The legacy of LEMAS: Effects on police scholarship of a federally administered, multi-wave establishment survey

Abstract

Purpose
– Since 1987, the Bureau of Justice Statistics (BJS) has periodically collected data from police agencies in the USA and disseminated these data as the Law Enforcement Management and Administrative Statistics (LEMAS) series. The purpose of this paper is to outline LEMAS's impact on criminal justice scholarship by describing the nexus between policing scholarship and LEMAS, and by analyzing the LEMAS constructs and variables used by researchers in refereed journal articles.

Design/methodology/approach
– A systematic review of the literature is undertaken to better comprehend how scholars use LEMAS variables and constructs. In total, 114 peer-reviewed journal articles were analyzed to parcel out variables and constructs derived from LEMAS data.

Findings
– The paper's analysis reveals that LEMAS is the second-most used BJS data series and the majority of authors use LEMAS to measure elements of organizational structure but not organizational behaviors, outcomes, or outputs.

Originality/value
– The study is the first to systematically identify all peer-reviewed journal articles that utilize LEMAS data. Police organizational research is unique in the fact that most authors agree on the operationalization of variables and constructs.

Full text

The legacy of LEMAS
Effects on police scholarship of a federally
administered, multi-wave establishment survey
Matthew C. Matusiak
Department of Criminal Justice, College of Health and Public Affairs,
Orlando, Florida, USA, and
Bradley A. Campbell and William R. King
Department of Criminal Justice and Criminology,
Sam Houston State University, Huntsville, Texas, USA
Abstract
Purpose – Since 1987, the Bureau of Justice Statistics (BJS) has periodically collected data from police
agencies in the USA and disseminated these data as the Law Enforcement Management and
Administrative Statistics (LEMAS) series. The purpose of this paper is to outline LEMAS’s impact
on criminal justice scholarship by describing the nexus between policing scholarship and LEMAS,
and by analyzing the LEMAS constructs and variables used by researchers in refereed journal articles.
Design/methodology/approach – A systematic review of the literature is undertaken to better
comprehend how scholars use LEMAS variables and constructs. In total, 114 peer-reviewed journal
articles were analyzed to parcel out variables and constructs derived from LEMAS data.
Findings – The paper’s analysis reveals that LEMAS is the second-most used BJS data series and
the majority of authors use LEMAS to measure elements of organizational structure but not
organizational behaviors, outcomes, or outputs.
Originality/value – The study is the first to systematically identify all peer-reviewed journal articles
that utilize LEMAS data. Police organizational research is unique in the fact that most authors agree
on the operationalization of variables and constructs.
Keywords Organizational structure, Policing, Organizational theory, LEMAS
Paper type Literature review
Introduction
Prior to the early 1990s only a few scholars had investigated police agency structure
using macro-level data and comparative frameworks, but these studies were
noteworthy in their uniqueness (Henderson, 1975; Langworthy, 1986; Slovak, 1988).
Before 1994 macro-level, comparative police organizational research was the exception
rather than the rule. Since 1994, however, such studies of police agency structure have
flourished. This growth in the comparative study of police agencies can be attributed
in large part to the Bureau of Justice Statistic’s (BJS) collection and dissemination of the
Law Enforcement Management and Administrative Statistics (LEMAS) data. LEMAS
is a periodic establishment survey of US police agencies that has been administered in
1987, 1990, 1993, 1997, 1999, 2000, 2003, and 2007. The results of LEMAS surveys
appear as BJS reports and as raw data available on the Inter-university Consortium for
Political and Social Research (ICPSR) web site. The wide availability of LEMAS data
has greatly facilitated comparative tests of agencies’ operations and structure and
the inclusion of structural measures in studies that do not directly address agency
structure. As we will soon describe, we identified 114 peer-reviewed journal articles
that used LEMAS data since 1994[1]. Simply, the quality and ease of access to LEMAS
data have made LEMAS a cornerstone of police organizational research.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1363-951X.htm
Received 4 December 2013
Revised 12 March 2014
Accepted 13 March 2014
Policing: An International Journal of
Police Strategies & Management
Vol. 37 No. 3, 2014
pp. 630-648
rEmerald Group Publishing Limited
1363-951X
DOI 10.1108/PIJPSM-12-2013-0117
630
PIJPSM
37,3

LEMAS’s impact on criminal justice scholarship is described by analyzing the
LEMAS constructs and variables used by researchers in refereed journal articles. We
begin with a brief review of police organizational data collection efforts preceding
the first LEMAS survey in 1987. We then describe the LEMAS legacy, and compare its
impact on scholarship in relation to other popular BJS data series, such as the National
Crime Victimization Survey (NCVS). The elements of the LEMAS data that are most
often used by researchers, and LEMAS’s role in fueling police research are described
next. We employ a systematic literature review, identifying studies that used LEMAS
data and then classify studies based on the LEMAS variables and constructs used.
The results convincingly show that the majority of studies that employ LEMAS data
use it to measure organizational structure. A smaller sub-set of studies use LEMAS
to measure law enforcement agency operations. We conclude with recommendations
for BJS and for researchers who use LEMAS data.
Antecedents to LEMAS: multi-wave single point of contact (SPOC)
establishment surveys of police organizations
LEMAS represents a particular type of data collection effort called a multi-wave,
single point of contact (SPOC) establishment survey. Multi-wave surveys are “surveys
that are repeated over time” (Maguire, 2002, p. 41), a definition that includes
panel designs, repeated cross-sectional designs, and hybrid surveys (Maguire, 2002).
We refer to LEMAS as a SPOC because it obtains information about an organization by
asking questions of one person within that organization[2]. Establishment surveys
collect information about an institution or organization’s structure or operations,
as opposed to asking about individual-level information such as a respondent’s
experiences or attitudes (Maguire, 2002). SPOC establishment surveys are generally
better at measuring agency structure (such as size and specialization), operations and
processes (such as policies and due process safeguards), and rudimentary measures of
organizational outputs and inputs (counting calls for service and arrests). SPOC
establishment surveys are less optimal for measuring the dosage and fidelity of
organizational processes, impacts, or outcomes, such as the quality of internal
processes or community policing activities. As a general guide, if a phenomenon can be
counted or quantified with relative ease by a point of contact, that phenomenon is a
good candidate for a SPOC establishment survey (King et al., 2011).
A range of multi-wave SPOC establishment surveys of police agencies have been
conducted since 1930 (see Uchida et al., 1986; Maguire, 2002 for reviews). Some of these
surveys are no longer administered, while other surveys continue to collect data
regularly. For example, the “National Science Foundation’s Research Applied to
National Needs (RANN)” data, collected by Ostrom (1978) included information
from interviews with police administrators from 80 Standard Metropolitan Statistical
Areas was collected just once (Langworthy, 1986). Further, Maguire (2003) surveyed
all law enforcement agencies with more than 100 “full-time sworn officers” (Maguire,
2002, p. 50) two times (e.g. 1996, 1998), but the data have not been collected since 1998.
In regard to ongoing longitudinal surveys, most either do not gather a wide range of
organizational variables or the data are not widely disseminated. To illustrate,
the FBI Police Employee data report a few simple measures of agency structure:
counts of agency employees, differentiation among sworn (since 1930) and civilians
(since 1937), and by gender (since 1970) (Uchida and King, 2002). Other surveys,
such as Washington State University’s periodic survey of municipal police
agencies (see Maguire, 2002, p. 47) contain a greater number of indicators of
631
The legacy of
LEMAS

agency structure, but the data are not readily available through data sources such
as ICPSR.
While these data are useful for police organization scholars, they only provide
cross-sectional snapshots of police agencies or are limited in scope, capturing
information from agencies in just one state (Maguire, 2002). Prior research has called
attention to the importance of studying change in police agencies over time (King,
2009). More specifically, this research postulates that police organizational scholarship
advances through the use longitudinal data to better understand change and
consistency in police agencies over the life course (King, 1999, 2009). Because of its
longitudinal nature, LEMAS is well suited to address these problems by providing
scholars with longitudinal data on police organizations at a national level.
LEMAS
The first LEMAS survey in 1987 represented a major advancement in police
organizational data, in part due to the ready access of the data and the comprehensive
nature of the survey instruments[3]. For example, the 1997 LEMAS contains data
collected from 3,412 agencies with information regarding 534 variables. Waves of
LEMAS have been conducted approximately every 3-4 years over the past 25 years
and the survey instrument changed as questions were refined and retained, while other
questions cycled onto the instrument and were subsequently dropped (Langworthy,
2002). The utility of LEMAS for researchers has remained relatively constant over
the years, because a core group of pertinent constructs have been measured with each
successive wave making it an accessible and powerful research tool. The next section
addresses the utility of LEMAS data relative to other BJS data collection efforts.
LEMAS and other BJS data
LEMAS has had a significant impact on the number of police organizational research
articles published since 1994. LEMAS is also the second most used data set produced
by BJS. To support this contention, a comparison of BJS data collection efforts was
conducted to compare the number of published articles analyzing BJS data. Studies
using BJS data were identified by searching the ICPSR web site, which lists 21 BJS data
collection efforts[4]. ICPSR requests that authors who use BJS data provide a citation of
their published works to ICPSR. ICPSR then posts these citations to their webpage.
Table I identifies the six most frequently utilized BJS data collection efforts according
to these posted citations[5]. For a breakdown of all identified BJS data collection efforts,
please refer to Appendix 1. Note, the National Justice Agency List was not included in
the Appendix, because it was discontinued in 1995 and had only one ICPSR citation as
well as zero journal articles attributed to it as the data source.
Table I compares the use of LEMAS data to other BJS collection efforts.
Submissions to ICPSR were counted and categorized as journal article or thesis/
dissertation (labeled thesis). Non-academic publications were not counted. The first
scholarly publication utilizing LEMAS was published in 1994; therefore, a yearly
publication rate since 1994 was calculated for all BJS data collection efforts. The
publication rate was calculated by dividing the number of publications utilizing
the data sets from 1994 to 2013 by 20 (years). According to Table I, the NCVS is the
most frequently cited BJS data collection effort with an average of 12.35 articles
published per year. LEMAS is the second most frequently cited data set collected by
BJS with an average of 4.05 publications per year. As such, LEMAS appears to be
a valuable and frequently utilized data set collected by BJS. Through the publication
632
PIJPSM
37,3

rate of LEMAS displayed in Table I, it demonstrates the importance and utility of
LEMAS survey data in the survey’s current format, which solicits information from
law enforcement agencies related to organizational structures. Below, we explore the
specific constructs and variables derived from the LEMAS data and the frequency each
variable is used by researchers.
A systematic review of studies using LEMAS, 1987-2013
Published peer-reviewed studies using LEMAS data were analyzed to better
understand how scholars use LEMAS variables and constructs. Studies included in
this analysis were identified via two methods during April 2013. First, the ICPSR
catalogs were searched for studies that reported using LEMAS data. Second, “Google
Scholar” (scholar.google.com) was used to search for and identify studies utilizing
LEMAS data that were not reported to ICPSR. Together the two search methods
produced 1,005 possible publications. It is of note that the search results were not all
unique publications as multiple authors have utilized several waves of LEMAS data in
their research. Additionally, ICPSR and “Google Scholar” searches were largely
duplicative[6]. Further investigation identified 133 publications (80 from ICPSR and 53
from “Google Scholar”) that were probably peer-reviewed journal articles but were not
theses, dissertations, reports, or books. We winnowed this initial list of 133 articles to a
final population of 114 peer-reviewed studies that used LEMAS data by eliminating
non-peer-reviewed journal articles. Each of the 114 articles was analyzed to parcel out
variables derived from the LEMAS data. To identify organizational variables derived
from the LEMAS data the operationalization of those variables as suggested by
Langworthy (1986) and Maguire (2003) were utilized as the framework in this analysis.
The works of Langworthy (1986) and Maguire (2003) are the most complete collections
of defined organizational structure variables related to policing known to the authors.
Based on the operationalization of police organizational structure variables noted
above, the authors reviewed the abstract, methods, and results sections of each article.
All variables derived from the LEMAS data were recorded. When the naming of
variables deviated from the norm (i.e. the works of Langworthy and Maguire), the
variable was recorded under the framework of Langworthy (1986) and Maguire (2003)
based on how each author operationalized the variable in their individual research. The
analysis is structured to discuss the use of LEMAS variables in order of popularity
beginning with the variables and constructs appearing most frequently among the 114
articles (see Figure 1).
Instrument
Year
began
Citations
submitted to
ICPSR
Journal
articles Thesis
Publication
rate since
1994
National Crime Victimization Survey 1986 653 260 45 12.35
Law Enforcement Management and
Administrative Statistics 1987 148 81 24 4.05
Survey of Inmates of State and
Federal Correctional Facilities 1974 238 93 25 3.90
Census of State and Federal Adult
Correctional Facilities 1979 73 39 9 1.90
National Corrections Reporting Program 1983 153 43 11 1.85
Survey of Inmates in Local Jails 1972 75 30 2 1.10
Tabl e I.
Citations and publications
submitted to ICPSR using
BJS data sources
633
The legacy of
LEMAS

Organizational size
Agency size is the most commonly used structural variable in the LEMAS data.
Organizational size appears in 57 of the 114 articles (50 percent) included in the
analysis and indicators of size have been included in all eight LEMAS waves.
Although agency size can be operationalized in a variety of ways (Maguire, 2003), the
most common operationalization counts the number of organizational employees
(Kimberly, 1976; Maguire, 2003). Generally, scholars use LEMAS data to operationalize
agency size as the total number of full-time police employees (both civilian and sworn
personnel). Other, less frequent operationalizations of organizational size use only the
number of full-time sworn (see Zhao and Lovrich, 1997).
Employee demographics
Employee demographics refers to the aggregate demographic composition of
personnel employed by police organizations at the agency level, such as the percent
of sworn officers who are female (Eitle, 2005), or employees classified by race or
ethnicity. Sometimes the percent of employees for a specific category are treated as
a variable, such as when the percent of black or Hispanic police employees is utilized
to examine the demographic composition of police agencies (e.g. Lott, 2000). Other
researchers create single indicators of racial or ethnic heterogeneity for police agencies,
such as with a Ginni Index (Miller, 2013). Agency demographics were used in
37 studies (32.5 percent) that analyzed the LEMAS data.
Functional differentiation/specialization
Functional differentiation or specialization refers to the division of labor and tasks into
specific work groups within an organization (Langworthy, 1986; Maguire, 2003).
Maguire (2003) defines functional differentiation as the “degree to which the
organization divides and assigns its tasks into functionally distinct units” (p. 139). In
the police research using LEMAS, functional differentiation has been operationalized
as the number of special units (e.g. child abuse, gangs, repeat offenders) typically with
57
37
30
22 21 18 16
12 12 12 11 10 7666
31
21
0
10
20
30
40
50
60
Organizational Size
Employee Demographics
Functional Differentiation
Budget/Slack Resources
Formalization
Educational Requirements
Occupational Differentiation
Collective Bargaining/Unionization
Computer/Technology
Vertical Height
Task Scope
Training
Administrative Intensity
Equipment
Pre-Employment Screening
Spatial Differentiation
Community Policing Construct
Other(s)
Figure 1.
Number of peer-reviewed
journal articles utilizing
LEMAS variables and
constructs (1994-2013)
634
PIJPSM
37,3

at least one full-time employee. The LEMAS survey lists a number of specialized units
(e.g. the 1997 LEMAS lists 17 different units). Researchers typically use this list to
create a summated index of specialization where agencies with a unit are coded “one”
while those without a unit are coded “zero.” Functional differentiation was found in
30 studies (26.3 percent) using data gathered by the LEMAS survey instruments.
Budget
LEMAS has regularly collected data on police agency budgets. These data are usually
captured as an overall operating budget (2003 LEMAS) or expenditures by categories
(1997 LEMAS). We include studies that use LEMAS variables on asset forfeiture in this
category as well (Worrall and Kovandizic, 2008). Some researchers use these data to
measure budget (Wilson and Buckler, 2010), but others use budget per population
to measure environmental capacity (Miller, 2013) or resource availability (Randol,
2012). Overall, the sample of studies using LEMAS included 22 articles (19.3 percent),
which included budget in their analysis.
Formalization
Formalization refers to the “formal written rules, policies, standards, and procedures”
(Maguire, 2003, p. 16) that regulate the behavior of organizational employees.
Formalization is typically treated as a summated index that counts the presence or
absence of specific agency rules and relations (e.g. Eitle and Monahan, 2009). For
example, the 1997 wave of LEMAS asked about 15 specific policies that aim to
standardize the conduct of police personnel. In total, 21 (18.4 percent) of the 114 articles
included in the present analysis utilized formalization variables derived from
LEMAS data.
Educational requirements
LEMAS data usually include a categorical variable that examines the educational
requirements for new police recruits mandated by each agency. Educational
requirements captured by the LEMAS data appeared in 18 studies (15.8 percent),
and they are often taken as an indicator of agency professionalism (Smith, 2004).
Civilianization/occupational differentiation
In police organizational research “occupational differentiation reflects the degree to
which an organization relies on specially trained and skilled workers” (Langworthy,
1986, p. 65). Langworthy (1986) argues that occupational differentiation in police
agencies can be measured adequately using the proportion of civilian (i.e. non-sworn)
employees working within an agency. Other researchers have opted to treat civilian
employees as a unique variable called civilianization, which is operationalized as
the proportion of full-time employees who are non-sworn (see Eitle, 2005). Regardless
of whether the variable is called occupational differentiation or civilianization,
civilianization appeared in 16 articles (14.0 percent) using LEMAS data.
Collective bargaining/unionization
Between 1987 and 1993, the LEMAS surveys asked about the unionization of police
employees with a wide-range of questions. Over time the number of unionization
questions has decreased until later waves of LEMAS report only whether
collective bargaining is authorized, which is a question present in all LEMAS waves
except 1999. Earlier waves of LEMAS asked whether membership in different
635
The legacy of
LEMAS

police unions or associations was authorized (1997), or if police membership in various
unions (e.g. regional, state, etc.) was formalized (1987-1993). Researchers have used
unionization variables in 12 studies (10.5 percent).
Computer/technology
LEMAS reports data on the different types of computer or advanced technology
systems available in police agencies, such as the presence of mobile data terminals
(LEMAS 1997), if the agency uses computers for racial recognition (LEMAS 2007) or if
patrol officers have access to computerized warrant records in the field (LEMAS 2007).
The presence of computer and advanced technology data from LEMAS was found in
12 studies (10.53 percent), and is used to examine the impact of these technologies
on police performance, such as by enhancing ability of police to solve crimes
(Roberts, 2012).
Vertical height
Police agency employees may be stratified into various, differential hierarchies,
including hierarchies of formal, organizationally sanctioned power (King, 2005). For
comparative organizational police scholars, the three most often measured aspects
of formal hierarchy have been segmentation, concentration, and vertical height
(Langworthy, 1986). The LEMAS data have consistently provided a measure of vertical
height: the “vertical distance, or the amount of social space, between those at the
bottom of an organization and those at the top” (Maguire, 2003, p. 15). Researchers
regularly use the difference in pay between the chief executive and bottom-level
employees as a proxy for vertical height (King, 1999; Langworthy, 1986; Maguire,
2003). In the current analysis, 12 studies (10.5 percent) using LEMAS data to assess
vertical height were identified.
Task scope
Task scope in policing studies refers to the “primary functions performed by the police
department” (Maguire, 2003, p. 122). Typically, studies examining the structure of
police organizations measure task scope by asking agencies to indicate whether their
agency provides specific services to citizens in their jurisdiction. For example,
Maguire (2003) operationalized this variable by asking agencies if they were
responsible for performing 28 tasks such as patrol, enforcing traffic laws, investigating
crimes, operating jails, and addressing citizen complaints (Maguire, 2003). Agencies
performing a larger number of tasks were considered to have a greater task scope than
agencies responsible for a more limited range of police functions. Task scope was
observed in 11 (9.6 percent) of the 114 articles included in the current analysis.
Training requirements
The LEMAS data usually report the length of the basic academy training (Eitle and
Monahan, 2009), field training required for new police employees, or in-service training
for COP (Morabito, 2010; Rosenbaum et al., 2011). Training variables are often used as
indicators of police professionalism and were found in ten studies (8.8 percent) using
LEMAS data.
Administrative intensity
Administrative intensity, sometimes referred to as administrative overhead, is usually
measured as the proportion of employees within a police agency responsible for
636
PIJPSM
37,3

administrative tasks. Maguire (2003) defines police administrative positions as, “The
chief executive and his or her assistants, and all other personnel who work in an
administrative capacity, including finance personnel, and internal affairs,” as well
as technical support staff such as “dispatchers, clerks, and other personnel providing
support services, including communications and training” (p. 147). Early versions of
LEMAS (1987-1999) counted the number of full-time and part-time sworn and civilian
employees assigned to administration. These questions have not appeared in 1999 and
later waves of LEMAS[7]. Despite the absence of these variables in later waves of
LEMAS, the current study identified seven (6.1 percent) examining administrative
intensity using LEMAS data collection efforts preceding the 1999 LEMAS wave.
Equipment
The LEMAS surveys have always collected information on police equipment such as
vehicles and firearms. Scholars have published on the use of police helicopters and
airplanes (Langton, 2012), bicycles (Holian, 2007), and various police firearms, impact
devices, and armored cars (King, 2000). Overall, the LEMAS equipment variables were
found in six (5.3 percent) of the articles included in this analysis.
Pre-employment screening
Processes used by police agencies to assess and screen potential new employees
consistently appear in the LEMAS data. These screening criteria may include personal
interviews, psychological evaluations, or medical examinations, among others. Found
in six studies (5.3 percent) these pre-employment screening criteria are usually taken
as indicators of agency control and are used in composite measures (Eitle and
Monahan, 2009).
Spatial differentiation
Spatial differentiation generally refers to the geographic spread of police resources
and it is usually operationalized as the distribution of patrol resources in space
(Langworthy, 1986; Maguire, 2003). Spatial differentiation has been measured by
LEMAS (in 1997, 1999, 2000, and 2003) as counts of the number of: police precinct
stations, fixed sub-stations, and mobile stations. Initial LEMAS attempts at counting
the number of patrol beats produced serious data irregularities in the 1990 and 1993
waves, but these issues were later addressed in the 1997 LEMAS survey (see Maguire,
2003, pp. 141-142). Most researchers have been content to use the number of police
stations as a proxy for spatial differentiation (e.g. Eitle, 2005) and have not used
the beat count data. Spatial differentiation was identified in a total of six (5.3 percent)
of the articles analyzed.
Community policing
Beginning in 1993, waves of LEMAS have included measures of community policing
using a range of questions including the presence of community policing officers and
practices in police organizations. Scholars have utilized these variables to approximate
the implementation of community policing. Appearing in 31 studies (27.2 percent), the
most frequently used LEMAS community policing variables include counts of COP
officers, the presence of a written COP plan, whether the agency encourages problem
solving, and the use of fixed COP patrol beats. It should be noted that studies using any
of these variables were included in the variable category “Community Policing.” Had
these variables been examined individually in the current study, their use in prior
637
The legacy of
LEMAS

research would have been too infrequent for inclusion in the analysis. Some studies
use multiple variables to create a summated index of COP implementation (Eitle and
Monahan, 2009; Randol, 2012), while other studies use multiple variables separately in
their analysis (Schnebly, 2008). Further, some studies rely on two (Morabito, 2010) or
three (Lilley and Hinduja, 2007) dichotomous variables to measure community policing.
Other variables
In total, 21 studies (18.4 percent) included 22 LEMAS variables or constructs that are
not discussed above, because these variables appeared fewer than five times in the
literature. Some of these less frequently used LEMAS variables include citizen use of
force complaints (Hickman and Piquero, 2009), citizen review boards (Miller, 2013),
and agency responses to terrorism (Roberts et al., 2012). If individual community
policing variables were independently considered, all community policing variables
would have fallen under this “other” category. It was believed that labeling community
policing variables as “other” would be misleading to the reader, which is why all
community policing variables were included within one community policing category.
LEMAS as a sampling framework
Some authors use the LEMAS sampling framework to conduct their own surveys of
police agencies. Some of these scholars also use LEMAS variables, and when they did,
they are included in the research detailed above. Studies that only used LEMAS as a
sampling framework, however, are included in this section. In all we counted three
articles where LEMAS was used as a framework without using LEMAS variables
(Archbold, 2006; Burns et al., 2004; Kadleck, 2003).
LEMAS as an adjunct to police organizational studies
Our analysis indicates that scholars generally rely on a relatively small number of
variables and constructs in the LEMAS data, most of which are indicators of police
agency structure. In other words, researchers use the same palette of variables and
constructs when they publish articles using data derived from the multiple waves of
LEMAS. The most popular LEMAS items are indicators of organizational structure
including size, functional differentiation, budget, and formalization. Other common
LEMAS constructs are measures of community policing, employment requirements
and screening, and equipment questions involving computers. It is also illustrative to
point out that data from most LEMAS questions do not appear in the literature. Given
the breadth of LEMAS, the majority of questions in the survey are never used, or used
quite rarely by scholars. For example, questions concerning drug testing of police
recruits or employees are ubiquitous in LEMAS. The 1990 LEMAS survey had 43
variables concerning drug testing policies for police employees and the 1993 LEMAS
survey had 45 variables devoted to drug testing[8]. Data from these drug testing
questions, however, have only appeared once in a peer-reviewed journal article (Eitle
and Monahan, 2009). Further, when LEMAS instruments do not include many of the
organizational variables (e.g. LEMAS 1999), the data are used by very few scholars
(see Figure 2). As such, future waves of LEMAS may benefit from the inclusion of the
consistently used organizational variables discussed in the current study.
Variation in the measurement of police organizations
Normal science requires orthodoxy of methods, constructs, and variables (Donaldson,
1996). In other words, just as physicists should agree about the methods for measuring
638
PIJPSM
37,3

velocity and time, organizational scholars should agree about how to measure
organizational size and formalization. LEMAS has contributed greatly to police
organizational research by providing easily accessed data that include a range of
organizational variables and constructs. The operationalization of and terminology
for these variables and constructs is relatively well established in the literature. Our
review of 114 articles that use LEMAS data reveals, however, some variation in the
terms applied to some constructs and variables. The majority of authors use
established, common operationalizations (e.g. Langworthy, 1986; Maguire, 2003) to
measure police organizations. A handful of authors, however, use some common terms
in novel or varying ways. These exceptions are rare, but worth noting. In one
manuscript the variable vertical height (the social distance between the bottom and top
of the organizational hierarchy) was later referred to as formalization (the presence
of codified rules) and as vertical segmentation (the number of command ranks). In a
second manuscript, another group of authors measured civilianization with LEMAS
data, but called it “specialization differentiation” and later in the same manuscript,
“functional specialization.”
We cite these examples of deviations from the norm in naming organizational
constructs in order to highlight the importance of using terms consistently and
accurately. A cumulative body of knowledge about police organizational structure may
only advance when scholars share a common understanding of the constructs and
variables associated with this research. For the past 26 years, LEMAS has advanced a
consistent police organizational research agenda by providing comprehensive data on
police agency structures in the USA.
Recommendations
Several relevant policy recommendations can be informed by the systematic literature
review conducted in this study. First and foremost, the periodic collection of LEMAS
data should continue. This periodic collection, however, should be conducted at
planned intervals rather than as resources become available from the Department of
Justice. The data have fueled a vibrant and healthy academic industry in the empirical
study of, and theory testing with, police agencies. Second, the most often used variables
11
22
27
3
37
13
38
24
6
0
5
10
15
20
25
30
35
40
1987 1990 1993 1995 1997 1999 2000 2003 2007
Figure 2.
Number of peer-reviewed
journal articles utilizing
each LEMAS wave
(1987-2007)
639
The legacy of
LEMAS

and constructs in LEMAS should be regularly collected to maintain the capacity to
conduct longitudinal research of police agencies. Some key indicators of organizational
structure, such as administrative intensity, should be added back to the LEMAS
survey instruments. Alternative or improved measures of other, often used constructs,
should be utilized. For example, employee counts and employee demographics
are the two most often used constructs in the LEMAS data. The 2007 LEMAS survey,
however, gathers data on police employee race only for sworn employees but not for
civilian employees, which was a departure from prior iterations of the survey. On the
other hand, the 1997 LEMAS survey collected data on employee race and gender for
both sworn and non-sworn personnel. We recommend that the 1997 LEMAS questions
for employee race and gender be used in future LEMAS surveys. Overall, inconsistency
in survey question format and variables impedes the ability for researchers to conduct
time series investigations related to police organizational structure. It is recommended
that the LEMAS survey be comprised of a core structural component, supplemented
by various other constructs, a recommendation in agreement with Groves and Cork
(2009). Finally, there are numerous improvements that could be used in future LEMAS
surveys, but we urge innovation with an eye toward the most important and useful
constructs for police organizational scholars as identified by the literature.
The use of LEMAS data in the literature ranges from the clear proponent to the
casual user and everything in between. While some academicians have built their
vitas off the figurative back of LEMAS data, others have simply utilized LEMAS as a
sampling frame for their own unique data collections efforts. Although the above
analysis identifies the importance of organizational variables within the LEMAS
data, additional testing should be undertaken by researchers to address the construct
validity of the standardized measures (Groves and Cork, 2009). As prior researchers
have shown (King et al., 2011), data collection efforts are subject to inconsistencies.
As such, further testing is likely to improve reliability of individual organizational
variables, which would in turn benefit this entire line of research. Additionally,
research in this area enjoys a common operationalization of variables, when
researchers deviate from the norm, they should thoroughly describe what the deviation
seeks to address and how. Finally, it is important to note for organizational researchers,
both in criminal justice and the broader discipline, that these LEMAS data provide
a rich environment for organizational theory testing. LEMAS data are widely available
and easily accessed allowing for replication and validation of prior research.
Conclusion
Police organizational research is a relatively unique literature in which most authors
agree on the operationalization of variables and constructs, which have been
summarized concisely by Maguire (2003). As noted above, LEMAS is an invaluable
data source for police organizational researchers. This value, however, must be framed
in the context that some LEMAS survey waves have proved to be more beneficial
to academic researchers than others. Those waves containing more detailed
organizational variables and constructs (1993, 1997, 2000) have been utilized more
frequently as sources of academic research. As we have noted above (especially
Figure 1) the strength of the LEMAS survey in policing research is related to the
organizational structure variables contained within the survey.
It is imperative that the BJS LEMAS Survey continue to regularly collect detailed
information related to police organizations. As it is the authors’ understanding that the
forthcoming wave of LEMAS data collection diverges from the norm of soliciting
640
PIJPSM
37,3

organizational structure data, there is potential for a LEMAS data collection effort that
will be shunned by academic researchers. The usefulness of LEMAS is evidenced
through its publication rate since 1994, which is second in BJS data collection efforts
only to the NCVS. As demonstrated above, there is great consistency with few
exceptions in how organizational variables are operationalized and utilized. Future
waves of LEMAS should take this consistency and the publication trends exhibited
by the current analysis into consideration when tailoring subsequent LEMAS data
collection efforts. Additionally, future LEMAS researchers should test and retest the
reliability and validity of organizational measures to lend additional support to
the utility of police organizational variables identified above[9].
Notes
1. These 114 journal articles were cited a total of 2,576 times as of April 2013.
2. In reality, a survey may be routed within a police agency to multiple respondents who each
respond to different questions but the number of respondents is small and we think it fair to
call such surveys SPOCs. A SPOC is thus different from surveys distributed to many, most,
or all of an organization’s employees, such as surveys designed to assess organizational
climate.
3. LEMAS has typically used two survey instruments: a short instrument distributed to a
sample of smaller agencies, and a longer survey administered to a census of large agencies.
The 1999 LEMAS used a single instrument, and 1997 used three instruments. See
Langworthy (2002) for a thorough review of LEMAS methodology and survey content.
4. We define a data collection effort as any BJS project, including those spanning multiple
years. Thus, LEMAS and the NCVS include different data files that cover multiple years and
each effort counts as one data collection effort.
5. Table I counts citations that were submitted by study authors to ICPSR. As we will see in the
next section, the ICPSR citation counts do not capture all studies that use ICPSR data
because some authors do not report their publications to ICPSR. Despite this limitation,
we think ICPSR study counts are a reasonable proxy for how often data are used.
6. ICPSR and “Google Scholar” were utilized because they were the most complete centralized
repositories relevant to LEMAS publications. Databases such as EBSCO host (n¼156) were
consulted; however, search results duplicated prior ICPSR and “Google Scholar” results
as well as returning fewer unique publications than the methods selected for further
investigation.
7. The elimination of questions about administrative employees after 1999 is ironic, given that
LEMAS is an acronym for “Law Enforcement Management and Administrative Statistics.”
8. The 1997 LEMAS survey had 15 variables about employee drug testing, the 1999 LEMAS
had zero variables, and the 2000 and 2003 LEMAS surveys had one variable each.
9. See Appendix 2 for a complete list of studies that were included in the analysis of the use of
LEMAS.
References
Archbold, C.A. (2006), “Police legal advisors in the USA: past, present, and future”, Police Practice
and Research: An International Journal, Vol. 7 No. 1, pp. 61-76.
Burns, R.G., Whitworth, K.H. and Thompson, C.Y. (2004), “Assessing law enforcement
preparedness to address internet fraud”, Journal of Criminal Justice, Vol. 32 No. 5,
pp. 477-493.
641
The legacy of
LEMAS

Donaldson, L. (1996), “The normal science of structural contingency theory”, in Clegg, S.R.,
Hardy, C. and Nord, W.R. (Eds), Handbook of Organization Studies, Sage Publications,
Thousand Oaks, CA, pp. 57-76.
Eitle, D. (2005), “The influence of mandatory arrest policies, police organizational characteristics,
and situational variables on the probability of arrest in domestic violence cases”, Crime &
Delinquency, Vol. 51 No. 4, pp. 573-597.
Eitle, D. and Monahan, S. (2009), “Revisiting the racial threat thesis: the role of police
organizational characteristics in predicting race-specific drug arrest rates”, Justice
Quarterly, Vol. 26 No. 3, pp. 528-561.
Groves, R. and Cork, D. (2009), Ensuring the Quality, Credibility, and Relevance of U.S. Justice
Statistics, National Academies Press, Washington, DC.
Henderson, T.A. (1975), “The relative effects of community complexity and of sheriffs upon the
professionalism of sheriff’s departments”, American Journal of Political Science, Vol. 19
No. 1, pp. 107-132.
Hickman, M.J. and Piquero, A.R. (2009), “Organizational, administrative, and environmental
correlates of complaints about police use of force: Does minority representation matter?”,
Crime & Delinquency, Vol. 15 No. 1, pp. 3-27.
Holian, M. (2007), “Compstat, community policing and the science of success: a market-based
approach to police management”, Economic Affairs, Vol. 27 No. 4, pp. 23-29.
Kadleck, C. (2003), “Police employee organizations”, Policing: An International Journal of Police
Strategies & Management, Vol. 26 No. 2, pp. 341-351.
Kimberly, J.R. (1976), “Organizational size and the structuralist perspective: a review, critique,
and proposal”, Administrative Science Quarterly, Vol. 21 No. 4, pp. 571-597.
King, W.R. (1999), “Time, constancy, and change in American municipal police organizations”,
Police Quarterly, Vol. 2 No. 3, pp. 338-364.
King, W.R. (2000), “Measuring police innovation: issues and measurement”, Policing: An
International Journal of Police Strategies & Management, Vol. 23 No. 3, pp. 303-317.
King, W.R. (2005), “Toward a better understanding of the hierarchical nature of police
organizations: conception and measurement”, Journal of Criminal Justice, Vol. 33 No. 1,
pp. 97-109.
King, W.R. (2009), “Toward a life-course perspective of police organizations”, Journal of Research
in Crime and Delinquency, Vol. 46 No. 2, pp. 213-244.
King, W.R., Cihan, A. and Heinonen, J. (2011), “The reliability of police employee counts:
comparing FBI and ICMA data, 1954-2008”, Journal of Criminal Justice, Vol. 39 No. 5,
pp. 445-451.
Langton, L. (2012), “Engaging in a more complete assessment of the operations of airborne
police units: a research note”, Police Practice and Research: An International Journal,
Vol. 15 No. 1, pp. 17-34.
Langworthy, R.H. (1986), The Structure of Police Organizations, Praeger, New York, NY.
Langworthy, R.H. (2002), “LEMAS: a comparative organizational research platform”, Justice
Research and Policy, Vol. 4, pp. 21-38.
Lilley, D. and Hinduja, S. (2007), “Police officer performance appraisal and overall satisfaction”,
Journal of Criminal Justice, Vol. 35 No. 2, pp. 137-150.
Lott, J.R. Jr. (2000), “Does a helping hand put others at risk? Affirmative action, police
departments, and crime”, Economic Inquiry, Vol. 38 No. 2, pp. 239-277.
Maguire, E.R. (2002), “Multiwave establishment surveys of police organizations”, Justice
Research and Policy, Vol. 4, pp. 39-60.
642
PIJPSM
37,3

Maguire, E.R. (2003), Organizational Structure in American Police Agencies: Context, Complexity,
and Control, State University of New York Press, Albany, NY.
Miller, K. (2007), “Racial profiling and postmodern society: police responsiveness, image
maintenance, and the left flank of police legitimacy”, Journal of Contemporary Criminal
Justice, Vol. 23 No. 3, pp. 248-262.
Miller, K. (2013), “The institutionalization of racial profiling policy: an examination of
antiprofiling policy adoption among large law enforcement agencies”, Crime and
Delinquency, Vol. 59 No. 1, pp. 32-58.
Morabito, M.S. (2010), “Understanding community policing as an innovation: patterns of
adoption”, Crime and Delinquency, Vol. 56 No. 4, pp. 564-587.
Ostrom, E. (1978), Decision Related Research on the Organization of Service Delivery
Systems in Metropolitan Areas: Police Protection, University of Michigan, Ann
Arbor, MI.
Randol, B.M. (2012), “The organizational correlates of terrorism response preparedness
in local police departments”, Criminal Justice Policy Review,Vol.23No.3,
pp. 304-326.
Roberts, A. (2012), “Motor vehicle recovery: a multilevel event history analysis of NIBRS data”,
Journal of Research in Crime and Delinquency, Vol. 49 No. 3, pp. 444-467.
Roberts, A., Roberts, J.M. and Liedka, R.V. (2012), “Elements of terrorism preparedness
in local police agencies, 2003-2007 impact of vulnerability, organizational
characteristics, and contagion in the Post-9/11 era”, Crime & Del inquency, Vol. 58 No. 5,
pp. 720-747.
Rosenbaum, D.P., Graziano, L.M., Stephens, C.D. and Schuck, A.M. (2011), “Understanding
community policing and legitimacy-seeking behavior in virtual reality: a national study of
municipal police web sites”, Police Quarterly, Vol. 14 No. 1, pp. 25-47.
Schnebly, S.M. (2008), “The influence of community-oriented policing on crime-reporting
behavior”, Justice Quarterly, Vol. 25 No. 2, pp. 223-251.
Slovak, J.S. (1988), Styles of Urban Policing: Organization, Environment, and Police Styles in
Selected American Cities, New York University Press, New York, NY.
Smith, B.W. (2004), “Structural and organizational predictors of homicide by police”,
Policing: An International Journal of Police Strategies & Management, Vol. 27 No. 4,
pp. 539-557.
Uchida, C.D. and King, W.R. (2002), “Police employee data: elements and validity”, Justice
Research and Policy, Vol. 4, pp. 1-9.
Uchida, C.D., Bridgeforth, C. and Wellford, C.F. (1986), “Law enforcement statistics: the state of
the art”, American Journal of Police, Vol. 5 No. 1, pp. 23-43.
Wilson, S. and Buckler, K. (2010), “The debate over police reform: examining minority support
for citizen oversight and resistance by police unions”, American Journal of Criminal
Justice, Vol. 35 No. 4, pp. 184-197.
Worrall, J.L. and Kovandizic, T.V. (2008), “Is policing for profit? Answers from asset forfeiture”,
Criminology & Public Policy, Vol. 7 No. 2, pp. 219-244.
Zhao, J. and Lovrich, N. (1997), “Collective bargaining and the police: the consequences for
supplemental compensation policies in large agencies”, Policing: An International Journal
of Police Strategies & Management, Vol. 20 No. 3, pp. 508-518.
(The Appendix follows overleaf.)
643
The legacy of
LEMAS

Appendix 1
Instrument
Year
began
Citations
submitted
to ICPSR
Journal
articles Thesis
Publication
rate
since 1994
National Crime Victimization Survey 1986 653 260 45 12.35
Law Enforcement Management and Administrative Statistics 1987 148 81 24 4.05
Survey of Inmates of State and Federal Correctional Facilities 1974 238 93 25 3.90
Census of State and Federal Adult Correctional Facilities 1979 73 39 9 1.90
National Corrections Reporting Program 1983 153 43 11 1.85
Survey of Inmates in Local Jails 1972 75 30 2 1.10
Capital Punishment in the USA 1972 72 19 5 0.90
National Jail Census 1970 43 18 1 0.60
Federal Justice Statistics Program 1979 71 10 4 0.50
Annual Survey of Jails 1982 65 9 2 0.40
National Prosecutors Survey 1990 31 8 1 0.40
Census of State and Local Law Enforcement Agencies 1986 21 7 2 0.35
Census of Public and Private Juvenile Detention, Correctional,
and Shelter Facilities 1971 39 10 3 0.30
National Judicial Reporting Program 1986 54 8 3 0.25
Offender Based Transaction Statistics 1979 28 13 3 0.25
Annual Probation Survey and Annual Parole Survey 1980 4 2 1 0.20
Survey of Campus Law Enforcement Agencies 1995 7 4 1 0.20
Expenditure and Employment Data for the Criminal Justice System 1971 59 4 0 0.15
Annual Survey of Jails in Indian Country 1998 12 0 0 0.00
National Prison Rape Statistics Program 2007 6 0 0 0.00
Table AI.
Citations and publications
submitted to ICPSR using
all BJS data sources
644
PIJPSM
37,3

Appendix 2. Peer-reviewed journal articles utilizing LEMAS variables and
constructs (1994-2013)
(1) Bishopp and Worrall (2009)
(2) Briggs et al. (2008)
(3) Bromley and Reaves (1998a)
(4) Bromley and Reaves (1998b)
(5) Burns et al. (2004)
(6) Burruss and Giblin (2009)
(7) Chamlin and Sanders (2010)
(8) Chappell et al. (2006)
(9) Choi (2011)
(10) Choi and Choi (2012)
(11) Cordner and Cordner (2011)
(12) D’Alessio et al. (2005)
(13) Dalehite (2008)
(14) Decker and Pyrooz (2010)
(15) DeLone (2007)
(16) Dichter et al. (2011)
(17) Donohue and Levitt (2001)
(18) Eitle (2005)
(19) Eitle and Monahan (2009)
(20) Eitle et al. (2005)
(21) Falcone and Wells (1995)
(22) Garicano (2010)
(23) Garicano and Heaton (2010)
(24) Gau et al. (2013)
(25) Greene and del Carmen (2002)
(26) Haberman and King (2010)
(27) Hassell et al. (2003)
(28) He et al. (2005)
(29) Helms (2007)
(30) Helms (2008)
(31) Helms and Costanza (2009)
(32) Helms and Gutierrez (2007)
(33) Hickman and Piquero (2009)
645
The legacy of
LEMAS

(34) Holcomb et al. (2011)
(35) Holian (2007)
(36) Holmes et al. (2008)
(37) Hummer (2004)
(38) Jaramillo et al. (2005)
(39) Jenness and Grattet (2005)
(40) Johnson (2013)
(41) Jones-Webb and Wall (2008)
(42) Kadleck (2003)
(43) Kaminski and Stuckey (2009)
(44) Katz et al. (2002)
(45) Kim and Guzman (2012)
(46) Kim and Mengistu (1994)
(47) King (2007)
(48) King (1999)
(49) King (2000)
(50) King (2009)
(51) King and Lab (2000)
(52) Langton (2012)
(53) Lee et al. (2010)
(54) Lilley and Hinduja (2006a)
(55) Lilley and Hinduja (2006b)
(56) Lilley and Hinduja (2007)
(57) Lonsway (2003)
(58) Lord et al. (2009)
(59) Lott (2000)
(60) MacDonald (2002)
(61) Maguire (1997)
(62) Maguire (2009)
(63) Maguire et al. (2003)
(64) Marschall and Ruhil (2007)
(65) Marschall and Shah (2007)
(66) Martin (1995)
(67) McCabe and Fajardo (2001)
(68) McCluskey and McCluskey (2004)
646
PIJPSM
37,3

(69) Miller (2007)
(70) Miller (2013)
(71) Morabito (2008)
(72) Morabito (2010)
(73) Morash et al. (2006)
(74) Morash et al. (2006)
(75) Murphy and Worrall (1999)
(76) Nicholson-Crotty and O’Toole (2004)
(77) Paoline and Sloan (2003)
(78) Paoline et al. (2000)
(79) Pyrooz (2012)
(80) Pyrooz et al. (2010)
(81) Randol (2012)
(82) Redmond and Baveja (2002)
(83) Roberts (2008)
(84) Roberts (2012)
(85) Roberts and Block (2012)
(86) Roberts and Roberts (2007)
(87) Roberts and Roberts (2009)
(88) Roberts et al. (2012)
(89) Rosenbaum et al. (2011)
(90) Sass and Mehay (2003)
(91) Sass and Troyer (1999)
(92) Schnebly (2008)
(93) Sever (2001)
(94) Sever and McSkimming (2004)
(95) Sharp (2006)
(96) Sharp and Johnson (2009)
(97) Sklansky (2006)
(98) Smith (2003)
(99) Smith (2004)
(100) Smith and Holmes (2003)
(101) Sozer and Merlo (2012)
(102) Weisburd and Lum (2005)
(103) Weitzer (1999)
647
The legacy of
LEMAS

(104) Weitzer (2000)
(105) Wilson (2003)
(106) Wilson (2004)
(107) Wilson and Buckler (2010)
(108) Wilson and Zhao (2008)
(109) Worrall (1998)
(110) Worrall (2001)
(111) Worrall and Kovandzic (2007)
(112) Worrall and Kovandzic (2008)
(113) Xie and Lauritsen (2012)
(114) Zhao and Lovrich (1997)
About the authors
Dr Matthew C. Matusiak, PhD, is an Assistant Professor at the University of Central Florida,
after recently graduating from the Sam Houston State University. His research interests include
organizational theory, police organizational structure, and police investigative processes. He is
also interested in working with law enforcement organizations to conduct program evaluations
or assist the organization from an academic standpoint. Dr Matthew C. Matusiak is the
corresponding author and can be contacted at: matthew.matusiak@ucf.edu
Bradley A. Campbell, MA, is currently in the PhD program in criminal justice at the Sam
Houston State University. Bradley’s Research interests are mainly focussed on policing,
particularly police organizations, police investigations, police socialization processes, and police
training. Further interests include qualitative research methods and eyewitness identification
procedures.
Dr William R. King, PhD, is an Associate Professor and Associate Dean for Research and
Program Development at the Sam Houston State University. Between 1997 and 2009 he served as
a Faculty Member and as the Director of the Crime and Justice Research Laboratory at Bowling
Green State University, in Ohio. His research interests include the quantitative and comparative
study of police organizational structure, the process of criminal investigations, and studying
forensics systems from an organizational and theoretical perspective.
To purchase reprints of this article please e-mail: reprints@emeraldinsight.com
Or visit our web site for further details: www.emeraldinsight.com/reprints
648
PIJPSM
37,3