zhang, & Hoover, 2013b). An example would be the nation-wide adoption
of a mandatory arrest policy by law enforcement agencies during the 1980s
(Johnson, 2007; Sherman, & Berk, 1984). Additionally, domestic violence has
received a great deal of academic attention, leading researchers to examine not
only its correlates but also how the justice system responds to it.
The importance of police response time has been highlighted in the literature
(Eck, & Rosenbaum, 1994; McEwen, Connors, & Cohen, 1986). Studies found
that police rapid response increased citizens satisfaction with police (McEwen
et al., 1986) and arrest rates (Cihan, Zhang, & Hoover, 2012; Clawson & Chang,
1977; Isaacs, 1967; Kansas City (MO) Police Department & United States of
America, 1978). Sherman (1992) pointed out that practitioners recognize
response time as an indispensable factor to evaluate police performance.
Hence, police response time to calls for service (CFS) has been one of the crucial
criteria to evaluate police work (Kelling & Moore, 1988; Sherman, 1992).
However, police response time in domestic violence has been underexplored.
Findings in the previous studies suggest that there exists a relationship
between police response time and ecological characteristics (Mladenka & Hill,
1978; Stevens, Webster, & Stipak, 1980). For example, Klinger (1997) posited
that police would become more cynical and their vigor to respond to citizens
demands would be diminished after responding to repeated CFS from a high
crime neighborhood. Accordingly, the current study attempts to examine the
situational factors aecting police response time to domestic violence incidents.
Examining police response time from an ecological perspective provides insight
into the association of police strategy for patrol operations and allocation of
patrol resources.

Literature Review
Theoretical and Empirical Bases
Police behavior has been examined from the perspective of several theories.
Rational choice theory argues that police behavior is driven primarily by legal
factors in order to secure legitimacy in their presence as well as enforcement
(Lee, Zhang, & Hoover, 2013a). Literature has identied common legal factors
including property damage, victim injuries, and weapon involvement (Alpert,
Dunham, & MacDonald, 2004; Avakame & Fyfe, 2001; Belknap, 1995;
Lee et al. 63

Feder, 1996; Finn & Stalans, 1995; Friday, Metzgar, & Walters, 1991; Garner,
Buchanan, Schade, & Hepburn, 1996; Hall, 2005; Kaminski, DiGiovanni, &
Downs, 2003; Kane, 2000; Terrill, Paoline, & Manning, 2003; Worden, 1995).
On the other hand, extra-legal factors, such as individual social class, gender,
or race, may be predictive of police behavior (Avakame, Fyfe, & McCoy, 1999;
Lee et al., 2013a). Also, police decision patterns show a signicant variation
between neighborhoods based on demographic composition and economic indi-
cators (Lee, 2014). It could be anything from ocers developing callousness to
repeated CFS to a systematic proling against specic group of individuals or
neighborhood, as conict theorists argue (Black, 1980; Klinger, 1997; Lawton,
2007; Sampson, 1986). In this regard, social disorganization theory appears to tie
in closely with police behavior (Kubrin & Weitzer, 2003; Sampson, 2011). First,
it argues that negative characteristics of a neighborhood such as poverty,
unemployment, and residential instability, weakens the communitys ability to
provide informal social control, thereby leading to more crime (Shaw & Mckay,
1969). The ndings in empirical studies support that neighborhood characteris-
tics were signicantly associated with crime and delinquency (Hipp & Boessen,
2013; Pals & Kaplan, 2013; Sampson & Lauritsen, 1994; Swatt, Varano, Uchida,
& Solomon, 2013). Accordingly, high volume of CFS may originate from rela-
tively marginalized neighborhoods and render ocers to question the eective-
ness of prompt response to calls. Second, it provides dierent dimensions of
neighborhood prole that could be the basis for intentional and systematic use
of discretion by police ocers. For example, ocers may be more responsive to
calls from neighborhoods with engaged and auent residents than ones from
neighborhoods with low informal social control among residents, due to a pol-
itical bias against marginalized group. Therefore, it is necessary to examine the
relationship between police response time and neighborhood characteristics.

Police Response Time: Its Importance in Policing


Kelling and Moore (1988) pointed out that police response time plays an import-
ant role in evaluating the eectiveness of police work. A growing body of
research has stressed the importance of police response time. McEwen et al.
(1986) found in three dierent cities that rapid response to citizen CFS increased
the level of citizen satisfaction with the police. The result was consistent with
Percys (1980) nding that rapid police response reduced citizens chance of
being victimized and consequently increased their satisfaction with police ser-
vices. A later study by Brandl and Horvath (1991) examined 436 crime victims
evaluations of the police and conrmed the result.
As well as the eect of response time on public satisfaction with the police,
response time is associated with arrest probability. Clawson and Chang (1977)
found that rapid response was related to high rates of on-scene arrests.
It was consistent with the result of the earlier study conducted by Isaacs (1967).
64 Police Quarterly 20(1)

Even though other studies reported no relationship between police response time and
arrest rates (Kansas City, 1978; Pate, Ferrara, Bowers, & Lorence, 1976; Spelman &
Brown, 1984), recent studies found the same result as Isaacs (1967) and Clawson and
Chang. Using the data obtained from a U.K. police agency in 1996, Blake and
Coupe (2001) examined 407 burglary incidents. They concluded that rapid police
response increased the arrest of burglars at the scene. The later study of Coupe and
Blake (2005) reported the same result. Using 406 burglary cases which only included
arrest incidents from the United Kingdom, the study found that response time was
signicantly and positively associated with arrest rates after controlling other situ-
ational factors, such as time of the incidents, workload per patrol unit, stage of
burglary, and origin of service calls. A recent study by Cihan et al. (2012) found
the same result from the analysis of in-progress burglary CFS data.

Factors Associating With Response Time


Along with the recognition of the importance of response time, previous
studies also attempted to examine the factors aecting police response time.
For example, the study of Mladenka and Hill (1978) suggested that there was
a relationship between police response time to CFS and other factors such as
socioeconomic status, race, and day of the incidents. Especially, the study found
that police response time was shorter in neighborhoods with low income and a
high level of minority population. The result of the study conducted by Steven
et al. (1980) was consistent. The analysis of 21,957 CFS across 16 census tracts in
York, Pennsylvania, revealed that police response time varied by census tracts
when the eects of other factors such as a situational factor and type of call were
taken into account. In the result, the situational factor time of the incidents
and type of call appeared to be signicantly related with police response time.
Blake and Coupe (2001) reported ndings consistent with Steven et al. (1980);
response time was shorter during the night time than day time.
More recently, Cihan et al. (2012) examined the relationship between police
response time and neighborhood characteristics using in-progress burglary CFS
data in 2007, derived from the Houston Police Department and 2000 census
statistics, including 420 tracts in which 4,917 in-progress burglaries occurred.
In the study, police response time was measured by the time interval between the
time a police unit received a call from dispatch and the time of arrival of the unit
at the scene. The result of the multilevel analysis indicated that police response
time was shorter in disadvantaged neighborhoods and that arrest rates were
increased with shorter response time. This nding was consistent with a recent
study conducted by Cihan (2014) examining the relationship between neighbor-
hood characteristics and response time to in-progress assault calls.
Comparison of two cities produced the consistent result with previous litera-
ture regarding response time. Cihan (2014) examined police response time using
the 2006 CFS data collected from the Houston and Dallas police departments
Lee et al. 65

and compared the pattern of response time between the cities. The analysis
revealed that response time varied commonly by neighborhood characteris-
tics in the two cities. Social disorganization indicators, such as concentrated
disadvantage, immigrant concentration, and residential stability, were signi-
cantly associated with police response time. However, whereas the pattern of
faster response in neighborhoods with high immigrant concentration and less
residential stability was consistent in both cities, the pattern of response time in
concentrated disadvantaged neighborhoods was inconsistent. In Houston, con-
centrated disadvantage shortened police response time to CFS while response
time was longer in Dallas. The result of this study indicates that ecological
contexts are very inuential on police response time.

Police Response Time and Domestic Violence


The former executive director of Alameda County Family Justice Center
in California, Nadia Lockyer, stated that faster response time to domestic
violence can mean the dierence between life or death (Kuruvila, 2010).
Police response time could be crucial to prevent possible injuries or deaths of
victims in domestic violence cases. Previously, studies found that rapid response
to CFS was associated with positive outcomes including an increase of appre-
hension rate (Blake & Coupe, 2001; Clawson & Chang, 1977; Coupe & Blake,
2005; Isaacs, 1967) and of citizen satisfaction with police (Brandl & Horvath,
1991; McEwen et al., 1986; Percy, 1980). Studies have consistently stressed that
police response time is very inuential in policing.
Furthermore, the eects of ecological contexts on police response were also
found in domestic violence literature. For example, Wright and Benson (2011)
conducted multilevel analysis in order to examine the eect of neighborhood
characteristics on domestic violence. Using the data from the Project on Human
Development in Chicago Neighborhoods (PHDCN), their multilevel analysis
indicated that neighborhood factors were associated with domestic violence
after controlling individual level factors. Especially, the study found that domes-
tic violence was more likely to occur in disadvantaged neighborhoods.
This result is consistent with the study conducted by Lee et al. (2013a). They
examined the relationship between extra-legal factors such as situational and
neighborhood characteristics and arrest rates in domestic violence cases. Using
over 9,000 domestic violence cases obtained from the Houston Police
Department (HPD) in 2005 and the 421 census tracts derived from the 2000
U.S. Census, the multilevel analysis revealed that social disorganization indica-
tors, including concentrated disadvantage and immigration concentration, were
positively associated with the probability of arrest as well as situational factors
such as time of incidents, type of oense, and gender and race. Their ndings
suggest that police arrest decisions in domestic violence cases were aected by
neighborhood characteristics.
66 Police Quarterly 20(1)

The Current Study


The review of previous literature indicated that police response time had a sig-
nicant impact on positive outcomes of police work (Blake & Coupe, 2001;
Brandl & Horvath, 1991; Cihan et al., 2012; Clawson & Chang, 1977; Coupe
& Blake, 2005; Isaacs, 1967; Kelling & Moore, 1988; McEwen et al., 1986; Percy,
1980). Specically, literature pointed out that neighborhood characteristics were
signicantly associated with domestic violence (Lee et al., 2013a; Wright &
Benson, 2011) and that police resources such as patrol deployment may be
distributed disproportionately across neighborhoods (Kubrin & Weitzer, 2003;
Triplett, Gainey, & Sun, 2003). These connections suggest a relationship between
neighborhood characteristics and police response time to domestic violence CFS.
However, no study has examined the relationship between response time and
neighborhood characteristics in domestic violence cases. Examining the eect of
neighborhood characteristics is important in reviewing current police practices
regarding domestic violence. Therefore, this study examines what factors aect
police response time to CFS in domestic violence cases. Two research hypotheses
are developed based on theories and previous literature:

1. Complainants demographic information and situational factors are signi-


cantly associated with police response time to CFS in domestic violence cases.
2. Neighborhood characteristics (concentrated disadvantage, immigration con-
centration, racial heterogeneity, and residential mobility) derived from social
disorganization perspective and neighborhood call rates are signicantly asso-
ciated with police response time to CFS in domestic violence cases.

Methods
Houston Police Department Dispatch Protocol
Houston is the nations fourth largest city, policed by 5,000 sworn ocers
deployed in 15 substations (precincts). A centralized communications center
receives all 911 calls, including police, re, and EMS (operated by the Fire
Department). Call-takers assign call types based on a very extensive categoriza-
tion. Call types are assigned a priority code, ranging from an E, emergency with
lights/siren response, to a 9, delayed teleserve call-back. Domestic incidents are
divided into only two call types: Disturbance/Family/Weapon and Disturbance/
Family. Disturbance/Family/Weapon is a Priority 2 dispatch with a 5-minute
response time goal. The less serious Disturbance/Family has a Priority 3 rating
with an 18-minute response time goal. If there is an indication of violence,
whether a weapon is reported or not, the call goes out at least as a Priority 2.
For Priority 2 dispatches, patrol units do not respond lights and siren but do
respond expediently. Although not sanctioned, a frequent practice when
responding to a Priority 2 event is to proceed with caution through a sparsely
Lee et al. 67

busy red light, using the vehicles emergency lights but not the siren. The practice
is unusual for a Priority 3 call. For this analysis, response times to the Priority 2
Disturbance/Family/Weapon calls are compared with the Priority 3 Disturbance/
Family dispatches. Quite obviously, assessments by call-takers are not always
correct. However, for purposes of this research, it doesnt matterthe analysis
is based on the call priority code conveyed to the responding patrol unit.
In addition, HPD does not have a specialist unit exclusively for domestic
violence. Related specialist units become involved only on a referral basis, for
example, victim services. Approximately a third of patrol ocers have received
training in handling behaviors associated with mental illness and are designated
as a crisis intervention unit. If a mental health nexus is observed by the initial
response patrol unit, and neither ocer is crisis certied, by policy, an ocer
with crisis intervention certication is deployed. Regardless, if an assault is evi-
dent, per Texas statute an arrest is made.

Data
HPD responded to over 10,000 domestic violence incidents from September 2010
through August 2013. The current study examined cases from the entire calendar
year from January to December in order to maximize the richness and variation
in the variables across incidents and across neighborhoods. The preliminary
screening excluded the cases with other racial groups than White, Black, and
Hispanic Latino, due to their relative scarcity. After screening out cases with
missing variables, the nal sample counted 10,880 incidents.
The current study used census tract as a neighborhood unit to warrant sub-
stantial variation across units.1 The incidents in the sample were from across 438
census tracts in Houston. Physical address for each incident, which was included
in the database provided by HPD, was geocoded to identify the corresponding
census tract. Characteristics at the neighborhood level were extrapolated from
information available through the 2010 U.S. Census.

Measurements
Dependent variable. Response time was identied as the dierence between initial
call-in time and the responding ocers time of arrival at the scene. It was
converted into seconds, ranging from 0 to 2844 (c.f., Table 1). For the sake of
minimal skewness and homogenous error variance, response time was trans-
formed using natural-logarithm.

Independent variables

Level 1 Demographic and situational factors. Weapon use increases the level
of seriousness of the incidents (Smith, 1986). Thus, the presence of weapon
68 Police Quarterly 20(1)

Table 1. Descriptive Statistics for Situational Variables


(N1 10,595).

Variables Min. Max. Mean/% SD

Response time 0 5.65 1.82 0.72


(Natural-Logged)
Weapon 0 1 0.19% 0.40
Weekend 0 1 0.39% 0.49
Nighttime 0 1 0.47 0.50
Rush hours 0 1 0.19% 0.39
Complainant/Caller
Age 1 99 35.66 14.18
Male 0 1 0.29 0.45
Black 0 1 0.47 0.50
Hispanic 0 1 0.40 0.49

during an altercation would instantly escalate the sense of urgency for the
responding ocer in domestic violence cases (Lee et al., 2013a). Accordingly,
it may result in faster response. Incidents were categorized into two groups: no
information concerning weapon (reference category) or presence of weapon
(coded as 1) was reported at the time of initial call. Table 1 shows that about
one in ve domestic altercation incidents involved a weapon used against com-
plainants by suspects.
Time of the day and the day of week for each incident were included as
explanatory variables based on the literature that suggested signicant variation
in the amount of CFS, which is likely to inuence police discretion in the way
they respond to the calls. Incident time was dichotomized into whether
the domestic violence was reported to the police during the daytime, between
6:00 a.m. and 10:00 p.m. or during the nighttime between 10:00 p.m. and
6:00 a.m. Next, incidents that occurred between 6:00 a.m. Monday through
5:59 p.m. Friday were grouped into weekdays, and 6:00 p.m. Friday through
5:59 a.m. Monday were grouped into weekends. Weekend was operationalized
as the window of time beginning at 6 p.m. on Friday because previous literature
included Friday nights in the weekend. Friday nights are associated with alcohol
consumption and homicide (Pridemore, 2004) and alcohol and domestic violence
(Roman & Reid, 2012). In addition, the current study acknowledged the need for
considering the impact of trac congestion at certain times of the day2 since
trac status appeared to be a limitation that may impact police response time in
response time literature (see Cihan, 2014; Cihan et al., 2012). An indicator was
created to indicate rush hour, which is designated as between 7 a.m. and 9 a.m.
and 4 p.m. and 6 p.m. on weekdays. As shown in Table 1, slightly more cases
Lee et al. 69

were reported in the daytime than in night time, and a majority was reported on
weekdays. Less than one in ve cases were reported during the rush hour.
The data set also contained demographic information for both suspects and
complainants. It was decided to exclude suspects information. First, domestic
violence is generally intra-racial and between two opposite genders such that
complainants information would be a reliable estimate of the suspects infor-
mation (Lee et al., 2013a). Second, in the initial call from the dispatcher, ocers
are usually given the complainants demographic information but not necessarily
the suspects information. In the current research on response time, it was
deemed that the value of suspects information does not oset the loss of data
from excluding cases that are missing suspects information.3 Accordingly, age,
gender, and race of the complainant were included in the current study. Age was
a raw variable with an average of 36, and there were more female than male
complainants. Race was classied with two dummy variables for Black (47%)
and Hispanic Latino (41%) complainants, respectively (c.f. Table 1).

Level 2 Neighborhood factors. Partially adopting social disorganization


theory, we identied multiple indicators of neighborhood characteristics for
each census tract. They included the proportion of population to the total popu-
lation that was as follows: (a) below the poverty level, (b) receiving public assist-
ance, (c) younger than age 18, (d) Black, (e) Latino, (f) foreign born, (g) age 16
or older and unemployed, and (h) living at the same residence for over 5 years.
In addition, the proportion of households to the total households that was (i)
female-headed and (j) owner-occupied were included. As the result of a principle
axis factor analysis, three underlying factors with eigenvalues greater than 1 were
identied.
The rst factor was accounted for by indicators for poverty, density of juven-
iles, Black residents, unemployment and female-headed families, with most load-
ings greater than 0.76 (c.f., Table 2). The second factor was shared by two
indicators with high factor loadings (>0.89): Latino residents and foreign
born population. The third factor also had high factor loadings (>0.91) for
continuous residing and home ownership. The three factors approximated the
three dimensions of neighborhood characteristics dened by Sampson et al.
(1997); concentrated disadvantage, immigration concentration, and residential
instability, respectively.4
Additionally, Houston is a racially diverse city with an average of over 40%
of non-White residents per census tract. Racial heterogeneity as another dimen-
sion of social disorganization was calculated using Gini-Simpson index with
percentages of White, Black, and Hispanic population for each census tract,
where higher value indicates larger racial diversity.5 As shown in Table 2, at
least one census tract recorded 0, suggesting that only one race dominates the
entire census tract, while another census tract recorded 0.77, showing the largest
diversity level in Houston.
70 Police Quarterly 20(1)

Table 2. Descriptive Statistics for Neighborhood Characteristics (N2 412).

Factor
Variables loading Min. Max. Mean SD

Concentrated disadvantage 1.69 4.93 0.31 0.96


Below poverty level 0.882
Receiving public assistance 0.889
Less than age 18 0.557
Black 0.765
Unemployed 0.852
Female-headed households 0.793
Immigrant concentration 1.90 2.38 0.02 1.04
Latino 0.949
Foreign-born 0.898
Residential instability 1.04 5.47 3.07 1.01
Same house for 5 years 0.927
Owner-occupied households 0.919
Racial heterogeneity 0 0.77 0.47 1.97
Call rate (Natural-Logged) 0.19 6.14 4.15 0.63

Last, based on previous ndings that most demands for service calls are
generated from highly disorganized neighborhoods (Braga et al., 2010; Cihan
et al., 2012; Klinger, 1997; Sherman et al., 1989; Weisburd et al., 2004), it is
hypothesized that relatively frequent and repeating call for service in disorga-
nized neighborhoods would generate callousness among ocers and generate
resultant delay in response to calls. The call rate was calculated as the number of
cases per 10,000 people for each census tract and ranged from 0.12 to 46.30
across the census tracts (c.f. Table 2).

Analytical Strategy
The nature of the nested structure of the data set (incidents are nested under
census tracts) warranted the use of hierarchical linear modeling (HLM), which
improves the regression by performing an analysis with multilevel data and taking
into account the random eect (Raudenbush & Bryk, 2002).6 Each incident rep-
resented the Level-1 unit of analysis, whereas each census tract represented the
Level-2 unit of analysis. The involvement of weapon, time, and complainants
demographic information in each case were treated as Level-1 variables. Four
dimensions of neighborhood characteristics and 911 call rate for each census
tract were treated as Level-2 variables. The two modelsbase and fullwere
designed to determine how these variables inuence the response time.
Lee et al. 71

Using the one-way ANOVA statistical method, Base Model tested whether
census tracts varied in the average response time. No variable was included, and
Base Model was expressed as:

Level 1 : Yij(Response Time) = 0j+eij


Level 2 : 0j  00 + 0j

where b0j is the mean response time in the census tract j, eij is the random eect
pertaining to the individual i in census tract j, g00 is the grand-mean of the
response time across all census tracts, and 0j is the random error associated
with the census tract j.
Next, all Level-1 variables were included in the Full Model. All but age gap
were uncentered for meaningful interpretation of the coecients and the inter-
cept.7 Also, the Level-1 intercept (b0) was regressed against Level-2 variables,
which are the four dimensions of neighborhood characteristics and 911 call rate.
Full Model was expressed as:

Level-1:
Yij b0j + b1j (Weapon) + b2j (Weekend) + b3j (Night time) + b4j (Rush Hour)
+ b5j (Age) + b6j (Male) + b7j (Black) + b8j (Hispanic) + eij
Level-2:
b0j g00 + g01 (Concentrated disadvantage) + g02 (Immigration concentration)
+ g03 (Residential stability) + g04 (Racial heterogeneity) + g05 (911 call rate)
+ u0j
b1 g10
...
B8 g80

Findings
The Base Model conrmed the existence of statistically signicant variation in
response time across census tracts, warranting the Full Model. Table 3 presents
estimates of regression coecients from the Full Model analysis. As expected,
the presence of weapon at the scene was associated with a lower response time
(b1 0.12). Concerning the temporal condition, calls made on weekends
(b2 0.05) and at night time (b3 0.09) had signicantly less response
time. Rush hour did not appear to be associated with response time. For demo-
graphic attributes, interestingly, a police ocers response time was only asso-
ciated with a complainants race. More specically, Hispanic complainants
(b4 0.05) experienced signicantly faster police response time. Other com-
plainants demographic information including age, gender, or race other than
Hispanic was not associated with police response time.
72 Police Quarterly 20(1)

Table 3. Multilevel Model for Response Time.

Full model

Fixed effects b SE t value

Level 1(situational)
Intercept 2.00 .09 23.53**
Weapon .12 .02 6.80**
Weekends .05 .02 3.23**
Nighttime .10 .02 6.19**
Rush hours .02 .02 1.06
Complainant/Caller
Age .00 .00 .41
Male .01 .02 .90
Black .03 .02 1.34
Hispanic .05 .02 1.98*
Level 2 (neighborhood)
Concentrated disadvantage .06 .01 4.38**
Immigration concentration .03 .01 2.88**
Residential mobility .03 .01 2.95**
Racial heterogeneity .05 .07 .71
Call rate .01 .02 .47
Random effects Variance df 2
ei 0.70 10581
u0 0.14 406 724.972
*p < .05. **p < .01.

Most of the social disorganization characteristics for census tracts appeared


to have a signicant impact on response time (c.f., Table 3). Police responded
quicker to a case occurring in a census tract of more disadvantage, more immi-
grants, and greater residential instability. Call rate and racial heterogeneity in
census tract were not signicantly associated with the average response time in
the corresponding census tract.

Discussion
This study provides valuable information to understand police behavior on
domestic violence incidents. Basically, ndings are generally supportive of
both hypotheses; complainants demographic information, situational factors,
and neighborhood characteristics seem associated with police response time.
However, caution is required. First, this study used only domestic dispute
Lee et al. 73

CFS data from one jurisdiction, Houston, Texas. Second, some of the results
were inconsistent with other research examining response time on other crime
types (Blake & Coupe, 2001; Cihan et al., 2012; Clawson & Chang, 1977; Coupe
& Blake, 2005; Isaacs, 1967). Police response time could possibly be dieren-
tiated by type of call (Stevens et al., 1980). Third, this study could not control
deployment concentration of police units and the geographic concentration of
incidents. These factors may inuence response time (Cihan, 2014).
Even though this study includes limitations, the results of this study deserve to
receive an attention. The results reveal that police response time was signicantly
inuenced by the race of complaints and other situational and neighborhood
conditions. Surprisingly and inconsistent with prevalent assumption that mino-
rities have been discriminated against by police, the Hispanic minority received
faster response from the police. Consistent with the ndings of previous research
regarding police response time on in-progress burglary (Cihan et al., 2012)
and on in-progress assault, police response time to CFS for domestic vio-
lence incidents was shorter during the night time and on weekends. The result
also reinforces previous ndings that domestic violence incidents are more
likely to occur during the night time (Cohn, 1993; Olson et al., 1996) and
on weekends (Roman & Reid, 2012). Response time was quicker when
a weapon was reportedly involved in the incident. The result is consistent with
previous ndings that weapon involvement in domestic violence cases is asso-
ciated with police behavior (Lee et al., 2013b). The presence of a weapon escal-
ates the seriousness of any incident (Smith, 1986). In Houston, the level of
priority of call is increased to a 2 from a 3 with weapon involvement as explained
in the Methods section.
Moreover, this study found that social disorganization indicators
(concentrated disadvantage, immigration concentration, and residential mobil-
ity) have a negative relationship with police response time. In other words,
police response to CFS for domestic violence cases were faster in socially and
economically stressed neighborhoods. This result is consistent with the previ-
ous studies on police response time (Cihan, 2014). However, the result is
inconsistent with Klinger (1997)s ecological theory of police behavior. Based
on Klingers theory, police response time should be longer in socially and eco-
nomically stressed neighborhoods, since police would respond less vigorously
to demands from those neighborhoods due to the prevalent crime, which gen-
erate police ocers cynical view toward the residents and victims in the neigh-
borhoods. Indeed, the result of this study revealed that people residing
in disadvantaged neighborhoods experienced shorter police response to
their requests.
When taking into account that police do their work tactically and strategically
in order to operate patrols eectively, the results indicate that the police
are endeavoring to respond eectively for citizens CFS. Numerous police
departments have adopted new policing strategies such as concentrated patrol
74 Police Quarterly 20(1)

at hot-spots (Braga, Papachristos, & Hureau, 2014) or Compstat (Eterno &


Silverman, 2006). Police agencies deploy more patrol units during certain
times and days and in areas which generate high demands, in order to respond
eciently and quickly to high demands and to distribute fair workloads to o-
cers (Walker, 1992). Our ndings show that HPD successfully reected those
innovative strategies on their response system. In addition, our study found the
interesting result that the Hispanic minority experienced shorter response time as
opposed to an assumption that minorities receive less vigorous police response
(Black, 1976; Klinger, 1997). During the recent three decades, Houston has
experienced a drastic growth in its Hispanic population while the White popu-
lation decreased (The City of Houston, 2012). Numerous studies reported that
Hispanic population growth is associated with crime distribution and increase
(Alvarez-Rivera, Nobles, & Lersch, 2014; Davies & Fagan, 2012; Peterson,
Krivo, & Hagan, 2010; Shaw & Mckay, 1969; Steensmeier, Feldmeyer,
Harris, & Ulmer, 2011). When taking into account the results of previous stu-
dies, it is possible to assume that the characteristics of neighborhoods in
Houston are the reection of the increase of the Hispanic population. Since
social disorganization indicators were negatively associated with police response
time, it is considered as a consistent and reasonable nding, evidencing that
HPD endeavors to respond eectively to domestic violence incidents.

Policy Implications and Future Research


Important policy implications can be derived from this study. Police response
time is closely associated with citizens evaluation of police performance and,
further, attitudes toward police (Kelling & Moore, 1988; McEwen et al., 1986;
Sherman, 1992). The results of this study indicate that situational and neighbor-
hood conditions inuence police response time in domestic violence. The nding
may be consistent in other types of incidents. Police administrators are encour-
aged to take the ndings of this study into consideration when they deploy patrol
units. Careful review of current patrol strategy is suggested.
The ndings of this study suggest that the factors aecting police response
time could dier by type of call since some of the results were inconsistent with
the ndings of previous studies (see Cihan et al., 2012). Future research is
encouraged to examine police response time among other incident types.
In domestic violence cases, suspects are easily identied due to the close relation-
ship between victim and suspect. Furthermore, the implementation of manda-
tory arrest policy for domestic violence may have an impact on apprehension
probability regardless of response time. These may result in a dierent relation-
ship between response time and apprehension probability in domestic violence
cases. Therefore, it is suggested for researchers to examine the eect of response
time on apprehension probability in domestic violence and other types of crime.
Lee et al. 75

Declaration of Conflicting Interests


The authors declared no potential conicts of interest with respect to the research,
authorship, and/or publication of this article.

Funding
The authors received no nancial support for the research, authorship, and/or publica-
tion of this article.

Notes
1. There is not a straightforward measure of neighborhood because a delineation of
neighborhood boundaries has yet to be established in the literature. The current ana-
lysis used census tracts to approximate neighborhood, a method that is widely
accepted in criminological studies (see Bursik & Grasmik, 1993; Cihan, Zhang,
Hoover, 2012; Lee et al., 2013a, 2013b; Wooldredge, 2002).
2. We, the authors, thank numerous police practitioners who offered invaluable and
constructive suggestions during our presentation at the ACJS 2015 annual conference.
The decision to include rush hour in the analysis was based on such feedback.
3. About 20% of the cases were missing suspects information.
4. For consistency in the orientation of the estimated dimensions, residential stability was
reverse coded as residential instability.
5. The index
P is also known as the probability of interspecific encounter and is formulated
as 1  P2, where P is proportion of each race in the corresponding census tract. It
would estimate the probability of randomly picking two different races in the same
census tract.
6. Fixed effect modeling was also considered for its superiority to random effect model-
ing and implemented for the current study (Wooldredge, 2002). Specifically, fixed
effect model would completely control the census tract level variation and thus gen-
erate more accurate estimates for the incident level variable. However, the current
study found that the HLM results are congruent with fixed effect model results.
Also, HLM results, unlike fixed effect model results, show which of the neighborhood
characteristics was significantly associated with response time. For these reasons, it
was decided to report and discuss HLM results in the current study.
7. Age was centered at the mean value of each census tract, rendering the estimate for the
intercept (b0j) an expected response time when the complainant was White female and
at average age, and the call was made in the day time on a weekday.

References
Alpert, G. P., Dunham, R. G., & MacDonald, J. M. (2004). Interactive police-citizen
encounters that result in force. Police Quarterly, 7(4), 475488.
Alvarez-Rivera, L. L., Nobles, M. R., & Lersch, K. M. (2014). Latino immigrant accul-
turation and crime. American Journal of Criminal Justice, 39(2), 315330.
Avakame, E. F., & Fyfe, J. J. (2001). Differential police treatment of male-on-female
spousal violence additional evidence on the leniency thesis. Violence Against Women,
7(1), 2245.
76 Police Quarterly 20(1)

Avakame, E., Fyfe, J., & McCoy, C. (1999). Did you call the police? What did they do?
An empirical assessment of Blacks theory of mobilization of law. Justice Quarterly,
16(4), 765791.
Belknap, J. (1995). Law enforcement officers attitudes about the appropriate responses
to woman battering. International Review of Victimology, 4(1), 4762.
Black, D. (1976). The behavior of law. New York, NY: Academic Press.
Black, D. J. (1980). Production of crime rates. In R. J. Lundman (Ed.), Police behavior:
A sociological perspective (pp. 108129). New York, NY: Oxford University Press.
Blake, L., & Coupe, R. T. (2001). The impact of single and two-officer patrols on catch-
ing burglars in the act a critique of the audit commissions reports on youth justice.
British Journal of Criminology, 41(2), 381396.
Braga, A. A., Papachristos, A. V., & Hureau, D. M. (2010). The concentration and
stability of gun violence at micro places in Boston, 19802008. Journal of
Quantitative Criminology, 26(1), 3353.
Braga, A. A., Papachristos, A. V., & Hureau, D. M. (2014). The effects of hot spots
policing on crime: An updated systematic review and meta-analysis. Justice Quarterly,
31(4), 633663.
Brandl, S. G., & Horvath, F. (1991). Crime-victim evaluation of police investigative
performance. Journal of Criminal Justice, 19(2), 109121.
Bursik, R. J., & Grasmick, H. G. (1993). Economic deprivation and neighborhood crime
rates, 19601980. Law & Society Review, 27, 263.
Buzawa, E. S., & Buzawa, C. (2001). Traditional and innovative police responses to
domestic violence. In R. G. Dunham & G. A. Alpert (Eds.), Critical issues in policing
(pp. 216237). Prospect Heights, IL: Waveland Press.
Cihan, A. (2014). Social disorganization and police performance to burglary calls: A tale
of two cities. Policing: An International Journal of Police Strategies & Management,
37(2), 340354.
Cihan, A., Zhang, Y., & Hoover, L. (2012). Police response time to in-progress burglary a
multilevel analysis. Police Quarterly, 15(3), 308327.
Clawson, C., & Chang, S. K. (1977). The relationship of response delays and arrest rates.
Journal of Police Science and Administration, 5, 5358.
Cohn, E. G. (1993). The prediction of police calls for service: The influence of weather
and temporal variables on rape and domestic violence. Journal of Environmental
Psychology, 13(1), 7183.
Coupe, R. T., & Blake, L. (2005). The effects of patrol workloads and response strength
on arrests at burglary emergencies. Journal of Criminal Justice, 33(3), 239255.
Davies, G., & Fagan, J. (2012). Crime and enforcement in immigrant neighborhoods
evidence from New York City. The Annals of the American Academy of Political
and Social Science, 641(1), 99124.
Eck, J., & Rosenbaum, D. (1994). The new police order: Effectiveness, equity, and effi-
ciency in community policing. In D. Rosenbaum (Ed.), The challenge of community
policing: Testing the promise (pp. 323). Thousand Oaks, CA: Sage.
Erez, E. (1986). Intimacy, violence, and the police. Human Relations, 39(3), 265281.
Eterno, J. A., & Silverman, E. B. (2006). The New York city police departments comp-
stat: Dream or nightmare? International Journal of Police Science & Management, 8(3),
218231.
Lee et al. 77

Fagan, J. (1996). The criminalization of domestic violence: Promises and limits.


Washington, DC: National Institute of Justice.
Feder, L. (1996). Police handling of domestic calls: The importance of offenders presence
in the arrest decision. Journal of Criminal Justice, 24(6), 481490.
Finn, M., & Stalans, L. (1995). Police referrals to shelters and mental health treatment:
Examining their decisions in domestic assault cases. Crime & Delinquency, 41(4),
467480.
Friday, P., Metzgar, S., & Walters, D. (1991). Policing domestic violence: Perceptions,
experience, and reality. Criminal Justice Review, 16(2), 198213.
Garner, J. H., Buchanan, J., Schade, T., & Hepburn, J. (1996). Understanding the use of
force by and against the police. Washington, DC: US Department of Justice, National
Institute of Justice.
Hall, D. L. (2005). Domestic violence arrest decision making: The role of suspect avail-
ability in the arrest decision. Criminal Justice and Behavior, 32(4), 390411.
Hipp, J. R., & Boessen, A. (2013). Egohoods as waves washing across the city: A new
measure of Neighborhood. Criminology, 51(2), 287327.
Isaacs, H. H. (1967). A study of communications, crimes, and arrests in a metropolitan
police department. Task force report: Science and technology. Washington, DC:
Government Printing Office.
Johnson, I. M. (2007). Victims perceptions of police response to domestic violence inci-
dents. Journal of Criminal Justice, 35(5), 498510.
Kaminski, R. J., DiGiovanni, C., & Downs, R. (2003). The use of force between the
police and persons with impaired judgment. Police Quarterly, 7(3), 311338.
Kane, R. J. (2000). Police responses to restraining orders in domestic violence
incidents: Identifying the custody-threshold thesis. Criminal Justice and Behavior,
27(5), 561580.
Kansas City (MO) Police Dept, & United States of America. (1978). Response time ana-
lysis Executive summary. Rockville, MD: National Institute of Justice.
Kelling, G. L., & Moore, M. H. (1988). The evolving strategy of policing. Perspectives on
policing, #4. Cambridge, England: Harvard University, Kennedy School of
Government.
Klinger, D. A. (1997). Negotiating order in patrol work: An ecological theory of police
response to deviance. Criminology, 35(2), 277306.
Kubrin, C. E., & Weitzer, R. (2003). New directions in social disorganization theory.
Journal of Research in Crime and Delinquency, 40(4), 374402.
Kuruvila, M. (2010). Oakland police seek to cut response time: Dispatchers set priorities
on biggest emergencies. Retrieved from http://www.sfgate.com/bayarea/article/
Oakland-police-seek-to-cut-response-time-3189639.php
Lawton, B. A. (2007). Levels of nonlethal force an examination of individual, situ-
ational, and contextual factors. Journal of Research in Crime and Delinquency, 44(2),
163184.
Lee, J. (2014). Police use of nonlethal force in New York City: Situational and commu-
nity factors. Policing and Society, 114 (doi:10.1080/10439463.2014.989162)
Lee, J., Zhang, Y., & Hoover, L. T. (2013a). Police response to domestic violence:
Multilevel factors of arrest decision. Policing: An International Journal of Police
Strategies & Management, 36(1), 157174.
78 Police Quarterly 20(1)

Lee, J., Zhang, Y., & Hoover, L. (2013b). Profiling weapon use in domestic violence:
Multilevel analysis of situational and neighborhood factors. Victims & Offenders, 8(2),
164184.
McEwen, J. T., Connors, E. F., & Cohen, M. I. (1986). Evaluation of the differential police
response field test. US Department of Justice, National Institute of Justice.
Melton, H. C. (1999). Police response to domestic violence. Journal of Offender
Rehabilitation, 29(12), 121.
Mladenka, K. R., & Hill, K. Q. (1978). The distribution of urban police services. The
Journal of Politics, 40(1), 112133.
Olson, L., Anctil, C., Fullerton, L., Brillman, J., Arbuckle, J., & Sklar, D. (1996).
Increasing emergency physician recognition of domestic violence. Annals of
Emergency Medicine, 27(6), 741746.
Pals, H., & Kaplan, H. (2013). Long-term effects of adolescent negative self-feelings on
adult deviance: Moderated by neighborhood disadvantage, mediated by expectations.
American Journal of Criminal Justice, 38(3), 348368.
Pate, A., Ferrara, A., Bowers, R. A., & Lorence, J. (1976). Police response time: Its
determinants and effects. Washington, DC: Police Foundation.
Percy, S. L. (1980). Response time and citizen evaluation of police. Journal of Police
Science and Administration, 8(1), 7586.
Peterson, R. D., Krivo, L. J., & Hagan, J. (2010). Divergent social worlds. New York, NY:
Russell Sage Foundation.
Pridemore, W. A. (2004). Weekend effects on binge drinking and homicide: The social
connection between alcohol and violence in Russia. Addiction, 99(8), 10341041.
Roman, C. G., & Reid, S. E. (2012). Assessing the relationship between alcohol outlets
and domestic violence: Routine activities and the neighborhood environment. Violence
and Victims, 27(5), 811828.
Sampson, R. J. (1986). Effects of socioeconomic context on official reaction to juvenile
delinquency. American Sociological Review, 51, 876885.
Sampson, R. J. (2011). The community. In J. Wilson (Ed.), Crime and public policy
(pp. 293344). New York, NY: Oxford University Press.
Sampson, R. J., & Lauritsen, J. L. (1994). Violent victimization and offending:
Individual-, situational-, and community-level risk factors. Understanding and
Preventing Violence, 3, 1114.
Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent
crime: A multilevel study of collective efficacy. Science, 277(5328), 918924.
Shaw, C. R., & McKay, H. D. (1969). Juvenile delinquency and urban areas. Chicago, IL:
University of Chicago Press.
Sherman, L. W. (1992). Policing domestic violence: Experiments and dilemmas. New York,
NY: The Free Press.
Sherman, L. W., & Berk, R. A. (1984). The specific deterrent effects of arrest for domestic
assault. American Sociological Review, 49, 261272.
Sherman, L. W., Gartin, P. R., & Buerger, M. E. (1989). Hot spots of predatory crime:
Routine activities and the criminology of place. Criminology, 27(1), 2756.
Smith, D. A. (1986). The neighborhood context of police behavior. In A. J. Reiss, Jr &
M. Tonry (Eds.), Crime and justice: A review of research: Vol. 8: Communities and
crime (pp. 313341). Chicago, IL: University of Chicago Press.
Lee et al. 79

Spelman, W., & Brown, D. K. (1984). Calling the police: Citizen reporting of serious crime.
Washington, DC: U.S. Department of Justice.
Steffensmeier, D., Feldmeyer, B., Harris, C. T., & Ulmer, J. T. (2011). Reassessing trends
in black violent crime, 19802008: Sorting out the Hispanic Effect in uniform crime
reports arrests, national crime victimization survey offender estimates, and US pris-
oner count. Criminology, 49(1), 197251.
Stevens, J. M., Webster, T. C., & Stipak, B. (1980). Response time: Role in assessing
police performance. Public Productivity Review, 4, 210230.
Swatt, M. L., Varano, S. P., Uchida, C. D., & Solomon, S. E. (2013). Fear of crime,
incivilities, and collective efficacy in four Miami neighborhoods. Journal of Criminal
Justice, 41(1), 111.
Terrill, W., Paoline, J. A., & Manning, P. K. (2003). Police culture and coercion.
Criminology, 41(4), 10031034.
The City of Houston (2012). Race-ethnicity 19802010. Planning & Development
Department. Retrieved from http://www.houstontx.gov/planning/Demographics/
docs_pdfs/Cy/coh_race_ethn_1980-2010.pdf
Triplett, R. A., Gainey, R. R., & Sun, I. Y. (2003). Institutional strength, social control
and neighborhood crime rates. Theoretical Criminology, 7(4), 439467.
Walker, S. (1992). The police in America. New York, NY: Mcgraw-Hill.
Weisburd, D., Bushway, S., Lum, C., & Yang, S. M. (2004). Trajectories of crime at
places: A longitudinal study of street segments in the city of Seattle. Criminology,
42(2), 283322.
Wooldredge, J. (2002). Examining the (ir)relevance of aggregation bias for multilevel
studies of neighborhoods and crime with an example comparing census tracts to offi-
cial neighborhoods in Cincinnati. Criminology, 40, 681710.
Worden, R. E. (1995). The causes of police brutality: Theory and evidence on police use
of force. In W. A. Gellar & H. Toch (Eds.), And justice for all: Understanding and
controlling police abuses of force (pp. 3160). Washington, DC: Police Executive
Research Forum.
Wright, E. M., & Benson, M. L. (2011). Clarifying the effects of neighborhood context on
violence Behind Closed Doors. Justice Quarterly, 28(5), 775798.

Author Biographies
Jae-Seung Lee is an assistant professor of Criminal Justice in the Department of
Political Science, Criminal Justice, and Organizational Leadership at Northern
Kentucky University. His research focuses on policing strategy, police response
to crime, terrorism, and research methodology.

Jonathan Lee, PhD, is an assistant professor of Criminal Justice in the School of


Public Affairs at Penn State Harrisburg. His research interests lie in police
public relations and police decision making. His work has appeared in
Policing: An International Journal of Police Strategies and Management,
Policing and Society, Crime & Delinquency, and Security Journal, among others.
80 Police Quarterly 20(1)

Larry T. Hoover received his PhD from Michigan State University and has been
on the criminal justice faculty at Sam Houston State University since 1977, and
directs its Police Research Center. He is a past president of the Academy of
Criminal Justice Sciences, and the recipient of the Academys Founders Award,
as well as the O.W. Wilson Award from its Police Section. He was recognized by
Michigan State University by induction to MSUs School of Criminal Justice
Wall of Fame.

You might also like