INFORMATION TECHNOLOGY AND CRIMINAL INVESTIGATIONS: DOES IT

MATTER?

by

Hakan Hekim

A dissertation submitted to the faculty of

The University of North Carolina at Charlotte
in partial fulfillment of the requirements
for the degree of Doctor of Philosophy in
Public Policy

Charlotte

2009

Approved by:

____________________________
Dr. Vivian B. Lord

____________________________
Dr. Robert Brame

____________________________
Dr. Thomas J. Holt

____________________________
Dr. Cynthia C. Combs

i
 ii

©2009
Hakan Hekim
ALL RIGHTS RESERVED

ii
 iii

ABSTRACT

HAKAN HEKIM. Information Technology and Criminal Investigations: Does IT Matter?

(Under the direction of DR.VIVIAN B. LORD)

As an important police activity, criminal investigation heavily relies on recovery,

analysis and interpretation of information. Since information technologies facilitate

creation, storage, retrieval, process and dissemination of information, police departments

have extensively employed them in investigative activities. However, information

technologies’ impact on the outcome of criminal investigations is still unclear. This study

examines whether use of information technologies makes an impact on the outcome of

criminal investigations. The analysis results revealed that there is not a consistent

relationship between police departments’ use of information technologies and their

clearance rates. In addition to this finding, this study made an important methodological

contribution to the literature by showing that unbalanced data may be an important source

of concern for observational studies of police departments, and it should be carefully

investigated before conducting any inferential analysis.

iii
 iv

ACKNOWLEDGMENTS

First, I would like to acknowledge the Turkish National Police Scholarship that I

received for my graduate study, and thank to everyone who made it possible. Then I must

thank to my chairperson Doctor Vivian B. Lord for her support and supervision,

committee members Doctor Robert Brame and Doctor Thomas J. Holt for their insightful

comments and Doctor Cynthia C. Combs for her review. I also must thank to Doctor

Kathryn E. Scarborough from Eastern Kentucky University for her valuable contribution

to this study. Finally, I thank to my wife Nur for her support, patience and enduring

friendship.
 v

TABLE OF CONTENTS

LIST OF TABLES vii

LIST OF FIGURES viii

CHAPTER 1: INTRODUCTION 1

The Purpose and Significance of the Study 1

Scope of the Study 3

Overview of the Chapters 7

CHAPTER 2: LITERATURE REVIEW 8

Criminal Investigations 8

Definition and the Goals of Criminal Investigation 9

Factors Affecting the Outcome of Criminal Investigation 13

Police and Information Technologies 24

Definitions 24

Development of Law Enforcement Information Technologies in the U.S. 26

Types of Law Enforcement Information Technologies 29

Impact of Information Systems 40

CHAPTER 3: METHODOLOGY 57

Research Question and Hypothesis 57

Data 58

Variables 68

CHAPTER 4: RESULTS 77

Univariate Analysis 77

Bivariate Analysis 89
 vi

Multivariate Analysis 97

CHAPTER 5: CONCLUSION, LIMITATIONS AND POLICY IMPLICATIONS 115

REFERENCES 127

APPENDIX A: CRIMINAL INVESTIGATION SURVEY 143

APPENDIX B: CORRELATION COEFFICIENTS OF USE OF IT ITEMS 160

APPENDIX C. FACTOR ANALYSIS RESULT 166

APPENDIX D: DESCRIPTIVE STATISTICS 167

APPENDIX E: HISTOGRAMS OF ALL VARIABLES 170

APPENDIX F: CORRELATION ANALYSIS RESULTS 184

APPENDIX G: MODERATE AND HIGH IT DEPARTMENT COMPARISON 188

APPENDIX H: DECILES OF USE OF IT SCALE 193

APPENDIX I: TEN GROUPS OF LOGGED NUMBER OF OFFICERS 195

APPENDIX J: REGRESSION ANALYSIS RESULTS 197

APPENDIX K REGRESSION ANALYSIS OF MODERATE/HIGH IT GROUPS 201

APPENDIX L: REGRESSION ANALYSIS OF TEN GROUPS 209

APPENDIX M: ITEMS WITH OVERLAPPING AGENCY SIZE 224

APPENDIX N: ITEMS WITH NO OR POOR OVERLAPPING AGENCY SIZE 227

APPENDIX O: REGRESSION COEFFICIENTS OF USE OF IT ITEMS 232

 vii

LIST OF TABLES

TABLE 1: Classification of Police Information Technologies 31

TABLE 2: Police Departments’ Response Rates to the Impact of Science 60

and Technology Survey

TABLE 3: UCR Part I Offense Hierarchy 65

TABLE 4: Percentage of Crimes Cleared by Arrest or Exceptional Means 69

in 2005 and 2006

TABLE 5: Number of Zero Clearances 78

TABLE 6: Regression Coefficients of Logged Number of Officers 97

Variable Predicting Use of IT Scale

TABLE 7: 2005 and 2006 Murder and Vehicle Theft Models 101

TABLE 8: Complex Models of Murder and Vehicle Theft Clearance 105

Rates Fitted to 2005-2006 Moderate and High IT Departments
 viii

LIST OF FIGURES

FIGURE 1: DeLone & McLean IS Success Model 42

FIGURE 2: Mean values of 2005 and 2006 crime rates 80

FIGURE 3: Mean values of 2005 and 2006 clearance rates 81

FIGURE 4: Standard deviations of 2005 and 2006 clearance rates 82

FIGURE 5: Histograms and box plots of 2005 murder and motor vehicle 83
theft clearance rates

FIGURE 6: Histograms and box plots of 2006 murder and motor vehicle 84
theft clearance rates

FIGURE 7: Histograms and box plots of 2005 murder and motor vehicle 85
theft clearance rates after square root transformation

FIGURE 8: Histograms and box plots of 2006 murder and motor vehicle 85
theft clearance rates after square root transformation

FIGURE 9: Histogram of Use of Information Technology scale 86

FIGURE 10: Box plots of standardized values of six control variables 88

FIGURE 11: Scatter plot of logged number of officers and use of IT scale 92
variables

FIGURE 12: Scatter plot of logged number of officers and use of IT scale 95
variables, and box plot comparison of use of IT scale
across groups.

FIGURE 13: Regression coefficients of Use of IT scale estimated from 100

2005 and 2006 data

FIGURE 14: Box plot Comparisons of variables between moderate and 103
high IT departments

FIGURE 15: Regression coefficients of Use of IT scale of moderate and 104

high IT departments estimated from 2005 and 2006 data
FIGURE 16: 2005 - Regression coefficients of Use of IT scale after 107
dividing sample into ten groups
 ix

FIGURE 17: 2006 - Regression coefficients of Use of IT scale after 107

dividing sample into ten groups

FIGURE 18: Use of IT items that have good overlap and very poor or no 111
overlap

FIGURE 19: Bar graphs of use of information technology items’ 112

regression coefficients for murder and vehicle theft
offenses
 CHAPTER 1: INTRODUCTION

The Purpose and Significance of the Study

Investigative police work is mostly about the recovery, analysis and interpretation

of information about criminal offenses (Osterburg and Ward, 2007). During the

investigation of a crime, police take victim and witness statements, examine the crime

scene, collect physical evidence, analyze collected material to identify the offender or

offenders and then present the resulting information to the prosecutor. Since the success

of criminal investigation depends on quality information, the production of quality

information is the first objective of investigative work. As Luen & Al-Hawamdeh (2001)

state timely and accurate information is critical to the success of policing. In order to

increase the probability of generating quality information, the police employ information

technologies. Information technologies appear as important instruments of criminal

investigations because they facilitate creation, storage, retrieval, transfer, and application

of investigation-related information (Gottschalk, 2007). Moreover, information

technologies may help produce effective use of time devoted to criminal investigation by

automating some routine investigative tasks (Eck, 1983; Folk, 1971). Nevertheless, the

link between information technologies and the outcome of criminal investigation is not a

clear one. Therefore, the purpose of this study is to examine whether information

technologies really make any contribution to the outcome of criminal investigations.

1
 2

Although criminal investigation is an important law enforcement activity, it is one

of the least studied police functions. Early influential studies in the field were conducted

in the 1970s (e.g., Chaiken, Greenwood, and Petersilia, 1976; Greenberg and Lang, 1973;

Greenwood 1970). Those studies generally explored the criminal investigation process

and analyzed its effectiveness in solving crimes. Perhaps as a result of consistent negative

findings, researchers diverted their attention to the determinants of success of criminal

investigation in the 1980s, and this trend continued through the 1990s (e.g., Burrows and

Tarling, 1982; Marche, 1994; Welford and Cronin, 1999). Although Horvath, Meesig and

Lee (2001) argue that criminal investigation is again becoming a focus of interest among

criminal justice researchers and practitioners, the number of studies conducted in the field

has not increased much yet. This study intends to contribute to the discussion about the

effectiveness of criminal investigation by presenting an extensive review of the previous

literature and providing new empirical evidence.

Police use of information technology (IT) is another issue that has not received

much attention from scholars. One of the first systemic approaches to the topic was made

in the 1960s when President’s Commission decided to establish the Task Force on

Science and Technology to explore possible contributions of science and technology to

the fight against crime. Blumstein, who was the Director of the Task Force, stated that the

Task Force began its work “with the then widely accepted cliché that ‘if science and

technology can get a man to the moon, then certainly it must have some important

contributions to make in the realm of … controlling crime’” (Blumstein, 1994, p. 146) As

a result, 13 out of 200 recommendations published in the President’s Commission report,

“The Challenge of Crime in a Free Society,” were about use of science and technology

2
 3

within the criminal justice system. Then, in 1968, President Lyndon Johnson had

expressed his plan to bring the most advanced technology to the war on crime, and the

Law Enforcement Administration Assistance (LEAA) was established to deliver on that

promise. In the same year, President Johnson signed the Safe Streets Act, and until 1982

a huge amount of federal aid was sent to the state and local governments for their

technology investments. Today, information technologies are critical to the police. Davis

and Jackson (2004) reported that 100 percent of police departments serving cities with

populations of 250,000 or more were using Computer Aided Dispatch (CAD) systems,

and 92 percent of them were using in-field computers or terminals. Moreover, 71 percent

of all local police departments reported to have enhanced 911 systems, and 40 percent of

them were using in-field computers or terminals (Davis & Jackson, 2004). New

information technologies have been developed by different vendors and implemented in

law enforcement agencies, but their impact on police work is still a controversial issue,

and available literature is very limited (Abt Associates, 2001). This study provides

additional empirical evidence to the criminal justice community about information

technologies’ impact on police work.

In conclusion, this study intends to contribute to policing research by focusing on

the intersection of two understudied topics: criminal investigations and police

information technologies.

Scope of the Study

This study analyzes the relationship between use of information technologies by

U.S. police departments and the outcome of criminal investigations. Its unit of analysis is

police departments, and the dependent variable is the clearance rates achieved by the
 4

police departments. It must be noted that this study is not testing any causal relationship

between use of information technologies and clearance rates. In order to be able to decide

whether a causal relationship exists, three criteria must be considered: association,

temporal order and nonspuriousness. In other words, there must be covariation between

dependent and independent variables, variation in the dependent variable should occur

after the variation in the independent variable and variation should not be due to variation

in a third variable (Schutt, 2006). This study is an observational study and it only tests

whether there is an association between use of information technologies and clearance

rates; however, it cannot tell if there is a causal relationship between independent and

dependent variables.

Crime and clearance data of this study is collected by the Unified Crime

Reporting (UCR) Program. The UCR data and the use of clearance rate as a success

measure for police departments have been criticized for several reasons. First of all, UCR

data reports Part I crimes known to the police, and participation to the UCR program is

voluntary. In addition, there is not a monitoring mechanism in place to ensure the

accuracy of reported crimes. Therefore, it was argued that UCR data does not portray an

accurate picture of crime problem in the U.S (Chambliss, 2001; Lott & Whitley, 2003;

Martin & Legault, 2005). Despite of these critiques, UCR data is still used by researchers

(e.g. Helms, 2008; Holmes, Smith, Freng & Munoz, 2008; Nolan, 2004; O’Brian, 2003).

In addition, some researchers argued that UCR is a valid source of criminal statistics in

the U.S. (Gove, Hughes & Geerken, 1985; Hindelang, 1974). Use of clearance rate as a

measure of police performance is criticized mainly for two reasons. Firsty, in order to

look better on the statistics, police departments can manipulate the crime and clearance
 5

numbers (Chambliss, 2001). Secondly, clearance rates do not give any information about

the difficulty of cases and the amount of effort police put in them (Pare, Felson, Quimet,

2007). However, clearance rates are also still used by researchers mainly because of lack

of a better measure (e.g., Borg & Parker, 2001; Keel, Jarvis & Muirhead, 2009; Pare et

al., 2007).

Although some criminal investigation studies differentiate the stages in which

arrest or clearance occurs (e.g., Isaacs, 1967; Greenwood, 1970), this study takes criminal

investigation as a whole and makes no distinction between different stages of the

investigative process. Clearance is viewed as the final output of this process, and no

attempt is made to determine in which stage it occurs. This approach is appropriate for

this study for two reasons. First of all, the main concern of this study is information

technologies’ impact on the outcome of criminal investigation. Despite of the differences

in information technology use across the stages of criminal investigation, information

technologies can be utilized throughout the entire investigation process (Dunworth, 2004;

Manning, 1992). Therefore, a holistic approach to the criminal investigation process is

necessary to better assess information technologies’ impact. This approach is also used

by some other researchers in the field (e.g. Chan, 2001; Hauck and Chen, 1999; Manning,

2001) In addition to this, usually the rationale for analyzing preliminary and secondary

investigation stages separately is to focus particularly on one of the stages and to assess

the effectiveness of the patrol or detective work (e.g. Marshall, 1998; Nunn, 1994; Nunn

& Quinet, 2002). Since that is not a goal of this study, taking criminal investigation as a

whole is more appropriate.

 6

The independent variable of this study is use of information technologies by

police departments. The information technology concept has a very broad meaning, and

there are several different ways to classify information technologies used by the police.

Hogan and Radack (1997) defined information technology as the blend of hardware and

software used for storage, processing, transfer, display, management, organization, and

retrieval of information. The broadness of the definition depicts the variety of shapes and

forms that information technologies can take. In fact, the literature review will show that

police departments use information technologies for many administrative and operational

purposes. While administrative use of information technologies consists of traditional

support services such as personnel, budget, fleet management and inventory systems, the

operational use of information technologies is related to law enforcement and

investigative activities of the police. As will be discussed in the literature review,

investigative use of police information technologies can be analyzed within four areas:

records management, crime analysis, mobile computing and forensic analysis. In this

respect, the independent variable of this study is an index variable measuring police

departments’ use of information technologies in those areas. The independent variable

composes of 29 items collected by the “National Study of the Impact of Science and

Technology on the Process of Criminal Investigation in Law Enforcement Agencies”

project conducted by Eastern Kentucky University researchers.

DeLone and McLean’s (1992; 2003) information system success model is used in

this study. According to the model, information success can be measured at six levels:

information quality, system quality, service quality, use and intention to use, user

satisfaction and net benefits In addition, the model proposes a relationship between these
 7

six levels of success. DeLone and McLean’s information system success model is a

highly complex model; for that reason, researchers have generally tested parts of it, and

usually found supportive results (e.g. McGill, Hobbs & Klobas, 2004; Roldan & Millan,

2000; Seddon & Kiew, 1996). Similar to those studies, this study only looks at the

information system success at system use level.

Overview of the Chapters

This study consists of five chapters. The first chapter provides an introduction to

the study and includes purpose and significance of the study, scope of the study, and

overview of the chapters.

Chapter Two has two major parts. The first part provides a literature review of

criminal investigation. The definition and the goals of criminal investigation, and

empirical studies about the factors affecting the outcome of criminal investigation are

discussed in this part. The second part of Chapter Two provides a literature review of

police use of information technologies. This part includes development of police

technologies in the U.S., types of information technologies used by the police, and

empirical studies on the impact of information technologies in law enforcement

organizations.

In Chapter Three, the research question, research methodology, and variables of

the study are presented. Chapter Four provides data analysis and results. Finally, Chapter

Five provides a discussion of findings, limitations, and the policy implications of this

study.
 CHAPTER 2: LITERATURE REVIEW

The literature review chapter includes two major parts. In the first part, the

literature examining criminal investigation and its goals and the factors affecting the

outcome of criminal investigations are presented. In the second part, the literature

concerning the development and use of police information systems and the impact of

information technologies on police work is presented.

Criminal Investigations

Criminal investigation is one of the central tasks of law enforcement agencies. In

the public’s view, criminal investigation may be the most important police task because

most people only contact the police when they want to report a crime or become a victim

of a crime and expect the police to find the perpetrator or recover their property

(Cheurprakobkit, 2000). For that reason, it is not surprising to see that law enforcement

agencies devote a huge portion of their resources to criminal investigations (Chaiken et

al., 1976; Horvath, et al., 2001). Despite the importance of this factor, the number of

studies conducted in this field is limited.

In July 1965, President Lyndon B. Johnson appointed the President’s Commission

on Law Enforcement and Administration of Justice to examine the crime problem and the

use of the U.S. criminal justice system and new policies to curb crime. The Commission

issued a 308-page report titled “The Challenge of Crime in a Free Society” and

8
 9

encouraged Federal, State and local governments to fund criminal justice research1.

Although it was argued that the Commission’s recommendations related with criminal

justice research did not come to fruition (DeZee, 1995), research on criminal

investigation flourished in the 1970s after the publication of the President’s Commission

report. Early studies in this field attempted to describe and assess the effectiveness of the

criminal investigation process. Although the number of studies on this topic had

decreased by the 1980s, researchers continue to study various aspects of criminal

investigation, and a body of literature has developed.

In this part of the literature review, major findings of criminal investigation

research will be presented. The goal of this review is to better understand the criminal

investigation activity and its usefulness for the crime control objective of police. In

addition, the factors leading to success in criminal investigation will be laid out to assess

how information technologies may contribute to the investigative process. This part is

divided into two sections. First, the definition of criminal investigation will be provided,

and the major goals attributed to the criminal investigation process will be discussed.

Then, empirical studies about the factors affecting the outcome of criminal investigation

will be presented.

Definition and the Goals of Criminal Investigation

Generally, criminal investigation is viewed as a ‘truth finding’ process at the end

of which the crime is solved, and offenders are caught (Maguire, 2003). This is evident in

the most common definitions of criminal investigation. In the Department of Justice’s

“Managing Criminal Investigations Manual,” criminal investigation is defined as “the

1
The Commission stated that “Federal, State and local governments should make increased funds available
for the benefits of individuals or groups with promising research program and the ability to execute them”
(President's Commission on Law Enforcement and Administration of Justice, 1967: 277).

9
 10

total police effort to: 1) collect facts leading to the identification, apprehension, and arrest

of an offender, and 2) organize these facts to present the evidence of guilt in such a way

that successful prosecution may occur” (Cawley et al., 1977, p. 1). Similarly, Osterburg

and Ward (2007) defined it as “the collection of information and evidence for identifying,

apprehending and convicting suspected offenders” (p. 5). Finally, Gilbert (2004) defined

it in a rather general tone as “a logical, objective, legal inquiry involving a possible

criminal activity” (p. 37).

These three definitions of criminal investigation point to a common process that

begins with the determination of the offense, and is then followed by the collection of

information, identification of the offender or offenders with the help of available

information, and finally the presentation of the case for the prosecution. The truth-finding

approach depicts a mechanistic view of criminal investigation. This depiction suggests

that the truth about any crime can be uncovered by applying a set of well defined

investigative activities through meticulous police work. However, the truth-finding view

of criminal investigation is criticized for being unrealistic. Empirical studies of criminal

investigation showed that instead of trying to uncover ‘truth’ by focusing on the crime

scene of each offense, detectives usually pursue a suspect-centered approach in which

they try to construct a case against the suspects known by the police. According to this

view, criminal investigation is not a truth-finding process, but an interpretive activity in

which police try to construct the truth by continuously collecting and analyzing available

information (Maguire, 2003; Tong & Bowling, 2006).

The construction of truth approach claims that crimes are social constructs, and

criminal investigation is an interpretive activity which translates a ‘social reality’ into a

 11

‘legal reality’ (Maguire, 2003). Therefore, police interpretation of the facts is critical for

criminal investigations (Ericson, 1993). Sanders (1977) argued that the decision to

investigate an allegation depends on whether police interpretation of the facts constitutes

a real crime. In order to discern real crime from the “phony cases,” detectives develop a

sense of “what really happened” about each case, which is called “establishing a case”

(Sanders, 1977). Waegel (1981, 1982) called this process “case routinization” in which

detectives categorize cases as routine versus non-routine based on the information about

the victim, offense, and possible suspects. While routine cases were perceived as non-

productive cases and receive only a superficial investigative effort, non-routine cases

were seen as likely to produce arrest and are vigorously investigated. Therefore, the

detective’s construction of truth through interpretation of the facts, and the subsequent

designation of a case as phony or real (routine or non-routine) were central elements of

each criminal investigation

An important common point in the truth-finding and construction of truth

approaches is the centrality and importance of information for the investigative work. If

criminal activity is not detected by the police, usually some kind of information about a

crime initiates the investigation (Eck, 1989; Maguire, 2003). After this occurs, the police

collect additional information to understand what has happened, and to try to identify the

offender by analyzing and interpreting that information. Finally, the police organize and

prepare information for the prosecution in a way that will increase the suspect’s chance of

conviction. Due to the important role information plays in the investigative process, Innes

(2003) defined criminal investigation as “the identification, interpretation and ordering of

information with the objective of ascertaining whether a crime has occurred, and if so,
 12

who was involved and how” (p, 113). Innes’ ‘information work’ approach combines

truth-finding and construction of truth approaches and provides a better definition of the

police investigation. Moreover, the information work approach emphasizes the

importance of information for the success of criminal investigation and implies that

technologies helping police to better process information may be an important factor for

solving crimes2.

Innes (2007) identified three main stages of criminal investigation which

emphasize the importance of the information for the investigative process. The first stage,

‘identifying and acquiring,’ involves separating relevant information from all other

available information about a case. The second stage is ‘interpreting and understanding,’

which signifies translation of information into intelligence or knowledge. At this stage,

bits of information from separate sources are fitted together to form inferences and

hypotheses about the crime. Screening the cases is also done at this stage. The third stage,

‘ordering and representing,’ involves “configuring new knowledge with extant

knowledge held by the investigator(s) in a format that enables a solution to the question

that is the focus of the investigation to be established and communicated” (Innes, 2007, p.

255). This new knowledge enables investigators to identify and apprehend the offender or

offenders, and forward the case for prosecution. Therefore, information is highly

important to achieve the goals of criminal investigation.

While there are a variety of goals discussed in the literature (Kuykendall, 1982;

McDevitt, 2005), clearance is found to be the most important goal of criminal

investigation (Chaiken et al., 1976; Eck, 1989; Greenwood, 1970; Horvath et al., 2001;

2
Actually, Innes (2003) argues that “increasingly sophisticated information management and
communication systems” expand the organizational memory and, in turn, aid criminal investigations (p.
121).
 13

McDevitt, 2005; Waegel, 1981). There are several arguments about the importance of

clearance. First of all, clearance is important for the criminal justice system because

criminal justice system generally deals with people who were arrested and charged

(Maguire, 2003). Clearance is also important for crime prevention activities because

clearing crimes can prevent future crime by deterring future offenders (Maguire, 2003).

In addition, clearance is important for justice because justice is achieved when the police

catch the right offender and provide the necessary information to the prosecution (Eck,

1989). Finally clearance is important for the citizens’ trust in the police. By clearing

crimes, police reassure people that they are fighting against crime and protecting the

society.

Factors Affecting the Outcome of Criminal Investigation

This part presents studies about the factors that are important for the outcome of

criminal investigation. The reviewed studies are organized and presented according to the

major findings in this literature.

Importance of Preliminary Investigation

Studies on criminal investigation have consistently found that the quality of

preliminary investigation3 and information collected at this stage are crucial factors for

the outcome of the investigation (Burrows, 1986; Chaiken et al., 1976; Eck, 1983; Eck,

2008; Greenberg, Elliott, Kraft, and Procter, 1977; Greenberg and Lang, 1973; Morris &

Heal, 1981; Horvath & Meesig 1996; Isaacs, 1967; Skogan, & Antunes, 1979; Wilmer,

1970). Furthermore, some studies have stated that the single most important determinant

3
Criminal investigation is generally divided into two parts: preliminary and secondary investigation.
Preliminary investigation is generally consists of the search for the information that would help
investigators to solve the case (Cawley et al., 1977). Secondary investigation, on the other hand, is police
activity meant to identify and apprehend the offender or offenders and recover the losses (McDevitt, 2005).
 14

of whether or not a case would be solved is the information the victim or witness supplied

to the responding patrol officer (Chaiken et al., 1976, Isaacs, 1967; Reis, 1971). Although

the importance of preliminary investigation and the information collected at this stage

might be the most common and least challenged finding of criminal investigation

research, it has not much altered the way that patrol services are conducted. Chaiken, et

al. (1976) study that will be described further in the next section found that patrol officers

in 58% of surveyed agencies had limited involvement in criminal investigation in the

1970s. In a more recent study, Horvath et al., (2001) analyzed investigative practices on a

larger sample and compared their result with the Chaiken, et al. (1976) study results.

They found that patrol officers still had limited involvement in criminal investigations in

more than half of the surveyed agencies (Horvath et al., 2001).

Information about the suspect, availability of the witness statement and

fingerprints collected during the preliminary investigation were identified as the most

important determinants of success in Stanford Research Institute’s (SRI) weighted

screening model4 for burglary cases (Greenberg and Lang, 1973). This model was tested

by the Police Executive Research Forum on a sample of 1,200 cases from 26 police

agencies and achieved 85 percent success in predicting the outcome of follow up

investigations (Eck, 1983). SRI researchers had developed another case screening model

for robbery cases (Greenberg, et al., 1977). Information about the suspect and forensic

4
Case screening is “a mechanism that will facilitate making a decision concerning the continuation of an
investigation based upon the existence of sufficient solvability factors obtained at the initial investigation”
(Cawley et al., 1977, p. 37). In the weighted case screening approach, the investigative unit tries to
ascertain the solvability of a crime by attaching statistically or non-statistically derived values to each
solvability factor. In the unweighted case screening approach, the investigative unit tries to ascertain the
solvability of a crime by answering critical questions for the case.
 15

examination were again identified as the most influential factors determining the success

of investigation.

The importance of information is also salient in the variation of arrest and

clearance rates among different types of crime (Burrows & Tarling, 1982; Marche, 1994;

Pucket & Lundman, 2003; Jiao, 2007; Pare, Felson & Quimet, 2007). Skogan and

Antunes (1979) further argue that quality of information very much depends on the type

of crime due to the “duration of personal contact between the parties.” As the duration of

contact increases, the victim’s likelihood of giving useful information to the police also

increases, and those crimes are more likely to get solved. This explains why burglaries

are harder for police to solve, and personal crimes particularly ones involving greater

duration of personal contact between the parties are more likely to get cleared than

property crimes.

In sum, importance of preliminary investigation and the information collected in

this stage was consistently found as the most important determinant of a successful

criminal investigation.

Organization of the Investigative Unit

Organizational reforms are regularly implemented in law enforcement agencies

especially in investigative units to increase effectiveness (Eck, 1989). The impact of

generalist-specialist investigators and centralized-decentralized investigative units on the

outcome of criminal investigation are among the topics discussed in the literature. In one

of the earliest studies, Greenwood (1970) evaluated the effectiveness of apprehension

programs and operating units of the New York Police Department (NYPD). He also

analyzed the effectiveness of “detective patrols” in that study. NYPD’s detective patrols
 16

consisted of plain clothe investigators who were patrolling areas known as dangerous to

detect and apprehend offenders. He reported that detective patrols were more successful

in making primary arrests5 than the traditional investigative units.

Greenwood’s finding was challenged in Chaiken, et al.’s (1976) influential study,

which was also known as the RAND study. The RAND study claimed that the

organization of police investigators is not related to the variations in crime, arrest, and

clearance rates; however, the RAND study and particularly its analysis of how crimes

were solved was criticized by some researchers for its methodological weaknesses. For

example, Gates & Knowles (1976) argued that the major findings of Chaiken et al.’s

(1976) study were based on unreliable data collected from only one police agency

(Kansas City Police Department, Missouri). They argued that, although authors had

acknowledged this problem, they did not do anything to overcome or alleviate it6. RAND

study authors responded to these and other critics at different venues;7 however, their

arguments did not eliminate that particular methodological concern about their study.

On the other hand, despite their conclusion on the organization of the

investigative units, Chaiken et al. (1976) also stated that “investigative strike forces” had

a potential to increase the arrest rates for hard-to-solve offenses if they were employed in

their specialization fields. The difference of investigative strike forces from traditional

investigative units was that they were organized to conduct both preliminary and

5
The author defined primary arrest as the arrest made for the following crimes: murder and non-negligent
manslaughter, forcible rape, robbery, felonious assault, burglary, grand larceny, theft of a motor vehicle,
possession of stolen property, dangerous weapons, or burglar tools, and petit larceny.
6
See Chaiken et al. (1976) for validity problems of the Kansas City Detective Case Assignment File.
7
In one article, they stated that “[o]ur conclusions and especially our policy recommendations should not
be judged alone by whether they flow inevitably and exclusively from the data collected in our study.
Instead they should be appraised in terms of whether they are within reason correct or incorrect as cast
against a full backdrop of what is known about the criminal investigation process” (Greenwood, Chaiken &
Petersilia, 1976, p. 62). They also argued that many of their findings had been previously reported and
supported by other scholars.
 17

secondary investigation, and they were organized to perform proactive instead of reactive

investigation8. This argument conflicted with their previous finding stated above because

investigative strike forces have a different organizational structure than the traditional

investigative units.

Further evidence on the impact of the organization of investigative units was

provided by Bloch and Bell (1976) and Elliot (1978). Bloch and Bell (1976) examined

the effectiveness of Coordinated Team Patrols (CTP) of the Rochester Police

Department, New York. CTP was an investigative unit, which was comprised of patrol

officers and detectives working together at a specific geographical location with special

emphasis on improving arrest and clearance rates. They found that CTP teams were more

successful in terms of making arrest and clearing crimes than non-CTP teams. Similarly,

Elliot (1978) examined the effectiveness of the Crime Control Teams (CCT) of Syracuse,

New York. CCT was a special investigative unit responsible for all the stages of an

incident occurring in its areas. Elliot (1978) found that the clearance rate was higher for

the CCT areas than the non-CCT areas. When he compared the key factors that helped to

clear more cases in CCT areas, he found that CCT detectives had achieved greater citizen

involvement; they were better in intercepting ongoing crimes; and they conducted better

investigative work than non-CCT detectives.

In conclusion, the majority of studies found that organizational structure was

effective on the outcome of investigations. Alternative organizational structures were

effectively increasing the time spent on investigations, detectives’ motivation and

citizens’ involvement.

8
Reactive investigations are “investigations of crime events after they occur because a citizen initiates the
investigation through a request;” on the other hand, “proactive investigations are police initiated,
sometimes before any crime has been committed” (Eck, 1989, p 176)
 18

Agency Size and Workload of Investigators

There are several studies that tested the impact of agency size or workload of

detectives on the outcome of police investigations. Greenwood (1970) analyzed the

impact of caseloads on the arrest rates and found inconclusive results. He divided NYPD

precincts as low and high caseload precincts and found that low case load precincts had

higher assault arrest rates than high caseload precincts, but the high caseload precincts

had higher arrest rates for robbery and burglary cases. This finding suggested that the

impact of caseload on the solution of crimes might differ according to the type of crime.

Developing a computer simulation model, Folk (1971) reported that the time allocated to

a case by detectives has a substantial effect on overall case processing in the system. On

the other hand, Chaiken et al. (1976) claimed that differences in workload have no

appreciable effect on crime, arrest or clearance rates because investigators spend more of

their time doing clerical activities such as reviewing reports, documenting files, and

interviewing suspects on cases that have less of a chance of being solved.

Ostrom, Parks and Whitaker’s (1978) study took a different approach to the issue,

and used citizen evaluations of police services and crime rates as their dependent

variables and examined them to see whether they changed with patrol density. They

found that larger departments were not performing more effectively than others.

Conversely, small to medium sized departments were performing equally to or better than

larger departments. The authors argued that the citizen-to-police ratio was the critical

factor in this finding. Since small police departments were serving in smaller

communities, they could achieve a higher citizen-to-police ratio than large police
 19

departments, which were serving in larger communities. As a result, agency size was

found to indirectly affect the level of citizen satisfaction in their study.

Later studies have tested this proposal on larger samples by employing improved

models; however, they also arrived at conflicting conclusions. Burrows & Tarling (1982)

explored the determinants of clearance with British data collected from 41 forces and

found consistent and negative impact of workload on clearance rates after controlling for

crime and socio-economic variables. On the contrary, using 76 cities’ crime and arrest

data, Liska and Chamlin (1984) found that the impact of agency size on arrest rates was

usually insignificant and varied by type of crime and race, again, after controlling for

crime and socio-economic variables9.

After deeper analysis deeper, Cordner (1989) used the Unified Crime Report’s

(UCR) crime and clearance statistics and collected demographic data from municipal

police departments and county police or sheriff departments in Maryland. With similar

conclusions as Liska and Chamlin, he found that region and proportion of property crime

variables were significantly related to the clearance, but agency size, and workload

variables were insignificant. Cordner then divided his sample into metropolitan and non-

metropolitan cities and ran two separate analyses. The agency size variable was again

insignificant in both analyses. He concluded that the regional scale and community

complexity may be more important, and environmental variables may be more influential

than organization-level variables for clearance.

In sum, the majority of studies shows that there is not a clear relationship between

agency size and investigative effectiveness. Although agency size appears as an

9
Liska & Chamlin (1984) study investigated the impact of racial/economic composition of macrosocial
units on arrest rates. Therefore, they selected 76 cities for which residential segregation rates had been
calculated since 1940, and crime and clearance data was available.
 20

important factor for citizens’ evaluation of police services, larger agencies are not more

successful than smaller agencies in solving crimes.

Forensic Evidence

Although physical evidence and forensic examination is generally expected to

increase the effectiveness of criminal investigation, it is certain that it has been

underutilized. Peterson, Mihajlovic, and Gilliland’s (1984) study found that clearance

rates were higher in cases for which physical evidence was collected and analyzed.

Moreover, their analyses showed that physical evidence increased the odds for successful

case outcome. On the other hand, Chaiken et al. (1976) reported that most police agencies

collect more physical evidence than they could productively process. Moreover they

argued that latent fingerprints rarely provide the only basis for identifying a suspect.

Horvath and Meesig’s (1996) study supported the conclusion drawn in Chaiken et al.’s

(1976) study. They claimed that police investigate only a small percentage of crimes;

physical evidence is collected in only a small percentage of cases investigated; only a

small percentage of the collected evidence actually undergoes scientific analysis; and, in

most cases physical evidence is not determinative of case outcomes.

More recently, Jones and Weatherburn’s (2004) study also found that forensic

evidence is underutilized. Similarly, Burrows, Tarling, Mackie, Poole and Hodgson

(2005)10 reported that forensic evidence contributed to only four percent of clearances in

England and Wales in 2001. They attributed the little contribution of forensic evidence to

10
Results were published before the release of the formal report: Burrows, J., & Tarling, R. (2004).
Measuring the impact of forensic science in detecting burglary and auto crime offences. Science & Justice,
44(4), 217-222.
 21

the difficulties of recovery of useful contact trace material from the crime scenes11. They

argued that forensic evidence was a contributory factor, but not a critical factor in

criminal investigations. Police training and culture were pointed as two possible causes

for the underutilization of forensic evidence. Horvath & Meesig (1996) stated that since

detectives have limited training and understanding of physical evidence and scientific

analysis, they usually focus on the human aspects of the criminal investigations like

victim statements and suspect interrogations. Bradbury and Feist (2005), on the other

hand, stated that forensic analysis is still a contributory factor in criminal investigations,

but by the introduction of automated searching of fingerprint, DNA and other forensic

databases, forensic techniques began to help to identify unknown offenders as well. They

also argued that forensic evidence made an important contribution to the identification of

hard-to-solve cases especially when other forms of evidence were absent.

In sum, although it is underutilized, forensic evidence is effective when

incorporated into criminal investigations. Currently, forensic evidence is used mainly as a

support element in criminal investigations, and as a result, makes little contribution to the

clearance of the cases. On the other hand, the extension of its role in criminal

investigations requires changes in police training and culture.

The Social Factors

As argued by the construction of truth approach to the criminal investigation,

social context and victim characteristics might have an impact on officers’ interpretation

of incidents and execution of criminal investigations. The society’s impact on criminal

investigation was generally discussed within two areas: 1) detectives’ decision to initiate

11
They calculated that examination of 100 burglary and vehicle crime scenes would yield only seven
fingerprint, 2.6 SGM plus, and 1.4 LCN identifications (SGMplus and LCN are two different DNA analysis
techniques).
 22

an investigation, and 2) outcome of the investigations. For the first part of the discussion,

Sanders (1977) and Ericson (1993) claimed that detectives’ decision to investigate was a

result of victim, case and agency related factors. Waegel’s (1981, 1982) studies provided

more insight into this debate. He revealed that detectives’ decision to investigate depends

on the information available about a case. Waegel stated that detectives were facing two

kinds of pressure within the organization: paperwork and arrest production requirements.

In order to satisfy these requirements, detectives group cases as routine and non-routine

where routine cases are non-productive cases and receive only a superficial investigative

effort, and non-routine cases are productive cases and are vigorously investigated.

Waegel (1981, 1982) argued that social characteristics of the victim and the social

context are important for detectives’ interpretation of cases as routine and non-routine.

This issue was further tested by Bynum, Cordner and Greene (1982), and partial

support was provided. Researchers analyzed the investigative activities of a Midwestern

Police Department and found that the amount of existing evidence had significant impact

on detectives’ decision to investigate for property crimes. Furthermore, the victim’s

social class was significant for burglary crimes; a conclusion only partially supported

later by Brandl (1993). Although Brandl’s (1993) quantitative analysis results reported

that victim’s social structural characteristics did not have an impact on detectives’

decision to investigate, qualitative observations suggested that they were influential.

Based on observations, he argued that detectives’ decisions were affected by the victim

characteristics. Finally, Klinger (1997) argued that social characteristics of department’s

jurisdiction have an impact on officers’ vigor with respect to investigating crimes. In

 23

conclusion, available evidence indicates that social characteristics are influential on

detectives’ decision to investigate.

For the second part of the discussion, economic and demographic variables were

tested as the predictors of the outcome of criminal investigation. For example, Swanson

(1978) found that environmental variables (income inequality, percent nonwhite, males

aged 15-19) have a greater impact on the arrest rates than the police organizational

variables (centralization, specialization, professionalism, community relations

orientation, task orientation). Later studies on this debate also generally found supportive

evidence. Burrows & Tarling (1982) found that population and the proportion of the 15-

24 age group with respect to the total population had a negative impact on clearance.

Liska and Chamlin (1984) also found supportive evidence showing that the

racial/economic composition of cities substantially, and economic inequality moderately

affects arrest rates for violent crimes and property crimes. Pucket and Lundman (2003),

on the other hand, found that socioeconomic composition was insignificant, but race

composition was a significant predictor of homicide clearance. Pare et al. (2007) found

that community size, poverty, and types of crimes were significant predictors of

clearance. Therefore, the existing literature generally supports the argument that decision

to investigate a crime and the outcome of a criminal investigation is affected by victim

characteristics and the social context of department’s jurisdiction.

In consequence, review of criminal investigation literature showed that

preliminary investigation and the information collected in this stage were the most

important determinants of outcome of criminal investigations. This finding emphasizes

the importance of information for the criminal investigations. Social context of

 24

department’s jurisdiction, victim characteristics and the organization of investigative unit

were also found influential on the outcome of investigations. There were not sufficient

amount of studies about the impact of agency size and workload of investigators on

criminal investigation; however, existing studies generally report that those variables

were not influential on the outcome of investigations. Finally, forensic analysis was

generally found influential on the outcome of investigations, but since it was

underutilized, its impact was very small

Police and Information Technologies

This part describes development, use and impact of IT in U.S. law enforcement

agencies. It begins with the definitions of some important terms and the historical

development of IT use by police agencies in the U.S. Then, the types of information

technologies used by the police and IT implementation problems in law enforcement

agencies are described. Finally, studies on the impact of information technologies on

public domain, particularly on law enforcement agencies, are reviewed. Since the number

of studies about the impact of information technologies on law enforcement organizations

is low, a review of major studies conducted on other public agencies enriches our

understanding of information technologies.

Definitions

Definitions for information, information technology and information systems are

given in this section. Information has been defined in many ways. These definitions

include: facts organized to describe a situation or condition (Wiig, 1993); data with

relevance and purpose (Davenport & Prusak, 1997); anything that reduces uncertainty

about the next part of the message (Manning, 1992); the value attached or instantiated to
 25

a characteristic or variable returned by a function or produced by a process (Losee, 1997).

Within the police context, Innes defined information as the “meaningful data of potential

relevance to the investigative activities of the police, either related to a specific crime

being investigated, or some other matter in which the police may have an interest” (Innes,

2003, p. 113).

Since the information concept has a broad meaning, the domain of information

technology extends to almost every corner of life. WordNet dictionary defines

information technology as “the branch of engineering that deals with the use of

computers and telecommunications to retrieve and store and transmit information”

(WordNet). Similarly, Turk (2000) defined it as any technology that is used to deliver

data, information, and knowledge. Finally, Hogan and Radack (1997) offered a more

useful definition for this study. They defined it as the blend of hardware and software

used for storage, processing, transfer, display, management, organization, and retrieval of

information.

No single piece of information technology can satisfy the complex information

needs of contemporary organizations by itself. Therefore, different types of information

technologies interconnect and form information systems. O’Brian (2002, p. G-10) defines

an information system as “a system that accepts data resources as inputs and processes

them into information products as output.” Although using any information technology is

not a necessary condition for an information system to function, information technology

and information systems are usually associated with each other. As a matter of fact,

information technology has become a very important component of modern information

systems, and it is highly critical for the success of an organization. Information

 26

technology and information system concepts are so interconnected with each other that it

is not possible to completely separate them. As a result, this study views information

technology within the information system context as a part of an information system, not

as a separate entity.

Development of Law Enforcement Information Technologies in the U.S.

Although extensive use of technology in the police organizations did not start

until the 1970s (Brown, 2000), efforts for developing systematic information about the

crime began as early as the 1800’s (Seaskate, 1998). In 1834, Massachusetts became the

first state to collect data on crimes, and in 1850, the Federal Government began to collect

crime data in conjunction with the census. By the early 1900s, criminal statistical reports

were prepared by compiling police reports (Dunworth, 2000). These initiatives were the

initial steps in developing contemporary criminal justice information systems.

The early reformers of the American policing system tried to separate the police

from the politics, and the professional or traditional policing era began by the 1920s and

continued until the 1970s (Seaskate, 1998). August Vollmer, Bruce Smith, Harry

Fosdick, and Orlando Wilson were among the influential figures, and the use of the

polygraph machine and fingerprint and handwriting classification systems began in this

period (Manning, 1992). Also in this era, police adoption of the automobile and two-way

radios occurred, and the FBI established its own forensic laboratory in 1932 (Seaskate,

1998). In 1929, the International Association of Chiefs of Police (IACP) implemented the

Uniform Crime Reporting (UCR) system, transferring the system to the FBI in 1930.

President Hoover launched the National Commission on Law Observance and

Enforcement, also known as the Wickersham Commission, in 1929. As the first national
 27

initiative to evaluate the U.S. system of justice administration, the Commission

recommended the collection of accurate nationwide statistics on crime and the criminal

justice system (Dunworth, 2000; 2004).

In 1965, Lyndon B. Johnson appointed the President’s Commission on Law

Enforcement and Administration of Justice to examine the crime problem. The

Commission offered more than 200 recommendations, and 13 of them were about the

development of technologies related to police operations12. In order to be able to develop

and use effective criminal justice technologies, huge amounts of federal subsidies were

provided to local and state law enforcement agencies. The Law Enforcement Assistance

Administration (LEAA) was established by the Omnibus Crime Control and Safe Streets

Act of 1968 to manage the federal funds given to the criminal justice system (Dunworth,

2000; 2004). LEAA sought to spur the computerization of policing. This policy change

had been coupled with advances in information technologies. Beginning from the 1960s

through the 1990s, information technologies became more effective, efficient and

affordable. Moreover, software manufacturers began to develop many useful programs

for business and government use13.

The professional model of policing has generally been characterized by its narrow

focus on traditional law enforcement and crime control policies (Greene, 2000). In the

1970s, police agencies began to challenge this criticism by encouraging the community’s

involvement in crime prevention activities (Seaskate, 1998). Computerization of law

enforcement continued in this era and during this period, the National Crime Information

12
The Commission stated that “scientific and technological revolution that has so radically changed most
of the American society during the past few decades has had significantly little impact on the criminal
justice system” (President's Commission on Law Enforcement and Administration of Justice, 1967, p. 245).
13
See Brown (2000) for further discussion on the diffusion of technological innovations into the criminal
justice system
 28

Center (NCIC), 911 system and Automated Fingerprint Identification System (AFIS)

were developed (Seaskate, 1998)14. Established in 1982 by the FBI, NCIC aimed to

provide criminal justice information such as criminal record history information, and

information about fugitives, stolen properties, and missing persons to the criminal justice

agencies. This development further encouraged the adoption of computer technology

(Davis & Jackson, 2004). In the same year, the FBI initiated the National Incident Based

Reporting System (NIBRS) project to improve its crime reporting system (UCR) and

began accepting NIBRS data in January 1989. NIBRS provides very detailed incident-

based information about crimes; however, implementation of NIBRS has been a slow

process mainly because of funding problems and lack of motivation of implementing

agencies15. As of 2003, 5,271 agencies in 23 states were certified as NIBRS participating

agencies (Bureau of Justice Statistics, 2004)

With the dissolution of LEAA, the National Institute of Justice (NIJ) became the

research and development arm of the criminal justice system. NIJ pursues its

technological research and development goal through the Office of Science and

Technology (OST) and touts the development of lightweight body armor and DNA

analysis among its important achievements (Dunworth, 2004).

Law enforcement agencies received funding from different federal and local

agencies for their information technology needs, and those grant programs usually

required them to focus on specific programmatic objectives. As a result, a “patchwork” of

various computer systems that “are applicable only for very specific purposes, are unable

to share information with other justice agencies, and serve only individual components of

14
See Klug, Peterson, & Stoney (1992) for more information on development, implementation and impacts
of AFIS.
15
See Faggiani & Hirschel (2004) for an extended discussion of the issue
 29

state and local governments” were implemented throughout the U.S. (Davis & Jackson,

2004, p. 29). For this reason, integration of criminal justice information systems became

a priority for the criminal justice community. In 1998, the Crime Identification

Technology Act (CITA) which provided funding for five years for integration of criminal

justice information systems was enacted. However, integration problems were cited as

major factors leading to the 9/11 terrorist attacks and even today continue to be critical

problems (National Commission on Terrorist Attacks, 2004). The variety and complexity

of police information systems are important factors impeding the research and integration

of criminal justice information systems; therefore, it will be useful to explore the types of

information systems used by law enforcement agencies.

Types of Law Enforcement Information Technologies

There are numerous hardware and software systems developed for law

enforcement agencies, and a concise overview of some important systems will be

provided in this section. However, prior to this, it is useful to look at how researchers

have grouped law enforcement information technologies in their studies. There are

several law enforcement IT system classifications offered in the criminal justice

literature16. For example, Colton (1972) grouped police information technologies around

four main functional areas: Police patrol and daily operations, investigative analysis,

administration management, and program/planning and evaluation. Ackroyd, Harper and

Hughes (1992) grouped investigative information technologies under the management

category, which consists of command and control, crime reporting, management of

16
Similarly, police technologies are also discussed in the literature and several classifications are offered.
See Nogala (1995), Nunn (2001), and Manning (2003).
 30

information, and duty state systems17. Similarly, Stevens (1995) defined three kinds of

law enforcement information systems: data retrieval systems, analytical systems, and

process control systems. Manning (1992) offered a different categorization as primary,

secondary and tertiary information technologies. Primary technologies mainly consist of

communication systems and more related with the preliminary investigation stage;

secondary information technologies are used to support ongoing cases and management

functions and more related with the secondary information stage; and tertiary information

technologies consist of analytic systems. Although tertiary information technologies are

also related with the secondary investigation stage, they are mainly used for analysis,

evaluation and strategic planning. Finally, Dunworth (2005) offered a more updated

classification and presented law enforcement information technologies as crime analysis,

records management systems, offender histories and offender identification, mobile data

terminals, and forensic technology.

The examination of these different classifications revealed that investigative use

of information technologies can be grouped according to following four areas:

communications and remote data access, records management, crime analysis, and

forensic analysis. Table 1 shows these areas and organizes other classifications according

to these groups. For example Colton’s police patrol and daily operations category

involves communication and mobile computing and records management groups

combined. The investigative analysis group, on the other hand, involves both crime

analysis and forensic analysis groups. Table 1 shows that other classifications also

generally fit to this new classification scheme.

17
Duty state systems are generally used to keep records of status information of personnel such as work
hours, location information, activities, etc.
 31

Table 1
Classification of Police Information Technologies
Ackroyd et Dunworth
Areas Colton (1972) al. (1992) Stevens (1995) Manning (1992) (2005)

Communication Police patrol Primary

Command Process control Mobile data
and Mobile and daily information
and control systems terminals -
Computing operations technology
(patrol Records
activities,
management,
operations, Secondary
Records Crime Data retrieval criminal
traffic information
Management reporting systems histories and
incidents, technology
offender
CAD, etc)
identification
Investigative Management Tertiary Crime
Analytical
Crime Analysis analysis of information analysis,
systems
(Criminal information technology UCR, NIBRS
histories,
modus
operandi (Secondary
Forensic Forensic
system, information)
Analysis technology
fingerprint technologies)
systems)

However, it must be noted that modern law enforcement information systems are

usually integrated with each other to increase the information sharing capability and

effectiveness of police operations18. For example, a 911 system can communicate with

the CAD system which enables the call center officer to assign the appropriate patrol

officer to the call. All the information produced by the 911 and the CAD systems can be

transferred to the records management system for further analysis of investigators or for

archiving purposes. Moreover, local, state or federal level records management systems

can interact with crime analysis applications. Therefore, it should be taken into account

that usually there are not solid borders between categories, and some technologies can be

18
Weller et al. (2001) argues that the goals of current IS projects are 1) developing inter-agency
collaboration as a way of doing business, 2) developing a capacity to create policy based on data and
information, 3) creating system-wide approaches to solving problems, and 4) involving the community in
the criminal justice policy (Weller et al., 2001).
 32

recorded within more than one category19. A brief explanation of those four areas would

be helpful in understanding the use of information technologies in law enforcement

agencies

Communication and Mobile Computing

Communication and mobile computing technologies are used for wireless receipt

and transmission of information to and from officers on foot or in patrol cars (Dunworth,

2004). Police departments generally use laptop computers and different handheld devices

such as cell phone and personal digital assistant (PDA) for mobile access. Mobile

computing is mainly employed to query databases, to communicate with the CAD system

and peers, to prepare and file incident reports and to get information on job related topics.

One of the important applications used with mobile computing is automated field

reporting (AFR). The AFR application enables officers to prepare incident reports at the

scene, to electronically submit these to a supervisor for approval and then to submit them

to the RMS. AFR applications have useful features such as drop-down menus, spell-

checking, prefilled fields, pre-population of multiple forms and error correction to

maximize integrity of the information stored in the RMS (Harris & Romesburg, 2002).

As new technologies are developed, mobile computing becomes more efficient, and more

departments are benefiting from it (Dunworth, 2005).

Records Management

Records Management System (RMS) is probably the most important part of a law

enforcement information system. RMS is mainly used for storage, retrieval and

19
An insightful presentation of system integration in law enforcement information systems is provided by
Dean and Gottschalk (2007). In their matrix of technological and policing functions, they visually show
that some law enforcement IT systems are designed to perform more than one policing or technological
functions together.
 33

processing of information about every aspect of police business (Dunworth, 2005). A

typical RMS stores and processes information regarding incidents, persons, vehicles,

locations, arrests/bookings, traffic related incidents, property, evidence and many other

things. Moreover, RMS provides access to a wide range of external RMS applications,

which stores different kinds of information like offender histories and identifications.

Major RMS applications that can be used as an extension to agency records are as

follows:

1. National Crime Information Center (NCIC). The NCIC system houses a wide

range of criminal justice information including criminal record history, fugitives, stolen

properties and missing persons information. NCIC data are compiled by the FBI, federal,

state, local and foreign criminal justice agencies, and authorized courts (FBI, 2009c).

2. Violent Criminal Apprehension Program (ViCAP). ViCAP keeps information

related to the violent crimes; specifically homicides, sexual assaults, missing persons, and

unidentified human remains (FBI, 2009b). All the information submitted to ViCAP

system is compared with stored information, and similar cases are reported to the users.

3. Regional Information Sharing System (RISS). RISS is a federally funded

program providing support to law enforcement agencies. Its main purpose is to combat

illegal drug trafficking, identity theft, human trafficking, violent crime, terrorist activity,

and to promote officer safety (BJA, 2007). RISS is comprised of six multi-state centers

communicating on a secure network known as RISSNET.

4. National Incident Based Reporting System (NIBRS). As described above,

NIBRS is developed as a complement to the Uniform Crime Reporting (UCR) system’s

 34

shortcomings. It is an incident-based reporting system for crimes known to the police and

provides detailed information about offenses (FBI, 2009c).

In addition to these four systems, there are other RMS applications such as the

Interstate Identification Index (III) and the National Instant Criminal Background Check

System (NICS). “Modern RMS applications are focused on improving data accuracy and

speedy retrieval of information” (Harris & Romesburg, 2002, p. 253). Therefore, they are

integrated with other systems and allow them automatically to add new data to the system

and to query databases for administrative and operational tasks.

Crime Analysis

Crime analysis is “the systematic study of crime and disorder problems as well as

police related issues … to assist the police in criminal apprehension, crime and disorder

reduction, crime prevention and evaluation” (Boba, 2005: p. 6). Crime analysis consists

of the following three essential functions:

1) Assess the nature, extent, and distribution of crime in order to efficiently and

effectively allocate and deploy resources. 2) Identify crime suspect correlations to

assist investigations. 3) Identify conditions that facilitate crime and incivility so

that policymakers may make informed decisions about prevention approaches

(O’Shea and Nicholls, 2003, p. 8).

The crime analysis process consists of the collection, collation and analysis of

data, dissemination of results and finally incorporation of feedback from the users into

the analysis (Boba, 2005). Since information technologies have great potential to aid all

of these steps, there are many different IT applications that are used for crime analysis
 35

purposes. Two important IT systems that are closely related with crime analysis are as

follows:

1. Geographic Information Systems (GIS) and Crime Mapping. GIS is “a system

designed to store, manipulate, analyze and output map-based or spatial information” for

many purposes (Steinberg, 2005, p.7). GIS store information about places in the forms of

“layers”. This structure enables a GIS system to present any kind of information about a

location on a map. Therefore, GIS users can view either discrete levels of information for

a specific analysis or many levels of information together to get a detailed picture of the

location (Harris & Romesburg, 2002). In crime analysis, GIS is mainly used to produce

crime maps by merging information from crime and geographical databases. Crime

mapping applications synthesize crime related information from different sources and

visualize them on a map in a way that is easier to understand the relationships and trends

(Chu, 2001). Moreover, integrated with Global Positioning Systems (GPS) and the

Automatic Vehicle Location system (AVL), GIS can be used in deploying units and

analyzing patrol activities.

2. Computer-Aided Dispatch (CAD) System. The CAD System is generally the

first point of data entry “designed to handle all information related to receiving and

dispatching emergency calls for service” (Harris & Romesburg, 2002, p. 248). Although

McEwen (2002) argued that CAD systems were designed mainly for improving

operational efficiency and achieving crime control through arrest, modern CAD systems

are an important part of crime analysis. Integrated with the E-911 (Enhanced 911)

service, CAD systems can store incident location history and related information such as

warrants, number and type of prior calls to the location. Moreover, combined with a
 36

geographic database, CAD systems can pinpoint coordinates of a location on a map and

store geospatial information suitable for further analysis (Harris & Romesburg, 2002). In

addition, CAD Systems can be integrated with other systems. For example, with suitable

message switch configurations, local, state and federal databases can be automatically

queried both at the CAD center and from patrol cars. CAD systems can also be integrated

with Internet so that users can get e-mails from the system and connect to the system

from anywhere (Chu, 2001).

In addition to these systems, there are plenty of crime analysis software

applications available to support criminal investigations. These applications usually help

police to conduct link analysis, network analysis, event sequence analysis and transaction

pattern analysis (Adderley & Musgrove, 2001)

Forensic Analysis

As discussed in the first part of the literature review, forensic analysis has a small

impact on criminal investigations mainly because forensic evidence is not collected in

majority of cases and most of the collected evidence is not even analyzed (Burrows et al.,

2005; Chaiken et al., 1976; Weatherburn, 2004). Information technologies are mainly

used for information sharing purposes in forensic analysis. There are several information

sharing networks used by law enforcement agencies nationwide:

1. Integrated Automated Fingerprint Identification System (IAFIS). IAFIS is a

national fingerprint and criminal history system. It provides automated fingerprint search

capabilities, latent searching capability, electronic image storage, and electronic exchange

of fingerprints to the law enforcement agencies (FBI, 2008b). With the developments in
 37

imaging technologies and mobile computing systems, IAFIS inquiries can be made by

patrol officers on the scene.

2. National Integrated Ballistic Information Network (NIBIN). NIBIN is a

database of fired cartridge casing and bullet images. It allows partnering agencies to enter

digital images into the system and to correlate them against earlier entries via electronic

image comparison (ATF, 2009).

3. Combined DNA Index System (CODIS). CODIS stores DNA profiles of

convicted felons and unidentified DNA found in crime scenes (FBI, 2009a). It allows

forensic laboratories to add new records and compare them with the stored information.

In sum, there are many kinds of information technologies and different

information systems available for law enforcement agencies. Although most of the law

enforcement agencies are benefiting from those technologies (Davis & Jackson, 2004), it

is not possible to say that all the benefits of information technologies are realized.

Implementation problems still remain as the most important impediments, and they

prevent law enforcement agencies from taking advantage of at least some of the

opportunities provided by the information technologies. For that reason, it will be useful

to outline implementation issues in the law enforcement context.

Implementation Issues

In 2005, the FBI’s Virtual Case File project was cancelled with a loss of more

than $100 million. The U.S. Department of Justice Report pointed to the following

problems: “poorly defined and slowly evolving design requirements; overly ambitious

schedules; and the lack of a plan to guide hardware purchases, network deployments, and

software development for the bureau” (Goldstein, 2005, p. 1). The Virtual Case File
 38

project is the most recent significant example of information technology implementation

failures. Information technology implementation is difficult particularly for public

organizations because IT projects are 1) big projects, 2) they often cross the line of

authority, 3) they deal with technology and aligning technology to the agency’s needs and

4) they are time-sensitive (Harris & Romesburg, 2002). For these reasons, there are

several guides that are prepared and disseminated by different agencies to help law

enforcement agencies in designing and implementing information systems (e.g. Harris &

Romesburg, 2002; Institute of Policy Research, 2004; Weller, Martin, Price &

Wagenknecht-Ivey, 2001)

Researchers have generally pointed to the organizational structure of police

agencies as the root cause of implementation failures. In an early study, Colton (1979)

stated that complexities concerning the implementation of technology and the context and

nature of police work should have been considered while planning information

technologies for police agencies. Similarly, Ackroyd et al., (1992) stated that IT has

many potential benefits to police work; however, the internal and external characteristics

of police organizations hinder the successful implementation of IT systems. The

requirements of a successful IT implementation usually contradict work practices and the

organizational context of the police force. Therefore, Ackroyd et al., (1992) argued that

IT may make important contributions to police work, but IT innovation by itself is not a

solution to the policing problems.

Along the same line, Manning (2001) argued that because of the different context

of police organizations, technological solutions cannot be implemented as they are

implemented in other organizations. The social context, organizational structure and

 39

historical development of police organizations have an important impact on the success

of IT systems. He claimed that law enforcement technology can be most useful only for

crime analysis purposes. Manning argued that IT implementation in law enforcement

agencies should be accompanied by organizational change in order to be successful. This

point is reiterated in the Police Department Information Systems Technology

Enhancement Project (ISTEP) report which argued that community and problem oriented

policing require several fundamental changes in police agencies, and information systems

should be designed to support the new framework (Abt Associates, 2001).

However, organizational change is not easy, and IT implementation studies

reported that it is one of the biggest obstacles for law enforcement agencies. LeBeuf

(2001) argued that the investigation community “accepts the required changes at a snail’s

pace” (p. 19). He stated that information technologies that were completely new to the

police usually require an adjustment period, but this period has been seen as a loss of job

efficiency by some managers. As a result, sometimes officers were forced to adopt new

technologies without adequate training, which causes failure (LeBeuf, 2001). Similarly,

studies by SEARCH (1997) and Brown, McCabe and Wellford (2007) showed

stakeholder resistance to change, education and training as important challenges to the IT

implementation in law enforcement agencies.

Consequently, IT implementation is problematic for organizations, and it is even

more problematic for law enforcement organizations because of their formal structure

and organizational culture.

 40

Impact of Information Systems

This section reviews the literature on the impact of information technology. It

begins with a discussion of information system evaluation and DeLone and McLean’s

(2003) model. Subsequently, empirical studies about information technologies’ impact on

public and law enforcement organizations are presented. This section will end with a

discussion of Compstat.

A Model for Evaluation of Information System Success

Despite the huge amounts of money spent on IT each year20, its contribution to

organizational goals is not clear. As Nobel Laureate Robert Solow stated, computers are

everywhere except in the productivity statistics (as cited in Brynjolfsson, 1993). This

disappointing conclusion was called “productivity paradox.” In the early days of

information technologies, improving the efficiency of information processing was the

primary target of IT investments, and evaluation of information technologies was

relatively easy. Through the rapid development of information systems, information

technologies became a “competitive advantage”21 for particularly the private

organizations, and the role of information technologies “has been changed from one of

support to one of strategic importance” (Ballantine, Galliers & Straw, 1996). Meanwhile,

the nature of the benefits accrued from IT investments changed and became more

intangible. For this reason, Brynjolfsson, (1993) defined mismeasurement of inputs and

20
For example, according to the 2009 Federal Budget plans, the Department of Homeland Security would
get $400 million to protect the critical infrastructure and IT networks from hackers, $39 million in new
spending to standardize IT acquisitions and "streamline maintenance and support contracts," and $71
million in new spending for Immigrations and Customs Enforcement IT, including detainee tracking, case
management, data warehousing, and data center consolidation (Hoover, 2009).
21
A competitive advantage exists when the firm is able to deliver the same benefits as competitors but at a
lower cost (cost advantage), or deliver benefits that exceed those of competing products (differentiation
advantage). Thus, a competitive advantage enables the firm to create a superior value for its customers and
superior profits for itself (Porter, 1998).
 41

outputs as the most important cause of the productivity paradox22. The measurement

problem is particularly acute for the service sector and among white collar workers

whose input and output is harder to measure.

As a result, many different evaluation models for information technologies were

offered in the literature23. DeLone and McLean (1992) conducted an extensive review of

IT evaluation literature and developed a new model. They argued that the variety of IT

evaluation approaches can be seen as a result of the range of information system success

measures. This variety is understandable when we think that information is the most basic

output of an information system. DeLone and McLean (1992) argue that information can

be measured at different levels including the technical level, the semantic level and the

effectiveness level. Success at the technical level is measured by analyzing the accuracy

and efficiency of the system that produces the information. Success at the semantic level

is measured by investigating whether the information conveys the intended meaning.

Finally, success at the effectiveness level is measured by analyzing the effect of the

information on the receiver. Therefore, the variety of IT evaluation approaches and IS

success measures is a manifestation of the characteristics of the information. By taking

this into consideration, DeLone and McLean (1992) developed their causal model to

measure IS success (Figure 1)24. Since the DeLone and McLean IS success model will be

used in this study, an explanation of the model will be appropriate to better understand

this study and also IT evaluation studies in general.

22
Other explanations for the productivity paradox offered by Brynjolfsson (1993) are lags due to learning
and adjustment, redistribution and dissipation of benefits, and mismanagement of information and
technology. Brynjolfsson (1998) also argued that organizational change is essential for information
technology to be effective.
23
See Rocheleau (2005), Garson (2006) and Gunasekaran, Ngai, & McGaughey (2008) for a detailed
discussion of IT evaluation and major IT evaluation models.
24
DeLone and McLean proposed their original model in their 1992 study and updated it in their 2003
study. Figure 1 presents the updated model.
 42

INFORMATION
QUALITY

INTENTION USE
TO USE
SYSTEM QUALITY NET
BENEFITS
USER
SATISFACTION
SERVICE
QUALITY

Figure 1. DeLone & McLean information system success model

As shown in Figure 1, there are six categories of IS success measures according to

DeLone and McLean. The system quality and the service quality variables measure

information system success at the technical level. System quality variables are

engineering-oriented performance characteristics of the system. Adaptability, availability,

usability, reliability and response time are commonly used system quality measures. The

service quality, which was added to the model later, is measured by the assurance,

empathy and responsiveness of the service unit. The information quality variables

measure the information system success at the semantic level. Completeness, ease of

understanding, personalization, relevance and security are used to measure information

quality. These characteristics of the information system are usually measured from the

perspective of the user; as a result, they are fairly subjective.

The remaining three groups of variables measure success at the effectiveness

level. The use of information system is straightforward, and it measures the nature of use,

navigation patterns, number of visits, and number of transactions executed. Intention to

use was added to the model later because use of some systems might be mandatory. User
 43

satisfaction is the most widely used measure for information system success. It is usually

measured by user surveys. Numbers of repeat purchases or repeat visits were also used

for measuring e-commerce systems’ success. Finally, net benefits can be measured by

many outcome variables such as cost savings, expanded markets, incremental additional

sales, reduced search costs, and time savings. DeLone and McLean (2003: 19) argued

that “[t]he choice of where the impacts should be measured will depend on the system or

systems being evaluated and their purposes.” Therefore, not to complicate their model

with more success measures, they preferred to group all impact measures into a single

category called net benefits. According to the model, information system success can be

measured in all these six levels. Furthermore, there is a relationship between these six

success levels. For example, information quality affects the use of the system and user

satisfaction where use and user satisfaction affect the net benefit accrued from the

information system.

Since DeLone and McLean’s information system success model is a complex

model, researchers have generally tested parts of the model. For example, Seddon and

Kiew’s (1996) and Roldan and Millan’s (2000) studies tested the system quality and user

satisfaction relationship and found supportive results. Almutairi & Subramanian (2005)

tested system quality, information quality, system usage and user satisfaction and

reported supportive results. Particularly, they found a significant relationship between

system use and net benefit, which is measured as the effect of the information system on

users’ performance. Finally, Iivari (2005) was also tested system quality, information

quality, system usage, user satisfaction and individual impact and reported partial support
 44

for the model. Iivari (2005) also found significant relationship between system use and

net benefit relationship.

Information System Impact at the Technical and Semantic Level

Technical and semantic measurement levels consist of measures related to system

quality, service quality and information quality25. The PMIS studies generally report a

positive contribution of information technologies to system and information quality

(Danziger & Andersen, 2002). For example, Brown (1996) analyzed the hurdles to GIS

success by conducting a survey study on a nonrandomized sample of local governments

and found generally positive results at the technical level. Survey respondents were

satisfied with the accuracy and timeliness, but somewhat dissatisfied about the

availability of the data. To assess IS success at the semantic level, Kraemer, Danziger,

Dunkle and King (1993) analyzed data gathered by conducting surveys and interviews at

public agencies in 42 U.S. cities. They reported that generally users and managers found

computer-based information useful and important for their tasks.

There are several studies investigating IS success at the technical and semantic

level in law enforcement agencies. As presented below they report mixed findings about

success at those levels.

Technical Level.

O’Shea (1996) and Brown (2001) analyzed law enforcement IS success at the

technical level and arrived at contrasting conclusions. O’Shea (1996) conducted

interviews and focus group sessions with Chicago Police Department officers to explore

the information processing operations of the agency. His study generally depicted a

25
No studies were found about IT service quality in public organizations including law enforcement
agencies.
 45

negative picture of IS success. Participants of focus group sessions complained that

information was not accessible due to technical problems. Moreover, they contended that

they could not get timely information because of red tape. Brown (2001) evaluated

Charlotte-Mecklenburg Police Department’s Knowledge-Based Community Oriented

Problem Solving System by using survey and focus group methods before and after the

implementation of the system. The pretest group stated that crime-related information

was not readily available, and available information was not sufficient for the job.

Posttest analysis, on the other hand, showed that the accessibility, accuracy and

timeliness of the information increased after the implementation of the system.

Finally, Tien, Cahn, Neray, Einstein and Pei (2005) developed the Record Quality

Index (RQI) to measure the performance of criminal history records systems. RQI was a

function of 1) a set of outcome measures, 2) a timeliness measure, and 3) a completeness

measure. They conducted an analysis on national data by using RQI and found that the

state RQIs increased over time. In other words, data quality of criminal history records

systems have increased over time (Tien et al., 2005).

Semantic Level.

O’Shea (1996) also reported negative findings about IS success at the semantic

level. She argued that IS was potentially useful, but the collected information was

underutilized because officers could get desired information easily by scanning the hard

copies of case reports instead of querying the system. Therefore, according to O’Shea,

problems at the technical level were shading the success at the semantic level. Similarly,

LeBeuf (2001) argued that collected information was usually accurate, but quality of
 46

information was in question. He argued that most of the information collected by the

police was stored in databases without being used in any of the analysis.

In another study, Northrop, Kraemer & King (1995) investigated the use of

computerized search capability by officers, its value to the police and factors that affect

computer use by the police. They found that officers generally perceive computer

generated information as important for their jobs. Finally, Paulsen (2004) investigated

whether maps generated by a GIS system improve officers’ perception and understanding

of crime patterns and problem areas within a jurisdiction. Paulsen randomly assigned 40

officers into control and experiment groups and surveyed them before and after the

treatment. Paulsen (2004) found no statistically significant difference between control

and experimental groups’ understanding of crime patterns within their jurisdictions.

However, lack of training on the crime maps and the short time period of the experiment

should be considered when interpreting the results.

In conclusion, available technical and semantic level IS impact studies have

usually reported mixed results in terms of system and information quality.

Information System Impact at the Effectiveness Level

PMIS studies reported mixed findings about information systems’ success at the

effectiveness level. Several studies evaluated information system success by analyzing

economic measures. Henman’s (1996) analysis of computerization of Australia’s

Department of Social Security showed that efficiency did not increase with the increasing

computerization; moreover, the client-to-staff ratio decreased and administrative costs

increased. Lehr and Lichtenberg (1998) and Lee and Perry (2002), on the other hand,

used the Cobb-Douglas production function and found that computers increased the
 47

economic performance of the government. Heintze and Bretschneider (2000) found that

IT adoption had a positive and significant effect on local government performance.

Contrary to foregoing findings, Peled (2001) argued that information systems bring

inefficiency by increasing red tape in the organization.

In the law enforcement context, effectiveness studies generally focused on the net

benefit aspect of the question. There are only a few studies available about the use of law

enforcement systems. Northrop, Kraemer & King (1995) conducted a longitudinal study

of the use of computerized search capability by detectives and found that police use had

increased since 1976. They also reported that both the characteristics of innovation (user

friendliness of the systems and training) and characteristics of the user (computer literacy

and prior computer experience) had associations with officers’ use of computers. As

presented above, O’Shea (1996) argued that Chicago Police Department officers prefer to

use hard copies instead of querying the information system; however, DuBois, Skogan,

Hartnett, Bump and Morris (2007) study’s findings contradicted O’Shea’s (1996) study.

They conducted a survey study about the use of the Chicago Police Department’s Citizen

and Law Enforcement Analysis and Reporting (CLEAR) system. They stated that

analyzing use patterns is not meaningful because use of most of the system modules is

mandatory. In other words, officers must use the system to be able to do particular jobs.

For that reason, they analyzed the use patterns of the data warehouse for which use is

voluntary. They found that the data warehouse was used by 92% of the officers (DuBois

et al., 2007).

The net benefit is usually measured in some form of system or officer

performance such as usefulness of output of the system, arrest productivity, clearance

 48

productivity, etc. A review of effectiveness studies showed that the majority of these are

about communication and mobile computing technologies. There are only a few studies

on crime analysis and forensic technologies. Although there is not any study that

particularly tackles the records management systems, crime analysis studies can be

extended to that area because they are generally integrated together. Therefore, these

studies will be presented within two groups. In addition to them, there are several studies

that analyzed information technologies’ impact on the organizational structure of law

enforcement agencies. Those studies will be presented as the third group.

Crime analysis and forensic analysis.

In an early study, Zavala and Mullen (1970) found that modus operandi systems

can be effective in identifying offenders. Danziger and Kraemer (1985) examined the

effects of computing on the performance of police detectives and found that more than

80% of detectives experienced information benefits from computing, and nearly two

thirds of detectives indicated that computers assisted them in some of their arrests and

clearances. Hauck and Chen (1999) evaluated the performance of the Coplink Concept

Space application, which helps officers to uncover relationships between different types

of information. They conducted a field experiment and reported that the Coplink Concept

Space application is highly useful for investigative purposes. The Coplink project was

evaluated also by the Chen et al. (2003) study, and supportive results were reported

again. Wellford and Cronin (1999) examined factors affecting homicide clearance rates

by using UCR statistics and a separate survey instrument. They analyzed the more than

two hundred independent variables in their study. They found that officers’ use of

information systems in their daily job has relationship with homicide clearances.
 49

In a recent study, Braga and Pierce (2004) analyzed the impact of the Integrated

Ballistic Identification System (IBIS) in the Boston Police Department. Their analysis

showed that the IBIS system significantly improved the productivity of the Boston Police

Department’s Ballistic Unit, and it was associated with a six fold increase in the monthly

number of ballistic matches. In sum, available studies in this group generally reported the

positive net benefits of information technologies on investigative activities.

Communications and remote data access.

Police command and control systems are generally found effective. In a series of

studies, Colton (1972; 1980) reported that the rapid retrieval system increased the arrest

rates of the Long Beach Police Department, and the CAD system increased the response

time of both telephone operators and patrol officers in the San Diego Police Department.

Colton, Brandeau, and Tien (1983) stated that the police command control and

communication systems of a sample of cities were generally found successful in

performing different patrol functions. Morckel (2002) and Mayer (2009) reported

improvements in investigative operations after implementing GPS technology.

Similarly to command and control systems, mobile access systems are generally

reported as effective in clearing crimes. Nunn (1994) examined the impact of mobile

digital terminals (MDT) on the recovery rates of motor vehicle thefts. He argued that the

MDT system would increase the number of vehicle checks conducted by patrol officers;

and as the number of checks increased, the probability of identifying and recovering

stolen vehicles should also increase. He found that the post-MDT recovery ratio was

higher than the pre-MDT ratio in each of the MDT cities; the change in post-MDT ratios

in the MDT cities exceeded that of non-MDT cities; and the presence of MDT technology
 50

was significantly associated with higher percentages of motor vehicle thefts recovered,

controlling for the total level of thefts. Meehan (1998) also examined MDT’s impact on

patrol and investigative services and found that MDT significantly improved

effectiveness of the police. In another study about MDTs, Ioimo and Aronson (2003)

analyzed whether the records, investigations and police administration bureaus derive

measurable benefits from mobile computing. They used agency data and survey

responses collected from a medium-sized police department located in Arizona and

reported that computing improved the rate of recovery of stolen vehicles.

Marshall (1998) analyzed the impact of cellular digital packet data (CDPD)

technology on officers’ performance. Two cars from six local law enforcement agencies

were equipped with CDPD technology and tested for 10 days. Researchers found that,

although the test group worked less than the control group, they made 17.12% more

inquiries, and 18.94% more arrests/citations than the control group. On the other hand,

Nunn & Quinet (2002) also evaluated the impact of CDPD technology on problem

oriented policing (POP) at a state police agency. The main responsibility of POP officers

was forming partnerships with other law enforcement agencies and community groups to

understand and address the underlying causes of problematic issues. Therefore,

communications and information exchange were seen as crucial ingredients to a

successful POP program. They compared performance of POP officers using CDPD with

the performance of POP officers without CDPD, and found that CDPD use did not make

any significant difference between the control and experiment group26.

26
Nunn & Quinet (2002) defined four measures of job performance: 1) officers’ inquiry data, 2) data
compiled from productivity reports, 3) actual job performance evaluations and 4) qualitative observations
of the officers.
 51

Finally, McRae and McDavid (1988) conducted a cost benefit analysis of mobile

digital communications systems implemented at the Vancouver Police Department.

Because of the difficulty of assigning monetary values to criminal cases, they analyzed

arrests due to minor crimes. The equipment, installation, staff and maintenance costs

were incorporated into the analysis. Arrest data and system usage patterns were collected

with a survey administered to officers making warrant arrests. Researchers found that the

system generated positive net benefits. In addition to the foregoing, the system has

produced intangible benefits such as increased officer safety, inquiry speed and enhanced

message security.

IT’s impact on organizational structure.

Several studies discussed IT’s impact on the structure of law enforcement

organizations. Nunn (2001) studied the impact of IT on organizational resources by

analyzing Law Enforcement Management and Administrative Statistics (LEMAS) survey

data of the Department of Justice. He found that highly computerized agencies spend

more per capita on IT and have more technical staff and deliver services with fewer

officers per capita than agencies reporting lower levels of computerization. Over more

than a decade, Peter Manning (1992; 1996; 2001; 2004; 2008) has made important

contributions to this topic. He stated that there is a reciprocal impact between technology

and organization. He argued that the “technology in the workplace is molded and shaped

by the environment, the organizational structure, and the occupational culture of policing

more than they are shaped by technology” (Manning, 1992, p. 388). Therefore, he argued

that IT implementation in law enforcement agencies should be accompanied by

organizational change in order to be successful.

 52

Chan (2001) analyzed the extent to which information technology has modified

the accountability structure, the occupational culture, and policing practices at the street

supervisory and management levels of an Australian police department. She stated that,

although information technology had not altered the dominant style of policing much

within the department, it had altered the value of communicative and technical resources

because increased information processing and dissemination capacity had increased the

agency’s ability to analyze criminogenic factors and to implement POP more effectively.

Moreover, information technology had also altered the accountability structure and

reporting procedures, and it had restructured daily routines of operational policing. On

the other hand, similarly to the Manning, Chan acknowledged the dominance of the

traditional police style and stated that IT’s full benefit could only be realized after a

massive change of agency culture.

In sum, majority of studies found that IT is contributing to the organizations’

performance. However, as argued by DeLone and McLean, a variety of outcome

measures were used to evaluate the success of information systems. Therefore, it is not

possible to make a comparison between studies.

Compstat

Since Compstat has been acclaimed as an important police innovation involving

information technologies (Silverman, 2006), it deserves special attention. Walsh & Vito

(2004) considered Compstat innovation as the emergence of a new model of policing and

defined it as “a goal-orientated strategic management process that uses information

technology, operational strategy and managerial accountability to guide police

operations” (p. 57). O’Connell (2002) described Compstat as a managerial strategy which
 53

employs a “business-like” approach to crime fighting. Compstat has four core principles:

1) accurate and timely intelligence, 2) effective and appropriate crime prevention tactics,

3) rapid deployment of personnel and resources, and 4) relentless follow up and

assessment.

Basically Compstat consists of directed patrol, geographic accountability of

precinct commanders, and the use of information and mapping technology (Eck &

Maguire, 2005). The Compstat process has three important components. The first step is

the collection and compilation of detailed weekly crime statistics. The second step is the

distribution of Compstat reports to the operational managers. The Compstat report

contains a concise summary of current crime in each location and provides a comparison

with previous crime rates to better see established and emerging crime trends. Finally,

operational managers are empowered and held accountable (particularly through weekly

Compstat meetings with managers) for addressing crime and disorder in their areas (Vito,

Walsh & Kunselman, 2005)27.

Compstat has been praised by some researchers and practitioners as a major factor

in the decline of New York’s crime rates. For example, Silverman (1999, 2006) claimed

that Compstat is one of the three factors that contributed to New York’s crime drop. He

also maintained that a crime drop has been observed in other cities that implemented

Compstat as well. In New Orleans, homicide rates dropped 55 percent within five years.

In Minneapolis, homicides, aggravated assaults, robberies, burglaries and auto thefts

decreased more than ten percent in each case. In Baltimore, overall crime rates dropped

25 percent (Silverman, 2006).

27
See Shane’s (2004) “Compstat Implementation” for further information.
 54

However, there is a limited number of empirical studies on the impact of the

Compstat program, and those studies are limited to its impact on crime rates. Rosenfeld,

Fornango and Baurner (2005) discussed the crime prevention effects of Boston’s

Operation Ceasefire, New York’s Compstat and Richmond Virginia’s Project Exile.

Their analysis yielded significant results only for Richmond Virginia’s Project Exile. Eck

& Maguire (2005) argued that in order to make an empirical claim about the effectiveness

of Compstat (or any other similar innovation), there must be four types of evidence

present. First, any explanation of Compstat’s crime prevention effect should be theory-

based. Then, there must be a statistical association and temporal order between Compstat

and crime reduction. Finally, rival hypotheses about crime reduction must be eliminated.

They stated that Compstat’s crime prevention rationale and theoretical base can be

likened to the hot spots approach. However, Weisburd, Mastrofski, Willis and Greenspan

(2006) showed that statistical association and temporal order are problematic for New

York crime data because the crime drop alone does not show any causation, and the

crime drop in New York had begun before the Compstat implementation. Finally and

most importantly, Compstat was implemented together with some other innovations, so it

is not possible to differentiate the individual effects of each program on crime rates.

Weisburd , Mastrofski, McNally and Greenspan (2002) looked at Compstat’s

impact on the use of strategic problem solving by American police agencies. They

examined the diffusion of Compstat programs and the nature of Compstat models

throughout the United States. They found that strategic problem solving is not a result of

Compstat because some elements of strategic problem solving had begun to be

implemented before the Compstat programs. In addition, Weisburd et al. (2006) argued
 55

that “Compstat as it has actually been implemented by American police agencies has

focused more on reinforcing and legitimating the traditional bureaucratic military model

of police organization than on innovation in the practices of policing” (p. 298). Similarly,

Greene (2000) argued that Compstat is strengthening the traditional police outlook. In

sum, Compstat is an important development in the use of information by policing, but

evaluating its impacts is problematic, and available studies are not sufficient to make a

conclusion about its impacts.

Summary

The literature review demonstrates that criminal investigation is a kind of

information work, and recovery, analysis, and interpretation of information are crucial

tasks for the success of criminal investigation. Several factors have been found to affect

the outcome of criminal investigation. These are quality of preliminary investigation,

availability of information about the suspect, specialized investigative units, and social

characteristics of the community served. Forensic evidence is also found to be effective,

but due to its underutilization, its contribution to the investigative process is small.

Finally, there are mixed findings about the impact of agency size and workload of the

investigators on the criminal investigations.

One of the most important findings of criminal investigation literature for this

study was the importance of information in the investigative process. Numerous studies

reported that the information gathered during the preliminary investigation is the most

important determinant of the outcome of a criminal investigation. Moreover, some

studies claimed that other factors affecting the success of criminal investigation are

actually affecting police’s capacity of information processing first. For example, Burrows
 56

& Tarling (1982) argued that some types of crimes are hard to solve because the lack of

contact between victim and witness in those types of crimes decreases the likelihood of

recovering useful information about the suspect. Therefore, availability of information

and law enforcement agency’s recovery, analysis and interpretation capability appear as

crucial factors for the outcome of criminal investigations.

Since information is an important factor for the outcome of criminal

investigations, information technologies have a great potential to help officers in their

investigative duties. The literature review showed that information technologies have

been used for both administrative and operational purposes in law enforcement agencies

for almost four decades. While administrative use involves traditional services such as

budgeting, personnel or fleet management, operational use involves mainly investigative

activities. Within the criminal investigation sphere, information technologies are

employed for records management, crime analysis, communication and mobile

computing and forensic analysis. Although researchers argued that IT implementation

was problematic for law enforcement agencies, most of the IT evaluation studies found

that IT provides a positive contribution to various aspects of policing.

However, the number of available studies is not sufficient to make a safe

conclusion about information technologies’ impact on policing. Moreover, the diversity

of measures used in those studies prevents us from making a comparison between them.

For these reasons, this study will contribute to the field by providing new evidence about

information technologies’ impact on police work.

 CHAPTER 3: METHODOLOGY

Research Question and Hypothesis

The review of the related literature showed that information technologies are used

extensively in criminal investigations; however, their impact on the outcome of

investigations is not well known. This study is an exploratory research which aims to

shed more light onto the subject by employing an empirical model and testing it with the

data gathered from U.S. police departments. This research examines the following

question:

Does use of information technologies have an impact on the outcome of criminal

investigations?

As discussed in the literature review chapter, DeLone and McLean’s (1992; 2003)

IS Success Model argued that there was a relationship between use of information

technology and net benefits. The term Net Benefit was loosely defined by the authors,

and it can mean any improvement depending on the context of the information systems.

In the law enforcement context, clearance is one of the most important goals of

investigative activities (Chaiken et al., 1976; Eck, 1989; Greenwood, 1970; Horvath et

al., 2001; McDevitt, 2005; Waegel, 1981). Crime clearance in general is an important

goal of traditional policing, which was also influential on the design and implementation

of the majority of current police information systems (Greene, 2000). Therefore, this

study expects that the investigative use of information technology increases crime control

57
 58

abilities of law enforcement agencies. This research will test the following hypothesis:

H1: As police departments’ use of information technologies for investigative

purposes increases, the number of crimes cleared by them also increases.

However, it must be said at this point that this study does not test any causal

relationship between use of information technologies and clearance rates. In order to be

able to investigate causal relationship, three things must be considered. First of all, there

must be covariance between independent and dependent variables. Secondly, variation in

the dependent variable must be occurred after the variation in the independent variable

which is called temporal order. Finally, variation in the dependent variable should not be

due to variation in a third variable which is called nonspuriousness (Schutt, 2006). Since

this study is a one-group only post-test design, it lacks both the comparison group and the

baseline measurement. Therefore, it is not possible to control for temporal order and the

nonspuriousness. This study can show whether there is a relationship between use of

information technologies and clearance rate, but it cannot show whether use of

information technologies is the cause of variation in the clearance rates.

Data

This study is a cross-sectional study. Its unit of analysis is the police departments.

This study has three data sources: a self-administered survey, the Uniform Crime Report

(UCR) Program data and U.S. Census data.

The impact of the science and technology survey.

The biggest portion of the data is collected by the “National Study of the Impact

of Science and Technology on the Process of Criminal Investigation in Law Enforcement

Agencies” project (it will hereafter be referred to as the “Impact of Science and

58
 59

Technology Survey”) which was funded by National Institute of Justice (NIJ), and

conducted by Eastern Kentucky University (EKU) researchers. The purpose of the Impact

of Science and Technology Survey was to describe and assess the impact of science and

technology on the process of criminal investigation in law enforcement agencies. The

project began in 2005 and data collection ended in 2006. No publications have been

written based on the project data yet. EKU has approved the use of data in this project,

and EKU researchers delivered the survey dataset and other available documents via

email.

The self-administered survey, which can be found in Appendix A, consisted of

eleven parts. It was developed using some of the questions posed in the Chaiken et al.

study published in 1976. Questions consisted of demographic information of participating

agencies, agency procedures related to criminal investigation, formal investigative

training received by the officers, selection and use of investigators, use of crime

laboratory services, and available technologies. Survey instruments were sent to 1) the

top 200 largest law enforcement agencies, 2) all the state law enforcement agencies and

3) a random sample of municipal, county and campus agencies. 630 agencies received the

survey, and 280 of them responded. Table-2 illustrates the overall response rates achieved

for each group and overall. The response rate is higher among bigger police agencies and

state police agencies than municipal agencies, campus police and sheriff departments.

Forty-four percent of all agencies that were solicited to participate in the study responded

to the survey.
 60

Table 2
Police Departments’ Response Rates to the Impact of Science and Technology Survey
Number Number Number Response
Mailed Returned Received Rate (%)
Sheriff 62 2 18 29.03
Municipal 267 4 94 35.21
Campus 52 1 17 32.69
Top 200 200 8 121 60.5
State 49 1 30 61.22
Total 630 16 280 44.44

Four waves of mailings were used to collect the data. One mailing with the

survey, cover letter, and return self-addressed, stamped envelope was sent to respondents.

This mailing was followed by a reminder postcard. Another two waves of surveys were

sent to the agencies that had yet to respond. Totally 280 agencies responded to the

survey. Responding agencies were composed of 18 Sheriff, 94 municipal, 17 campus, 30

state and 121 of the top 200 agencies.

As stated above, there were 17 campus agencies in the sample. Differently from

the local law enforcement agencies, campus law enforcement agencies’ jurisdictions

comprise of campuses and their surrounding areas. Although Peak, Barthe and Garcia

(2008) reported that jurisdiction of campus law enforcement agencies has increased since

1986, 42% of agencies reported having jurisdiction only on their campuses. In addition,

campus law enforcement agencies differ from local law enforcement agencies by their

responsibilities. Peak et al. (2008) stated that in 2006 the top reported activity of campus

agencies was criminal investigation. However, in a BJS report Reaves (2008) showed that

there were differences between large and small campus agencies in terms of criminal

investigations. Reaves (2008) reported that 98% of large campus agencies conduct any
 61

type of property crime investigation, but this figure drops to 74% for small campus

agencies. Again 98% of large campus agencies conduct any type of violent crime

investigations, but only 66% of small campus agencies do the same. The biggest reported

difference between campus agencies was homicide cases. While 77% of large campus

agencies were conducting homicide investigations, only 27% of small campus agencies

were doing so. Since not all campus law enforcement agencies have equal investigative

responsibilities, and a big portion of campus agencies do not have investigative

responsibilities for some crimes, campus agencies were removed from the sample.

On the other hand, state police departments were not suitable for the objectives of

this study, either. The first state police agency was established in Pennsylvania as a

military-style police force to deal with labor and ethnic violence (Steverson, 2007). Other

states followed Pennsylvania, but the responsibilities of state police departments have

changed over time. Today state police are mainly responsible for 1) motor vehicle law

enforcement on state highways, 2) security for government personnel and property, 3)

criminal investigation as a supplement to local agencies, 4) criminal records collection

and distribution, 5) forensic services, 6) law enforcement training, 7) communication

system coordination, and 8) some special services such as emergency response (Purpura,

1996). Since, criminal investigation is not a major function of state agencies, they were

removed from the sample as well.

As a result, after removing 47 campus and state law enforcement agencies, 233

cases remained in the sample.

 62

UCR.

The second data source of this study is the Uniform Crime Reporting (UCR)

Program. The UCR Program is “a voluntary city, county, state, tribal, and federal law

enforcement program that provides a nationwide view of crime based on the submission

of statistics by law enforcement agencies throughout the country” (FBI, 2004a, ¶ 3). As

stated in the literature review, the UCR Program was first implemented in 1929 by the

International Association of Chiefs of Police, and transferred to the FBI in 1930. FBI’s

primary objective in collecting crime data is “to generate a reliable set of crime statistics

for use in law enforcement administration, operation, and management” (FBI, 2004c, ¶

1). Although the UCR Program is a good source of crime and clearance information for

practitioners and researchers, UCR data has important limitations that must be well

understood prior to any decision.

The UCR Program collects data by a series of administrative forms that is known

as the Summary Reporting System (SRS). There are eight data reporting forms currently

in use: 1) Return A: Report of Offenses known to the Police, 2) Supplement to Return A

(capture information on the value and type of the property stolen and recovered and the

nature of the offenses reported in Return A), 3) Age, Sex and Race of Persons Arrested,

4) Supplementary Homicide Reports (provides additional details about homicide cases),

5) the Law Enforcement Officers Killed and Assaulted forms, 6) the Hate Crime Data

Collection forms, 7) the Monthly Return of Arson Offenses Known to Law Enforcement,

and 8) Law Enforcement Employee Report (Barnett-Ryan, 2007).

This study will use the data collected by Return A which captures monthly

information about Part I offenses reported to the law enforcement agencies and any
 63

associated clearances. Offenses reported by this form include murder and non-negligent

manslaughter, manslaughter by negligence, forcible rape28, attempts to commit forcible

rape, robbery by weapon type, aggravated assault by weapon type, other assaults,

burglary –forcible entry, burglary – and unlawful entry (no force), attempted forcible

entry, larceny-theft and motor vehicle theft by type of vehicle. Agencies report how many

reports of offenses they received, and how many of them were found to be baseless or

unfounded after initial investigation. Then agencies calculate the number of actual

offenses by subtracting unfounded offenses from the total number of reported offenses

(FBI, 2004b).

In addition to the number of crimes reported to law enforcement, Return A

captures information about the clearances. First of all, it is important to note that the

clearances reported in the Return A on any given month may not have any relationship

with the offenses reported for that month. Therefore, clearance figures are not an

adequate measure of how many of the crimes that were committed in that month were

cleared. According to the UCR data, a case is cleared by two ways: arrest or exceptional

means. To be able to clear an offense by arrest, at least one person should be 1) arrested,

2) charged with the commission of the offense, and 3) turned over to the court for the

prosecution. In some situations, it becomes impossible to arrest and formally charge the

offender. For example, if the offender is dead (e.g., suicide or justifiably killed by law

enforcement or citizen) or if the victim or witness refuses to cooperate after the offender

has been identified, arresting and formally charging becomes impossible. According to

the UCR classifications, those cases are called exceptional clearance (FBI, 2004b). FBI

28
Forcible rape is defined as “the carnal knowledge of a female forcibly and against her will,” and statutory
rape, incest or other sex offenses are not included in this category (FBI, 2004b, p. 19).
 64

disseminates crime and clearance data in its annual publication called Crime in the

United States.

In order to overcome conceptual problems, the UCR Program has developed

uniform definitions for each offense type and devised a set of rules governing the

classification and scoring of offenses. To ensure data quality, the FBI provides a

handbook (UCR Handbook) to the participating agencies which explains those rules.

According to the UCR Program’s Hierarchy Rule, if more than one offense is committed

in an incident, only the most serious one is reported and related offenses are not reported.

Therefore, “when more than one Part I offense is classified, the law enforcement agency

must locate the offense that is highest on the hierarchy list and score that offense

involved and not the other offense(s) in the multiple-offense situation” (FBI, 2004b, p.

10). For example, if an individual breaks into a home to commit a theft and assaults the

owner of the house, law enforcement will report the incident as aggravated assault.

However, if there is a separation of time and place between multiple offenses then the

agency reports them individually29. The hierarchy rule was adopted to prevent the

multiple reporting of a criminal incident, but it obstructs the accurate depiction of the

crime problem by not measuring related crimes. The UCR hierarchy of Part I offenses is

presented in Table 3. Arson, motor vehicle theft and justifiable homicide are exceptions

of hierarchy rule. Arson is reported independently of the other Part I crimes. If an

incident contains both motor vehicle theft and theft of items located in the vehicle, only

29
This is called the Separation of Time and Place Rule. Here is an example to that rule given in the UCR
Handbook (FBI, 2004b): A man and a woman were parked at a secluded location. A gunman surprised
them and shot and killed the man when he resisted. He abducted the woman and drove across town to a
secluded area where he forcibly raped her. The police arrested the perpetrator at the scene. In this incident,
the Hierarchy Rule does not apply, and two crimes are reported separately because there is a separation of
time and place between two crimes.
 65

motor vehicle theft is reported. Justifiable homicide is defined as “the killing of a felon

by a peace officer in the line of duty or the killing of a felon, during the commission of a

felony, by a private citizen” (FBI, 2004b, p. 17). An incident of justifiable homicide and

the offender’s felonious act are reported separately.

Table 3
UCR Part I Offense Hierarchy
Offense Type Description
1. Criminal Homicide a. Murder and Nonnegligent Manslaughter
b. Manslaughter by Negligence
2. Forcible Rape a. Rape by Force*
b. Attempts to Commit Forcible Rape
3. Robbery a. Firearm
b. Knife or Cutting Instrument
c. Other Dangerous Weapon
d. Strong-arm—Hands, Fists, Feet, etc.
4. Aggravated Assault a. Firearm
b. Knife or Cutting Instrument
c. Other Dangerous Weapon
d. Hands, Fists, Feet, etc.—Aggravated Injury
5. Burglary a. Forcible Entry
b. Unlawful Entry—No Force
c. Attempted Forcible Entry
6. Larceny-theft
7. Motor Vehicle Theft a. Autos
b. Trucks and Buses
c. Other Vehicles
8. Arson a. Structural
b. Mobile
c. Other

After classifying the offenses, an agency should determine the count or the

number of offenses committed. In order to correctly score the offenses, it is important to

 66

distinguish crimes against the person from crimes against property. For scoring purposes,

homicide, forcible rape and aggravated assault are grouped as crimes against the person,

and robbery, burglary, larceny-theft, motor vehicle theft and arson are grouped as crimes

against property30. In the case of crimes against the person, one offense is counted for

each victim, and for crimes against property, one offense is counted for each distinct

operation or attempt—except in the case of motor vehicle theft for which one offense is

counted for each stolen vehicle and one offense for each attempt to steal a motor vehicle.

An exception to these standards is called the Hotel Rule. According to this rule “if a

number of dwelling units under a single manager are burglarized and the offenses are

most likely to be reported to the police by the manager rather than the individual tenants,

the burglary must be scored as one offense” (FBI, 2004b, p. 62).

Participation in the UCR Program is voluntary, and there are no rules that specify

the mandatory amount of data to submit for participating agencies. Although the UCR

Program has achieved participation of more than 90% of all police agencies, sometimes

some agencies cannot provide data due to computer problems, changes in record

management systems, personnel shortages, or a number of other reasons (FBI, 2004c). To

provide a complete picture of crime in the United States, the FBI imputes values for the

missing pieces of data caused by nonparticipation, incomplete participation and limited

participation (Barnett-Ryan, 2007). This process is called estimation by the FBI31.

There are two other issues about the UCR data that should be mentioned. First of

all, it is well known that not all crime is reported to the police; hence, not all crime is

30
This classification is used just for reporting purposes. The FBI uses a different classification in its crime
reports. In that classification, crimes against the person consists of homicide, forcible rape, robbery and
aggravated assault, and crimes against property consists of burglary, larceny-theft, motor vehicle theft and
arson (FBI, 2004c).
31
See Maltz (2007) for further information on the FBI’s estimation procedures
 67

known or recorded by the law enforcement agencies. Moreover, Part II crimes are not

recorded by the UCR Program32, and because of the Hierarchy Rule some Part I offenses

are discarded in multiple offense incidents as stated above. Secondly, the accuracy of

UCR data is dependent on the efforts of the reporting agency because the FBI does not

have an adequate control mechanism to ensure that the reported data is flawless. In

addition, it is also argued that UCR data is open to manipulation by participating

agencies. In order to look better in the statistics and to abate the political pressure, some

participating agencies may choose not to report all of the crimes occurring in their

jurisdictions (Chambliss, 2001). Therefore, UCR data must be interpreted by taking these

limitations into consideration.

For this study, 2005 and 2006 UCR datasets were downloaded from the Inter-

University Consortium for Political and Social Research (ICPSR) web site. UCR data

were incorporated into the analysis in three steps. First, 233 police agencies that had

responded to the Criminal Investigation Survey were looked up in the UCR database.

Crime and clearance data of 220 of the 233 agencies were located in the 2005 UCR

dataset and 216 of the 233 agencies were located in the 2006 UCR dataset. Then, 2005

and 2006 crime and clearance data were added to the Criminal Investigation Survey

dataset.

U.S. Census.

The third data source is the U.S Census Bureau’s 2000 Census data. Depending

on the jurisdiction of the responding police department, town, city and county level

poverty, income and race variables were queried at the 2000 Census database via the U.S.

32
Only arrest data involving the Part II crimes are reported to the UCR.
 68

Census Bureau’s American FactFinder website (U.S. Census Bureau, 2009). Census data

was added to the Impact of Science and Technology Survey dataset.

Variables

Dependent variable.

The dependent variable of this study is the clearance rates for Part I crimes which

include murder and non-negligent manslaughter, forcible rape, robbery aggravated

assault, burglary, larceny-theft, and motor vehicle theft33. Arson was not included in the

analysis because it is reported separately in the UCR Program and a lesser number of

agencies report arson statistics. In order to calculate the clearance rates, monthly crime

and clearance figures were summed to get the annual crime and clearance numbers for

each agency. Then, clearance rates were calculated by dividing number of clearances by

the number of crimes known to the police.

Violent crime and property crime clearance rate variables were calculated by

using the classification used by the FBI in its crime reports. According to that

classification the violent crime category consisted of murder and nonnegligent

manslaughter, forcible rape, robbery and aggravated assault. The property crime variable

consisted of burglary, larceny-theft, and motor vehicle theft. In addition to the violent and

property crime variables, an all crime clearance rate variable was also calculated. For

these three variables, crime and clearance numbers were calculated by adding respective

crime types and then calculating the clearance rate by dividing number of clearances by

number of crimes known to the police.

33
Information technologies may contribute to the solution of Part II crimes as well. However, Part II crimes
are not reported by the UCR Program, and they were not collected by the Impact of Science and
Technology Survey either.
 69

Clearance rates vary depending on the type of crime. Literature review showed

that Part I crimes are generally perceived as more serious, and more organizational

resources are devoted to their investigations. The percentage of crimes cleared by arrest

for 2005 and 2006 are given in Table 4 below. It shows that crimes against the person are

more likely to get solved than crimes against property. Although property crimes were

harder to solve due to lack of information about the suspect (Skogan & Antunes, 1979),

different amount of resources devoted to crimes against the person and crime against

property may also be influential on the difference between clearance rates.

Table 4
Percentage of Crimes Cleared by Arrest or Exceptional Means in 2005 and 2006

Motor
Forcible Aggravated Larceny Vehicle
Year Murder Rape Robbery Assault Burglary -theft Theft

2005 62.1% 41.3% 25.4% 55.2% 12.7% 18.0% 13.0%

2006 60.7% 40.9% 25.2% 54% 12.6% 17.4% 12.6

6ote. Published in Crime in the United States (FBI 2005a; 2006a)

Use of clearance rates as a measure of police performance has been criticized for

several reasons. First of all, the UCR does not differentiate among cases and accepts

every cleared case as equal; however, as discussed in the literature review chapter

investigation of some cases can be more difficult than others and may require more skill

and resources. An agency might score low on clearance rates just because it handles more

difficult cases than others. Unfortunately, UCR data do not provide any measure for the

difficulty of the case or the effort that is put on a case by the investigators. Furthermore,

as stated above, law enforcement agencies may distort the crime and clearance figures for

political reasons. Despite these problems, clearance rates are still one of the best available
 70

police performance indicators and used by researchers (e.g. Helms, 2008; Holmes et al.,

2008; Nolan, 2004; O’Brian, 2003).

Independent Variable.

As presented in the literature review chapter, the DeLone and McLean model

proposes a relationship between information technology use and net benefit. Intention to

use was also added to the model later because use of some systems might be mandatory.

The literature review showed that use of most of the law enforcement IT systems is

mandatory. In other words, in most of the departments, officers are forced to use the IT

system in order to accomplish the job. DeLone and McLean argued that the mandatory

nature of system use does not cause rejection of the system use measure as a success

variable. They stated that:

Even when the use is required, variability in the quality and intensity of this use is

likely to have a significant impact on the realization of the system benefits.

Furthermore, no system use is totally mandatory. .… Thus, whereas usage of

system may be mandatory at one level, the continued adoption and use of the

system itself maybe wholly voluntary, based upon management judgment at a

higher level. Management always has the option of discontinuing a system that is

not providing the desired results and benefits. (DeLone and McLean, 2003: p.16-

17).

The independent variable of this study measures IT use by police departments. As

argued by DeLone and McLean (2003), different levels of IT use are expected to impact

the outcome variable differently. In order to measure the extent of use of information

systems in the surveyed agencies, “use of IT scale” was constructed. The use of IT scale
 71

was composed of 29 items asking the availability of following information technologies

in responding police departments:

1. Cellular or mobile phones

2. Global Positioning System (GPS)

3. Personal Digital Assistant (PDA) or handheld computer

4. Blackberry

5. Mobile Data Computer/Mobile Data Terminal (MDC/MDT)

6. Intranet

7. Web pages or web-based applications

8. Computer Based Training (CBT)

9. Laptop computer

10. National Integrated Ballistics Information Network (NIBIN)

11. Integrated Ballistics Identification System (IBIS)

12. Violent Criminal Apprehension Program (ViCap)

13. Regional Information Sharing Systems (RISS)

14. National Crime Information Center (NCIC)

15. Crime reports on computer form

16. Arrest reports on computer form

17. Case disposition information on computer form

18. Prosecution disposition information on computer form

19. Court dispositions information on computer form

20. Summary crime statistics on computer form

21. Fingerprints on computer form

 72

22. Known offender on computer form

23. M.O. file on computer form

24. Mug shot on computer form

25. Organized crime intelligence on computer form

26. Narcotics intelligence on computer form

27. Sex offender information on computer form

28. Stolen property information on computer form

29. Stolen vehicles information on computer form

In order to determine whether those variables could be used to construct an index

variable, their correlations were examined. 29 variables were incorporated into the

analysis. Descriptive analysis results showed that there were some missing values in

those variables, so observations with missing value were dismissed from the analysis to

get more reliable results. Then, a correlation analysis was conducted. The correlation

coefficients are presented in Appendix B. The correlation matrix revealed that

information technology variables have generally moderate correlations with each other.

The standardized alpha coefficient (or Cronbach's alpha) was 0.87 for all the variables.

Although Kuder-Richardson Formula 20 was designed for reliability analysis of

dichotomous variables, Cronbach's alpha gives same results with dichotomous variables

(Traub, 1994). In this analysis, Cronbach's alpha measures how well the 29 variables

measure a single unidimensional latent construct. If data would have a multidimensional

structure, this value would usually be low. Since 0.7 or higher is generally considered

acceptable (Traub, 1994), we can say that the 29 information technology variables have a

unidimensional structure and can form an index variable.

 73

To further investigate this proposition, and to see if there is any structure in the

relationships between variables, a factor analysis was conducted. Factor analysis is an

exploratory method. It is mainly used to reduce the number of variables in a dataset and

to uncover the structure in the relationship between the variables. Basically, factor

analysis condenses intercorrelated variables into fewer dimensions, which are called

factors. There are different extraction methods used to do this. In this analysis, the

principal axis factoring method was utilized as the extraction method because principal

axis factoring seeks the least number of factors which can account for the correlation of a

set of variables. In addition, the varimax method, which is an orthogonal method, was

used as the rotation method. Since the varimax method (or orthogonal rotation methods)

increases each variable’s likelihood of loading on a single factor, it is more suitable for

the purpose of this analysis (Kim & Mueller, 1978).

There are two common methods in factor analysis that can be used to determine

the number of factors. The first method is to use eigenvalues as the cut-off value.

Eigenvalue measures the amount of variation in the total sample that is accounted for by

each factor, and eigenvalue greater than one is usually taken as the cut-off value. The

second method is to look at the scree plot of eigenvalues against the number of factors.

The point where the eigenvalues line begins to level off can be used as a cut-off point.

Both methods can be used to determine the number of factors, but the best method is to

use eigenvalues and scree plot together as a guide and to look at the interpretability of the

factors given by the analysis. If a factor does not have any meaning, in other words, if it

does not indicate any structure in the relationships between variables, then the solution

may be discarded. Another important output of factor analysis is factor loading which is
 74

the correlation coefficient between a variable and a factor. A factor loading greater than

0.3 is usually considered significant, and that variable is said to be loaded on that factor

(Kim & Mueller, 1978).

In this analysis, factors with eigenvalues greater than 1 were chosen in the first

run of the factor analysis. Eight factors were identified. 46.81 % of the total variance was

explained by these eight factors, and 53.19% of variance remained unexplained. Only one

variable did not load on any factor. Nine variables loaded onto the first factor, and they

were mainly communication or IT infrastructure related variables; however the

interpretation of the resulting distribution was difficult. The scree plot showed that three

factors accounted for most of the variance. Therefore, factor numbers were constrained to

three in the second run. Explained variance decreased to 32.93%. 15 variables loaded on

the first factor, 8 variables loaded on the second factor and 4 variables loaded on the third

one. 2 variables did not load on any factor.

In the third run, two variables that did not load on any factors were dismissed

from the analysis for better results. The variables that were dismissed from the analysis

were use of cellular/mobile phones and use of the National Crime Information Center.

The table of factor loadings of the third factor analysis can be found in Appendix C. The

explained variance slightly improved to 34.98%. Generally, variables that were related

with the IT infrastructure and availability of external databases loaded on the first factor.

Fingerprints, M.O. File, organized crime intelligence and narcotics intelligence items also

loaded on the first factor. However, the M.O. file, organized intelligence and narcotic

intelligence variables did not seem matched with this group. Moreover, the narcotic

intelligence variable loaded on the third factor at the same time. The second factor
 75

consisted of variables that were more related with record management systems and

information processing capabilities of the agencies. And the third factor consisted of

variables that were related with prosecution and court procedures. In sum, factor analysis

did not reveal any meaningful structure among the items. Because of this, the decision

was made to use all items on a single scale.

Control Variables.

Based on the review of literature, seven control variables were identified:

1. Department size: The department size variable is measured as the total

number of full-time sworn officers by the Impact of Science and Technology Survey. As

presented in the literature review, research on the impact of department size on clearance

rate is mixed. Larger agencies generally have more resources which may be helpful in

clearing crimes. In addition, larger agency personnel can build expertise on specific

crimes which may help solution of crimes. Therefore, larger agencies are expected to

have more clearances.

2. Proportion of sworn officers working at investigative functions: Department

size may not be an adequate measure to capture a law enforcement agency’s investigative

capabilities because not all of the officers are working at investigative units in a police

department. The proportion of officers working at investigative functions variable is

calculated by dividing the number of sworn officers working at investigative functions by

the total number of full-time sworn officers in the department. Both variables were

measured by the Impact of Science and Technology Survey.

3. Population Served: Community size is the population that is served by the

agency. It is measured by the Impact of Science and Technology Survey. Researchers

 76

have argued that urban areas provide more anonymity to offenders because in urban areas

it becomes more difficult for police to find witnesses and to know offenders and criminal

networks (Pare et al., 2007; Wilmer, 1970).

4. Poverty: Poverty variable measures the economic status of the community. It

is measured as the percentage of people below the poverty threshold as provided in the

U.S. Census data.

5. Median Income: The median income variable is included as another indicator

of economic status of the community. Moreover, median income variable is also used as

a proxy to the resources available to the police departments. Median income variable was

used as a proxy to the resources available to organizations in some other studies as well

(Kennedy, 2009; Zaid, 1967). Median income variable is taken from the U.S. Census

dataset.

6. Percent White: Percent white is another variable measuring societal

characteristics of police department’s jurisdiction. It is taken from U.S. Census data.

7. Crime rates: As discussed above, crime rates for Part I crimes reported in the

UCR statistics are used. Since the number of crimes increases the workload of the police,

detectives may not give adequate time to the cases and clearance rates may decrease.

Crime rate variable is calculated by dividing number of crimes known to the police to the

size of the population and scaling it up by a multiplier. In this study, 1,000 is preferred as

the multiplier as commonly used by law enforcement agencies such as FBI.

 CHAPTER 4: RESULTS

This chapter presents data analysis results in three parts. In the first part,

descriptive statistics of all variables are provided. Then, bivariate relationships are

examined in the second part. The third part presents multivariate analysis results. A

supplementary analysis of use of IT scale items is presented also within the third part.

Univariate Analysis

The unit of analysis of this study is law enforcement organizations. As presented

in the methodology chapter, 280 police departments responded to the Impact of Science

and Technology Survey. Responding departments were composed of 18 Sheriff, 94

municipal, 17 campus, 30 state and 121 top 200 departments. Since campus and state law

enforcement departments are different from local police departments in terms of

investigative activities (Reaves, 2008; Purpura, 1996), they were removed from the

dataset. As a result, 233 departments remained in the sample. The final dataset was

composed of 18 Sheriff, 94 municipal, and 121 top 200 departments.

The basic descriptive statistics are given in Appendix D. Descriptive statistic

tables show the number of missing values, valid N, minimum, maximum, standard error

of the mean, median and standard deviation for each variable. As discussed in the

methodology chapter, the clearance rate variable was calculated by dividing the number

of clearances achieved by the departments by the number of crimes known to the police.

77
 78

The examination of the descriptive statistics tables revealed that some agencies reported

zero clearances. Table 5 shows the frequency of zero clearances for each offense type.

For example, 107 departments reported that they did not clear any murder case in 2005

and 102 departments reported that they did not clear any murder case in 2006. Totally,

there are 514 zero clearances reported in 2005 and 511 zero clearances reported in 2006.

This means that 22.06% of reported clearances in 2005 and 21.93% of reported

clearances in 2006 are zero.

Table 5
6umber of Zero Clearances
2005 2006
Clearances Clearances
Murder 107 102
Rape 89 90
Robbery 71 76
Assault 32 33
Burglary 40 43
Larceny 35 37
Vehicle Theft 47 54
Violent Crimes 32 23
Property Crimes 31 23
All Crimes 30 30

There are two explanations for zero clearances. If a department reported zero

clearances but at least one crime for any offense category, then that means the department

did not clear any of the reported crimes. On the other hand, if a department reported zero

crime, and zero clearance, then that means the department does not have any crime to

clear for that year. Since division by zero is undefined, calculation of clearance rates

would yield missing values for both groups of cases. However, reporting missing values

78
 79

for the former group would bias the clearance rate variable by concealing some of the

agencies with low clearance rates. For that reason, if an agency reported a number of

crimes, but zero clearance for an offense category, its clearance rate value was defined as

zero. However, if an agency reported zero crime and zero clearance in an offense

category, its clearance rate value was defined as missing. For example, 93 departments

did not report any homicide clearance in 2005 and 94 departments did not report any

homicide clearance in 2006 because they did not have any homicide cases to report.

Therefore, murder clearance rates of 93 departments in 2005 and 94 departments in 2006

are defined as missing. Murder, rape and robbery clearance rate variables have been

affected more than others by these kinds of missing values. It is also important to keep in

mind that clearances reported in any given month or year may not have any relationship

with the offenses reported for that month or year as discussed in the methodology

chapter. In other words, it is highly possible that an offense can be reported as occurring

in one year and it can be reported as cleared in another year. As a result of this,

sometimes departments may report more clearances than the number of crimes they

reported for any offense type.

In order to learn more about the investigative performance of sampled

departments, crime and clearance rates are discussed herein. As presented in the

methodology chapter, crime rates are calculated for each department by dividing the

number of crimes known to the police by the size of the population and scaling it up by a

multiplier. In this study, 1,000 is preferred as the multiplier. The means of crime rates are

calculated by dividing the sum of the crime rates involving the departments in this study

by the number of reporting departments. Similarly, the mean of clearance rates is

 80

calculated by dividing the sum of the clearance rates over the departments in this study

by the number of reporting departments. Figure 2 presents graphs illustrating the mean

values of 2005 and 2006 crime rates.

50
2005 50
2006

40 40

30 30
Mean

Mean
20 20

10 10

0 0
Assault

Violent

All

Assault

Violent

All
Murder

Murder
Rape

Vehicle

Rape

Vehicle
Robbery

Burglary
Larceny

Property

Robbery

Burglary
Larceny

Property
Figure 2. Mean values of 2005 and 2006 crime rates

Figure 2 shows that mean of property crime rates is higher than violent crime

rates in both years. Assault and larceny crimes are the most prevalent crime types in

violent and property crime groups, respectively. The mean of assault is 12.023 in 2005

and 12.307 in 2006. The mean of larceny is 20.094 in 2005 and 20.117 in 2006. Murder

and rape offenses are the least prevalent crimes in the sample. The mean of murder

crimes is 0.054 in 2005 and 0.057 in 2006. The mean of rape crimes is 0.227 in 2005 and

0.24 in 2006. Vehicle theft is the least prevalent crime type in property crimes. Its mean

is 3.999 in 2005 and 3.882 in 2006.

Figure 3 presents graphs illustrating the means of 2005 and 2006 clearance rates.

Similarly to 2005 and 2006 crime rates, the mean values of 2005 and 2006 clearance rates

are also very close to each other. The graphs also shows that murder, assault and rape had
 81

the highest clearance rates both in 2005 and 2006. In other words, the violent crimes

category has a higher clearance rate than the property crimes category in the sample. The

mean of the murder clearance rate is 0.6 in 2005 and 0.619 in 2005. The mean of the

assault clearance rate is 0.538 in 2005 and 0.54 in 2006. The mean of the rape clearance

rate is 0.375 in 2005 and 0.365 in 2006. On the other hand, burglary, larceny and motor

vehicle theft had the lowest clearance rates in the sample. The means of burglary

clearance rates are 0.14 and 0.122, larceny clearance rates are 0.79 and 0.549, and motor

vehicle clearance rates are 0.196 and 0.177 in 2005 and 2006. 26% of all crimes were

cleared in 2005 and 25% percent of all crimes were cleared in 2006.

0.6
2005 2006
0.6
0.5 0.5
0.4 0.4
Mean

Mean

0.3 0.3
0.2 0.2

0.1 0.1

0.0 0.0
Assault

Violent

All

Assault

Violent

All
Murder

Murder
Rape

Vehicle

Rape

Vehicle
Robbery

Burglary
Larceny

Property

Robbery

Burglary
Larceny

Property

Figure 3. Mean values of 2005 and 2006 clearance rates

Figure 4 illustrates standard deviations of clearance rates for nine crime categories

and all crimes. Standard deviations are calculated by taking square root of the variance.

Standard deviations of 2005 and 2006 clearance rates are very close to each other, but

2005 figures are a little higher than 2006 figures indicating a little more variation in

clearing crimes in 2005. The standard deviations of violent crimes such as murder, rape,
 82

robbery and assault are larger than other crimes. Their standard deviations are 0.358,

0.314, 0.246, 0.246 in 2005 and 0.29, 0.281, 0.247, 0.251 in 2006. This indicates that

there is more variation in the sample in clearances of violent crimes. On the other hand,

standard deviations of property crimes such as burglary, larceny and motor vehicle theft

are smaller. Their standard deviations are 0.149, 0.112 and 0.209 in 2005 and 0.086,

0.091 and 0.177 in 2006. The low standard deviation values of property crime clearance

rates indicate that there is less difference among agencies in clearing property crimes.

2005 0.35
2006
0.35
0.30 0.30

0.25 0.25
Std. Dev.

Std. Dev.

0.20 0.20
0.15 0.15
0.10 0.10
0.05 0.05
0.00 0.00
Assault

Violent

All

Assault

Violent

All
Murder

Murder
Rape

Vehicle

Rape

Vehicle
Robbery

Burglary
Larceny

Property

Robbery

Burglary
Larceny

Property
Figure 4. Standard deviations of 2005 and 2006 clearance rates

Histograms of frequency distributions of all variables are given in Appendix E.

As presented in the literature review chapter, some studies found that use of information

technologies has an impact on murder and motor vehicle theft clearance rates (Meehan,

1998; Nunn, 1994; Wellford & Cronin, 1999). For that reason, murder and motor vehicle

crimes will be presented in more detail herein. Figure 5 and Figure 6 present the

histograms and box plots of frequency distributions of 2005 and 2006 murder and motor

vehicle theft clearance rates, respectively. The mean of 2005 murder clearance rates is 0.6
 83

and the mean of vehicle theft clearance rates is 0.196. Their standard deviations are 0.358

and 0.209, respectively. Figure 5 shows that the 2005 murder clearance rate variable has

a kurtosis problem. Five departments have a murder clearance rate bigger than 1 which

indicates that those departments cleared murder cases that were reported in previous

years. The motor vehicle theft variable is positively skewed. The clearance rate value of

five departments is 1.

Murder Vehicle Theft

70
30

60
25

50
Frequency

Frequency
20

40
15

30
10

20
10
5
0

0.0 0.5 1.0 1.5 2.0 0.0 0.2 0.4 0.6 0.8 1.0

Clearance Rates Clearance Rates

0.0 0.5 1.0 1.5 2.0 0.0 0.2 0.4 0.6 0.8 1.0

Figure 5. Histograms and box plots of 2005 murder and motor vehicle theft clearance
rates

Histogram and box plots of 2006 murder and vehicle theft clearance rates are

presented in Figure 6. The mean of 2006 murder clearance rates is 0.619 and that of

vehicle theft clearance rates is 0.177. Their standard deviations are 0.29 and 0.177,
 84

respectively. The histogram of the murder clearance rate shows that its distribution is not

normal. There are 3 departments with a murder clearance rate of over 1, indicating that

those agencies cleared homicides committed in previous years. The motor vehicle theft

variable is again positively skewed. The clearance rate value of two departments is 1.

Murder Vehicle Theft

35

80
30

60
25
Frequency

Frequency
20

40
15
10

20
5
0

0
0.0 0.5 1.0 1.5 0.0 0.2 0.4 0.6 0.8 1.0

Clearance Rates Clearance Rates

0.0 0.5 1.0 1.5 0.0 0.2 0.4 0.6 0.8 1.0

Figure 6. Histograms and box plots of 2006 murder and motor vehicle theft clearance
rates

The examination of other clearance rate variables showed that they are not

normally distributed. There are many transformation methods used to correct non-

normally distributed data. Natural log transformation and square root transformation are

frequently applied transformation methods. Log transformation is not suitable for

clearance rate variables because they have too many zero values. Since the log of zero is

undefined, log transformation would increase the number of missing values in clearance

rate variables. For this reason, square root transformation is preferred. Figures 7 and 8
 85

present square rooted 2005 murder and motor vehicle clearance rates. They show that

square root transformation did not completely solve the non-normality problem, but

brought variable distributions closer to the normal distribution. The histograms of square

rooted clearance rate variables are also presented in Appendix E.

Murder Vehicle Theft

50
40

40
30
Frequency

Frequency

30
20

20
10

10
0

0
0.0 0.5 1.0 1.5 0.0 0.2 0.4 0.6 0.8 1.0
Clearance Rates Clearance Rates

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0

Figure 7. Histograms and box plots of 2005 murder and motor vehicle theft clearance
rates after square root transformation

Murder Vehicle Theft

40
50

30
40
Frequency

Frequency
30

20
20

10
10
0

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0

Clearance Rates Clearance Rates

0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0

Figure 8. Histograms and box plots of 2006 murder and motor vehicle theft clearance
rates after square root transformation
 86

Figure 9 presents the frequency distribution of the Use of Information Technology

scale. As presented in the methodology chapter, there are 29 items in the Use of

Information Technology scale. They were collected by the Impact of Science and

Technology Survey. All the items are binary response variables where zero indicates that

the agency does not have that particular information technology and one indicates that the

agency has that technology. The Use of IT score is calculated by summing 29 responses

for each agency. The mean and median of the Use of Information Technology scale is

20.186 and 21. The histogram and box plot graphs show that the majority of cases are

clustered between 10 and 29, and there are only a few agencies’ scores below ten.
20
15
Frequency
10
5
0

0 5 10 15 20 25 30
Use of IT score

5 10 15 20 25 30

Figure 9. Histogram of Use of Information Technology scale

 87

There are seven control variables used in this study. The number of officers

variable is the number of sworn officers the department has. The population variable is

the number of people living within the jurisdiction of the police department. The number

of officers and population variables were measured by the Impact of Science and

Technology Survey. The percent below poverty variable is the percentage of population

living under the poverty line. The median income variable is the median income of

people living within the jurisdiction of the police department. The percent white variable

is the percentage of white people living within the jurisdiction of the police department.

The population, poverty, median income, and race variables were measured by the U.S.

Census. The proportion of sworn officers working at investigative functions is calculated

by dividing the number of sworn officers working at investigative functions by the total

number of officers. Finally, as presented above, the crime rate variable is calculated by

dividing the number of crimes known to the police by the size of the population and

scaling it up by a multiplier. Crime variables are collected by the UCR program.

Figure 10 presents the box plots of standardized values of the number of sworn

officers, population, poverty, median income, race and proportion of investigators

variables. Standardization converts variables to standard scores, but it does not change

their frequency distributions. Data analysis, on the other hand, was conducted with the

original data, not the standardized data. Standardization was applied just to be able to

view those variables on the same scale. Box plots of median income, percent below

poverty, percent white and proportion of investigators variables show that their

distributions can be considered as normal. Median income and proportion of investigators

variables have a lesser amount of variation than other variables. The mean and median of
 88

the median income variable are 41,949.232 and 37,954. The mean and median of the

proportion of investigators variable are 0.252 and 0.154. The mean and median of the

percent below poverty variable are 13.786 and 12.7, and the mean and median of the

percent white variable are 72.995 and 75.1. Their distributions are closer to normal

distribution than in the case of other variables.

# of
Officer

Population

Median
Income

% Below
Poverty

% White

Prop of
Invest

Logged #
of Officer

Logged
Population
0

10

Figure 10. Box plots of standardized values of six control variables

However, the distributions of number of officers and population variables are

problematic. Their median values are very close to their first and third quartiles indicating
 89

that there is less variation in the sample in terms of number of officers and population

variables. The examination of their frequencies showed that 93 departments have less

than 100 officers, and 197 departments have less than 1000 officers in the sample. 42

departments have more than 999 officers and two of them have more than 10,000

officers. The population variable also has a similar distribution. 110 departments were

serving jurisdictions with a population of less than 10,000. 210 departments were serving

jurisdictions with a population of less than 1,000,000. 24 departments were serving

jurisdictions with a population of more than 999,999 and one of these is serving a

jurisdiction with a population of 8,000,000. The mean and median of the number of

officers variable are 814.681 and 210, and the mean and median of the population

variable are 398,987.399 and 152,000. Both variables’ means are larger than twice of

their medians because of several big city police departments. In order to normalize these

variables, log transformation is applied to them. The box plots of standardized values of

logged number of officers and logged population variables are presented at the bottom of

Figure 10. Box plots show that log transformation approximately normalized their

distributions. Bar charts showing the frequency distributions of the logged variables can

be found in Appendix E.

Bivariate Analysis

Pearson product-moment correlation coefficients were calculated for all variables

and presented in Appendix F. The largest positive correlation coefficient between

clearance rate variables and use of IT scale is 0.193, and the largest negative correlation

coefficient is -0.169. In addition, there are eight negative correlation coefficients (assault,

vehicle theft, violent crimes and all crimes in both 2005 and 2006 data). The use of IT
 90

variable is significantly correlated only with rape clearance rates in both 2005 and 2006

data (r=0.176, p<0.05 and r=0.252, p<0.05). Other than rape clearances, the use of IT

variable is significantly correlated with 2006 robbery clearance rates; however, this

relationship is not significant in 2005 data. The direction of significant relationships is

positive indicating that agencies that score higher on the use of IT scale achieve

significantly more clearances in those crime categories than the agencies that score low

on the use of IT scale. On the other hand, the direction of the relationship is negative for

some of the clearance rates.

Correlation analysis of clearance rate and control variables also gave some

statistically significant results. The percent white population variable has the highest

number of significant correlation coefficients. Except rape and burglary offenses, it is

significantly correlated with 2005 and 2006 clearance rate variables, and the direction of

relationship is always positive. This indicates that agencies serving in jurisdictions that

have a higher percentage of white population are more likely to achieve clearances than

the agencies serving in jurisdictions that have a lower percentage of white population.

The magnitude of the correlation coefficient varies between 0.153 and 0.338. The logged

number of officers variable is significantly correlated with the rape, vehicle theft and

violent crimes categories in both 2005 and 2006 data. Interestingly, while the direction of

relationship is positive for rape clearances, it is negative for vehicle theft and violent

crime categories. This indicates that large agencies are more likely to clear rape offenses

than the small agencies, but they are less likely to clear vehicle theft crimes. Moreover,

they are less likely to clear violent crimes than the small agencies. The median income

and percent below poverty line variables have significant correlation coefficients for
 91

assault and violent crimes categories. The direction of relationship is positive for the

median income variable and negative for the percent below poverty line variable. This

indicates that as the median income variable increases and the percent below poverty line

variable decreases, clearance rates are likely to increase for assault and violent crime

offenses. Finally, the population variable has a significant relationship with only vehicle

theft clearances. The direction of the relationship is negative, which indicates that as the

population increases, vehicle theft clearances are likely to decrease. The proportion of

investigators variable has a significant relationship with burglary, vehicle theft and

property crimes categories for only 2005 data, but it does not have a significant

relationship with any variables from the 2006 data.

The use of IT variable has a strong correlation with the logged number of officers

and population variables. The correlation coefficients are above 0.6 and the directions of

relationships are positive indicating that larger agencies serving large populations are

more likely to score higher on the use of IT scale. In other words, larger agencies serving

large populations are more likely to use more information technologies than the small

agencies. Although relationships are weaker, the use of IT scale is significantly correlated

with other control variables with the exception of the proportion of investigators variable.

The direction of relationship with the percentage of white variable is negative indicating

that agencies whose jurisdictions have a large percentage of white population are less

likely to use information technologies. The correlation coefficients of the median income

and percent poverty variables suggest that agencies serving in more affluent communities

are more likely to score higher on the use of IT scale. The use of IT scale has an
 92

insignificant, weak and negative relationship with the proportion of investigators variable

(r=-0.08).

The strong correlation between the use of IT scale and the logged number of

officers deserves further attention. Correlation results show that large agencies are more

likely to use information technologies. Because of this relationship it may be difficult to

distinguish the impact of the use of information technology on clearance rates from the

impact of the agency size on clearance rates. Since there is a very strong correlation

between the logged number of officers and population variables (r=0.943) and both

variables have a strong relationship with the use of IT scale, it is sufficient to examine

just one of the two variables. Here we will look at the relationship between the use of IT

scale and the logged number of officers variable. Figure 11 presents the scatter plot of the

relationship between the logged number of officers and the use of IT variables.
10
8
Logged # of Officer
6
4
2
0

5 10 15 20 25 30
Use of IT
Figure 11. Scatter plot of logged number of officers and use of IT scale variables
 93

Figure 11 verifies the positive correlation between the logged number of officers

and use of IT scale variables. Figure 11 also shows that there is not sufficient number of

observations in some regions of scatter plot. There is a large empty region at the top left

portion of the scatter plot. It indicates that the sample does not have any large department

which scored low on the use of IT scale. Similarly, there is another empty region at the

bottom right portion of the scatter plot. It shows that the sample does not have any small

department which scored high on use of IT scale.

This finding shows that the data is unbalanced. Unbalanced data means that there

are unequal numbers of observations in the cells of a design (Sahai & Ojeda, 2005). In

this case, if we make a two-way table of deciles of use of IT and logged number of

officers variables, we would see that there were not equal numbers of observations in the

cells. Moreover, some cells would have zero observations. This constitutes a problem for

the multivariate analysis. Since there is not a sufficient number of small agencies that

score high and big agencies that score low on the use of IT scale, it is not possible to

control for the logged number of officers variable by just adding it into the analysis as a

control variable. In a multivariate analysis, the derived regression line would be the

product of just small agencies on one side, and it would be the product of just big

agencies on the other side. Since we will not be able to adequately control for the logged

number of officers variable, the regression coefficient of the use of IT variable is likely to

be overestimated (Gelman & Hill, 2006).

To further examine the imbalance problem and to find the corresponding areas of

number of officers and use of IT variables that drive the correlation high, the sample is

divided into deciles of the use of IT scale, and correlation coefficients are calculated. The
 94

scatter plots are presented in Appendix H. Scatter plots show that, except for the first

decile, the range of the use of IT variable is very small in the remaining nine deciles. For

example, in the second decile the minimum value of the use of IT scale is 13 and the

maximum value is 16, so its range is 3. The range of use of IT variable is smaller in other

deciles and it is 0 in the seventh and ninth deciles. For this reason, it is not very

informative to calculate the correlation coefficient for those deciles. The range of use of

IT scale in the first decile is 11. There are 18 cases in the first decile and the correlation

coefficient is 0.43. On the other hand, the imbalance problem can be seen clearly by

looking at the first and the tenth deciles. The first decile is populated by the small

departments and the tenth decile is populated by the larger ones.

Since deciles of use of IT variables did not give much information, the sample is

again divided into ten groups according to the logged number of officers variable. The

logged number of officer variable ranged between 0 and 10.52 (if an agency has only 1

officer, the logged number of officers variable becomes 0). Therefore, the boundaries of

ten groups are defined as follows: 0-09, 1-1.9, 2-2.9, 3-3.9, 4-4.9, 5-5.9, 6-6.9, 7-7.9, 8-

8.9 and 9-10.52. Scatter plots and correlation coefficients are presented in Appendix I.

Figure 12 presents the scatter plot of the logged number of officers and use of IT scale

variables, again. The logged number of officers variable is divided into ten groups and

the correlation coefficient and group sizes are provided on the graph. Figure 12 also

presents a box plot comparison of the use of IT variable in ten groups of the logged

number of officers variable.

 95

10
N=4

Logged # of Officer N=3

8 r=-0.027 N=23

r=0.073 N=57
6

r=-0.57 N=16
r=0.55* N=14
4

r=0.32* N=23
r=0.42* N=30
2

r=-0.016 N=22

N=3
0

5 10 15 20 25 30
Use of IT
1 2 3 4 5 6 7 8 9 10
Logged # of Officer

5 10 15 20 25 30
Use of IT

Figure 12. Scatter plot of logged number of officers and use of IT scale variables, and
box plot comparison of use of IT scale across ten groups.

As shown on the first graph, the correlation coefficients are stronger in the third,

fourth, fifth and sixth groups (it is significant for the third, fourth and fifth groups). The

direction of relationship of the correlation coefficient is generally positive, but it is

negative in the second, sixth and eighth groups. The scatter plot shows that correlation is

mainly driven by the third, fourth and fifth groups. The second graph clearly shows that

the distribution of the use of IT variable differs across the ten groups of the number of

officers variable. Positive correlation is also visible in the second graph. This graph also
 96

shows that distribution of the use of IT variable is overlapping in the 6th, 7th, 8th and 9th

groups. The first group, on the other hand, is clearly different than other groups.

For the final analysis of the relationship between the number of officers and use

of IT scale variables, a regression model was fitted in which use of IT scale is the

dependent variable and the logged number of officers variable is the independent

variable. The goal of regression analysis is to see the association between two variables.

Regression analysis was conducted for the groups of logged number of officer variable

that have at least 20 observations and for the entire sample The regression formula was

expressed as follows:

Use of IT Scale = a + β 1 ( Logged 6umber of Officer ) + e

Regression coefficients with brief model information are provided in Table 6. The

direction of relationship in all cases is consistent with the correlation coefficients. The

regression coefficient is large for the third and fourth groups and for the entire sample,

but it is smaller than one for the remaining groups. The regression model and coefficients

are significant for the third group and the overall model. Since the independent variable

was a logged variable, its interpretation is made in terms of percent change (Gelman &

Hill, 2006). Therefore, the third model suggests that a one percent increase in the number

of officers increases the use of technology scale by 0.068 units. Similarly, the entire

sample model suggests that a one percent increase in the number of officers increases use

of technology scale by 0.015 units. Although the impact is fairly small, it reveals the

existence of a relationship between the two variables which is important for the next part

of this study.
 97

Table 6
Regression Coefficients of Logged 6umber of Officers Variable Predicting Use of IT
Scale
2 3 4 7 8 All
Intercept 14.743 -1.175 3.380 17.108 25.543 12.708
# of Officer -0.253 6.848* 4.359 0.919 -0.262 1.56***
F 0.005 5.857* 2.334 0.293 0.015 125.1***
2
R -0.050 0.143 0.057 -0.013 -0.047 0.39
6ote. ***p< 0.001, **p< 0.01, *p< 0.05

In sum, the bivariate analysis revealed two problems. The first problem is the lack

of balance in the sample. As shown above, there are not enough departments which are

small but scored high on use of IT scale and vice versa. The second problem is the strong

correlation between the use of IT scale and the logged number of officers variables.

Multivariate Analysis

In this part, multiple regression analysis is employed with the ordinary least

squares (OLS) estimation method to investigate functional relationships among variables.

Multiple regression permits the estimation of the effect on the dependent variable of

changing one variable while holding the other regressors constant. Linear functional form

is preferred because examination of scatter plots of 2005 and 2006 clearance rate

variables and the use of IT variable did not reveal any nonlinear pattern in the

relationships between response and predictor variables. Two linear models are specified.

The first model is the reduced model with the square rooted clearance rate variable as the

response variable and the use of IT scale and number of officers variables as the predictor

variables. The regression equation of the reduced model is expressed as follows:

 98

Clearance Rate = a + β 1 (Use of IT Scale ) + β 2 ( Logged 6umber of Officer ) + e

The second model is the complex model with the use of IT scale as the independent

variable controlling for the six variables discussed above, except for the population

variable. The regression equation of the complex model is expressed as follows:

Clearance Rate = a + β1 (Use of IT Scale) + β 2 ( Logged 6umber of Officers ) +

β 3 ( Median Income) + β 4 ( Poverty ) + β 5 ( P. of Invest.) +
β 6 ( Race) + β 7 (Crime Rate) + e

The population variable was excluded from the model because of its strong correlation

with the number of officers variable (r=0.943); therefore, controlling for either one would

be sufficient for the analysis. The median income and percent below poverty line

variables have also a strong correlation with each other (r=-0.719). However, both

variables were kept in the model because besides measuring the socio-economic

condition of the population, the median income variable is also used as a proxy to the

resources available to the police departments. Therefore, both variables are important for

correctly measuring the environmental impact on clearance rates.

Three groups of regression analysis were conducted. First, regression parameters

were estimated by including all the departments into the analyses after dropping eight

departments scored below 10 on use of IT scale. As presented in the bivariate analysis,

those eight cases are small departments. Their median of the logged number of officers

variable is 1.59, and the largest logged number of officers variable in that group is 3.25.

This shows that this portion of the dataset is very much unbalanced. Since they cannot

contribute much to the analysis, those eight departments were dropped from the sample.

Each of the twenty clearance rate variables (ten variables from each year) were plugged

into reduced and complex models as dependent variables, so forty models were
 99

estimated. In the second step, police departments were divided into two groups as

moderate IT and high IT departments, and reduced and complex models were fitted to

both moderate IT and high IT departments. Therefore, eighty models were estimated in

this step. In the third step, the sample was divided into ten groups by the number of

officers variable as described above, and regression parameters were estimated for the

groups with enough sample size separately.

The results of regression analysis involving the whole sample are presented in

Appendix J. F tests were significant for 12 models of 2005 and 12 models of 2006 crime

data. Therefore, significant F test results reject the null hypothesis that all the regression

coefficients in the regression equation are zero. Adjusted R2 values are generally smaller

than 0.1 indicating that models can generally explain less than 10 percent of variance in

the dependent variable. Except for the rape models in both years’ data, complex models’

adjusted R2 values have improved compared to the reduced models’ adjusted R2 values.

Figure 13 shows the regression coefficients of the use of IT scale (R and C letters

at the end of offense types represent reduced and complex models). The bar graph shows

that the magnitude of regression coefficients ranges between 0.001 and 0.014. It is

generally higher in violent crime models than the property crime models, and it is higher

in 2006 data than the 2005 data. The use of IT scale variable is significant in only four

models. The direction of relationship is always positive indicating that high scores in the

use of IT scale are associated with higher clearance rates, but the impact is fairly small.

For example, the complex model of robbery fitted to the 2006 data has the highest

regression coefficient of the use of IT scale which is 0.014. Since the dependent variable

is square rooted, and its interpretation is difficult in this form, the regression coefficient is
 100

transformed back by taking its square (Cohen, 2003). As a result; one unit change in the

use of IT scale corresponds to 0.000196 or % 0.02 change in clearance rates.

2005 2006
0.016
Magnitude of Use of IT Regression

0.014 *

0.012 *
Coefficient

0.01
0.008
0.006 *
*
0.004
0.002
0

Figure 13. Regression coefficients of Use of IT scale estimated from 2005 and 2006 data
(***p< 0.001, **p< 0.01, *p< 0.05)

The agency size variable was significant in 12 out of 40 models, and surprisingly

its direction of relationship is generally negative. The percent white variable is significant

in 13 out of 20 models. The direction of relationship is positive indicating that clearance

rates are increasing as the percentage of white population increases. The proportion of

investigators variable is significant in only five out of 20 models, and its direction of

relationship is not consistent. The crime rate variable is significant in 6 out of 20 models

and its direction is generally positive indicating that clearance rates increase as crime

rates increase. Finally, it would be informative to look at details of several models. Table

7 presents murder and vehicle theft models of the 2005 and 2006 years.
 101

Table 7
2005 and 2006 Murder and Vehicle Theft Models
Murder 05 Murder 06 Vehicle 05 Vehicle 06
R C R C R C R C
Intercept 0.769 0.008 0.962 0.603 0.442 0.087 0.471 0.395
(0.202) (0.582) (0.16) (0.482) (0.075) (0.22) (0.075) (0.217)
Use of IT 0.002 0.004 0.007 0.006 0.006 0.005 0.005 0.002
(0.007) (0.008) (0.006) (0.007) (0.004) (0.004) (0.004) (0.004)
# of -0.015 0.011 -0.06** -0.032 -0.035*** -0.014 -0.037*** -0.012
Officer (0.022) (0.024) (0.017) (0.021) (0.009) (0.011) (0.01) (0.012)
Median 0.000 0.000 0.000 0.000
Income (0.000) (0.000) (0.000) (0.000)
% Below 0.007 0.000 0.003 -0.003
Poverty (0.011) (0.009) (0.005) (0.005)
% White 0.007* 0.004 0.002 0.002
(0.002) (0.002) (0.001) (0.001)
Prop. Of 0.076 -0.023 0.092 0.119
Inves. (0.177) (0.145) (0.056) (0.054)
Crime rate 0.126 -0.045 -0.007 -0.006
(0.363) 0.257 (0.004) (0.004)

F 0.252 1.429 5.641* 2.726* 7.114** 4.2*** 8.27*** 4.559***

DF Reg 2 7 2 7 2 7 2 7
DF Res 107 101 105 99 158 152 154 148
2
Adj. R -0.013 0.027 0.08 0.102 0.071 0.124 0.085 0.139
6ote. ***p< 0.001, **p< 0.01, *p< 0.05

Except for 2005 murder models, the F test is significant in all the models.

Complex models have higher adjusted R2 values. The use of IT variable is not significant

in any of the models. Except for the 2005 complex murder model, the number of officers

variable is negative and it is significant in three reduced models. The regression

coefficients of median income, percent below poverty line and percent white variables

are very small, their direction of relationship is not consistent and only the percent white
 102

variable is significant in one model. Finally, the crime rate variable also has a small

regression coefficient with generally negative relationship with clearance rates.

In sum, multivariate analysis of 2005 and 2006 crime and clearance data showed

that regression coefficient of use of IT variable was significant in only 4 out of 40

models. Although complex models generally improved the adjusted R2 values,

explanatory power is generally low in all models. Because of the imbalance problem

presented in the previous part, the interpretation of regression coefficients is problematic.

Since multivariate model cannot adequately control for the impact of logged number of

officers variable, regression coefficient of use of IT variable might be overestimated.

In the second stage of multivariate analysis, the sample was divided into two,

groups (after dropping 8 departments that scored below 10 on the use of IT scale) from

the midpoint of the use of IT scale. The first group consisted of the departments that

scored between 10 and 19, and the second group consisted of the departments that scored

between 20 and 29. Since the departments scored low on use of IT scale was dropped,

these groups are called as moderate and high IT groups. Two groups were compared to

each other before conducting any multivariate analysis. Box plot comparisons of

moderate and high IT departments (Appendix G) revealed that there is not much

difference between moderate and high IT departments in terms of crime and clearance

rates.

Figure 14 presents the box plot comparison of 2005-2006 all crime and clearance

variables and the six control variables. Box plot comparison shows that moderate and

high IT departments are very similar in terms of crime and clearance rates. There is an

important difference between the logged number of officers variables of moderate and
 103

high IT departments. High IT departments are generally the larger ones, and moderate IT

departments are generally the smaller ones. Although they have some overlap, the

population, median income and percent white variables also look different. High IT

departments are generally serving larger populations whose median income is somewhat

bigger than the jurisdictions that moderate IT departments are serving. On the other hand,

the percentage of white people is bigger in moderate IT departments than in the high IT

departments. High and moderate IT departments look similar in terms of the proportion

of investigators and poverty variables,

2005 All Crimes CLearance Rate % Below Poverty

H

H
M

0.0 0.2 0.4 0.6 0 10 20 30

2005 All Crimes Rate Median Income
H

H
M

0 50 100 150 200 20000 40000 60000 80000 100000 140000

2006 All Crimes CLearance Rate % White
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 20 40 60 80 100

2006 All Crimes Rate Logged Population
H

H
M

0 50 100 150 200 6 8 10 12 14 16

Proportion of Investigators Logged # of Officer
H

H
M

0.0 0.5 1.0 1.5 2.0 2.5 3.0 2 4 6 8 10

Figure 14. Box plot comparisons of variables between moderate and high IT departments

Multiple regression analysis was conducted in moderate and high IT departments,

and regression coefficients were estimated for reduced and simple models for both 2005
 104

and 2006 data. Eighty models were estimated. Regression coefficients, F values and R2

values are given in Appendix K. F test results showed that none of the models were

significant in 2005 moderate IT data, and 4 models (all reduced) were significant in 2006

moderate IT data. On the other hand, 12 models (4 reduced, 8 complex) were significant

in 2005 high IT data and 10 models (4 reduced and 6 complex) were significant in 2006

high IT data. Adjusted R2 values showed that the complex model did not increase

explanatory power for moderate IT departments’ data. In general, adjusted R2 values of

complex models are smaller than adjusted R2 values of reduced models. On the other

hand, adjusted R2 values of complex models increased in high IT department data. Figure

15 presents the regression coefficients of use of IT variable.

Moderate IT 05 Moderate IT 06 High IT 05 High IT 06

0.090

0.080 **
Magnitude of Use of IT Regression

0.070

0.060

0.050 *
Coefficient

0.040
*
0.030

0.020

0.010

0.000

-0.010

-0.020

Figure 15. Regression coefficients of Use of IT scale of moderate and high IT

departments estimated from 2005 and 2006 data (***p< 0.001, **p< 0.01, *p< 0.05)

As marked in the graph, the use of IT variable is significant only in three models.

Moreover, it is interesting that the direction of relationship becomes negative in some

models, particularly in 2006 reduced burglary, complex burglary and reduced vehicle
 105

theft models. On the other hand, there are more negative coefficients in high IT

department models. The coefficients of the number of officer variable are also weak. It is

significant in 11 out of 80 models. The direction of relationship is again negative in 67 of

80 models. Other variables’ coefficients are very small again, and generally do not have a

constant direction of relationship.

Detailed model information of the complex models of murder and vehicle theft

clearance rates are presented in Table 8. F tests are significant for only two models, and

moderate IT departments have higher adjusted R2 values in murder models, but high IT

departments have higher adjusted R2 values in vehicle theft models. As presented in

Table 8, all the coefficients are small again, and generally inconsistent in terms of the

direction of relationship. The use of IT variable has a positive impact on moderate IT

departments, but it sometimes has a negative relationship in high IT departments. The

same condition is visible for the number of officers variable. The percent white variable

always has a positive coefficient and except for one case, the crime rate variable always

has a negative coefficient.

Table 8
Complex Models of Murder and Vehicle Theft Clearance Rates Fitted to 2005-2006
Moderate and High IT Departments
Murder 05 Murder 06 Vehicle 05 Vehicle 06
Mod. High Mod. High Mod. High Mod. High
Intercept -1.911 -0.023 -0.712 0.264 -1.008 0.611 0.284 0.573
(3.172) (0.627) (1.857) (0.538) (0.571) (0.244) (0.583) (0.252)
Use of 0.049 -0.007 0.07 0.009 0.032* 0.002 0.001 -0.004
IT (0.051) (0.012) (0.035) (0.011) (0.014) (0.007) (0.016) (0.007)
# of 0.047 0.024 0.016 -0.041 -0.007 -0.021 -0.011 -0.017
Officer (0.083) (0.023) (0.044) (0.028) (0.026) (0.012) (0.028) (0.012)
Median 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Income (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
 106

Table 8 (Continued)
% Below 0.009 0.015 -0.003 0.009 0.031** -0.012* 0.006 -0.006
Poverty (0.043) (0.013) (0.028) (0.01) (0.01) (0.005) (0.01) (0.005)
% White 0.014 0.008** 0.008 0.005* 0.004 0.001 0.003 0.002
(0.017) (0.002) (0.008) (0.002) (0.003) (0.001) (0.004) (0.001)
Prop. Of 0.438 0.057 0.373 -0.06 0.017 0.122* -0.043 0.115*
Inves. (1.538) (0.174) (0.763) (0.144) (0.162) (0.052) (0.165) (0.053)
Crime 1.994 -0.026 -0.012 -0.013 -0.036* -0.003 -0.014 -0.005
rate (2.044) (0.368) (0.92) (0.266) (0.014) (0.003) (0.016) (0.003)

F 0.453 1.477 2.226 1.643 2.162 6.873*** 1.01 6.369***

DF Reg 7 7 7 7 7 7 7 7
DF Res 10 83 13 78 42 102 43 97
2
Adj. R -0.29 0.035 0.3 0.05 0.142 0.274 0.001 0.27
6ote. ***p< 0.001, **p< 0.01, *p< 0.05

Analysis of moderate and high IT groups did not provide much support for the

relationship between use of IT and clearance rate variables either. However, as presented

above, imbalance is still a problem for moderate and high IT groups. For that reason, the

sample is again divided into ten groups by the number of officers variable as done in the

bivariate analyses part. Regression coefficients were estimated for the groups that have at

least 20 observations in dependent variable. As a result, regression coefficients were

estimated for 35 reduced and complex models in 2005 data and 33 reduced and complex

models in 2006 data. Figure 16 and Figure 17 present the use of IT coefficients. A table

of regression coefficients with brief model information is presented in Appendix L

 107

Group 3 Group 4 Group 7 Group 8

0.07

Magnitude of Use of IT Regression 0.06

0.05

0.04
Coefficient

0.03 * *

0.02 *
* * *
*
0.01

-0.01

-0.02

Figure 16. 2005 - Regression coefficients of Use of IT scale after dividing sample into
ten groups (***p< 0.001, **p< 0.01, *p< 0.05)

Group 2 Group 3 Group 4 Group 7 Group 8

0.05
Magnitude of Use of IT Regression

0.04

0.03
Coefficient

0.02
*

0.01

-0.01

-0.02

Figure 17. 2006 - Regression coefficients of Use of IT scale after dividing sample into
ten groups (***p< 0.001, **p< 0.01, *p< 0.05)

There are 7 models fitted to the third group of 2005 data and 5 models fitted to the

third group of 2006 data. As presented above, the use of IT scale and number of officers

variables have strong and positively significant relationships in this group. Regression
 108

coefficients are consistently positive in the 2005 data and six of them are significant. In

fact, they are the only significant coefficients in the 2005 data. However, the regression

coefficients are not consistent in the 2006 data, and they become negative for vehicle

theft models. There are 8 models fitted to the fourth groups of 2005 and 2006 data. As

presented above, the use of IT scale and number of officers variables have a significant

relationship in the fourth group as well. The fourth group is consistent in terms of the

direction of regression coefficients in that they are positive in both 2005 and 2006 data;

however, none of them is significant. The seventh group also gives consistent results in

terms of direction of regression coefficients, but most of the coefficients are negative in

this group. Ten models were fitted to the seventh groups of 2005 and 2006 data. Except

for robbery and vehicle theft models, all the regression coefficients are negative. Finally

the eighth model is also consistent. Except for the vehicle theft model, all the coefficients

are positive in 2005 and 2006 data. The regression coefficients of robbery and violent

crime are significant in 2006 data. In sum, the examination of regression models did not

reveal any consistent pattern in terms of the relationship between use of IT variable and

clearance rates. Regression coefficients are small and the impact of the use of IT scale

variable changes within and among groups.

In conclusion, in this chapter we tested the relationship between use of IT and the

clearance rate variables. Because of the imbalance problem, dataset is divided into groups

and the relationship between use of IT and the clearance rate variables was investigated

in those groups as well. However, multivariate analyses did not reveal any significant

relationship between use of IT variable and the clearance rate variables. The use of IT
 109

variable was significant in only 16 out of 188 models. Moreover, the direction of

relationship is not consistent across the models.

Use of IT items.

As a final analysis, each individual item of the use of IT scale is analyzed. As

presented in the methodology chapter, there are 29 items in the use of IT scale. All of the

items are binary variables where zero means that the department does not have that

technology and one means that the department has it. To make the following discussion

more understandable the group that does not have the technology is called non-

technology group and the group that has the technology is called technology group.

Since previous analysis results showed that the use of IT scale is positively

correlated with the agency size variable, it was highly possible that only certain types of

agencies might use those individual technologies. To investigate this claim, departments

are divided into two groups as technology and non technology groups for each item.

Then, the technology and non-technology groups of each of the use of IT items were

compared with each other. Box plots and t-test analysis were used for comparison

purposes.

The examination of box plots revealed that, in general, the technology group is

not very much different from the non-technology group for each of the use of IT items in

terms of clearance rates. This finding is consistent for raw and square rooted clearance

rate variables of both 2005 and 2006 data. On the other hand, crime rate variables showed

a different pattern. For some use of IT items, departments that have the technology

usually have higher crime rates than the departments that do not have that technology.
 110

Box plot comparison of control variables revealed that departments are generally

different from each other in terms of the logged number of officers variable. In 26 of 29

items, the technology group’s median of the logged number of officers variable is larger

than that of the non-technology group. Box plot comparisons showed that there is more

overlap between departments in only three items (Appendix M), and t-test scores were

not significant for these groups showing that the logged number of officers variable is not

different with respect to these items. These items are availability of cellular/mobile

phones, availability of RISS, and availability of NCIC system. This finding shows that

sampled departments are not different from each other in using those three information

technologies.

On the other hand, the box plot comparison revealed that the logged number of

officers variable of technology and non-technology groups almost does not have any

overlap for five use of IT items (Appendix N). In addition, t-test statistics were

significant for them. Those items are availability of blackberry technology, availability

MDC/MDT (mobile computing) technology, availability of fingerprints on computer,

availability of mug shots on computer and availability of stolen vehicles on computer.

The median of the logged number of officers variable is larger in the technology group

for those five items indicating that generally larger agencies have those technologies. Box

plot comparisons of the use of IT items that have good overlap and very poor or no

overlap are presented in Figure 18.

 111

Cellular/Mobile Phones RISS

Y
N

N
0 2 4 6 8 10 0 2 4 6 8 10
NCIC PDA/Handheld Computer
Y

Y
N

N
0 2 4 6 8 10 0 2 4 6 8 10
MDC/MDT Fingerprints on computer form
Y

Y
N

N
0 2 4 6 8 10 0 2 4 6 8 10
Mug Shots on computer form Stolen Vehicles on computer form
Y

Y
N

N
0 2 4 6 8 10 0 2 4 6 8 10

Figure 18. Use of IT items that have good overlap and very poor or no overlap

The remaining 21 items are in between these two groups. Box plot comparisons

show that there is some overlap between technology and non-technology groups for those

items. However, they still look different from each other. In addition, the t-test statistic is

significant for 20 of them. Again, the median of the logged number of officers variable is

larger in the technology group than in the non-technology group indicating that,

generally, larger agencies have that technology.

Regression analysis was conducted to examine the relationship between use of IT

items and clearance rates. The regression equation is expressed as follows:

Clearance Rate = a + β1 (Use of IT Items) + β 2 ( Logged 6umber of Officers) +

β 3 ( Median Income) + β 4 ( Poverty) + β 5 ( P. of Invest.) +
β 6 ( Race) + β 7 (Crime Rate) + e

Each of the 20 dependent variables was regressed on each of the 29 use of IT item

variables. As a result, 580 regression models were fitted to 2005 and 2006 data. Bar

graphs of item coefficients were presented in Appendix O. Items in which there is an

overlap between departments for the number of officers variable (the first group), items
 112

in which there is poor or no overlap (the second group), and items in between these two

groups (the third group) are separated with a dotted line on the graphs. Bar graphs of

murder and vehicle theft models are also presented in Figure 19 below.

05 Murder 06 Murder
0.4
Magnitude of Use of IT Regression

0.3
*
0.2
Coefficient

0.1

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.1

-0.2

-0.3

05 Vehicle Theft 06 Vehicle Theft

0.2
Magnitude of Use of IT Regression

0.15

0.1 *
*
*
Coefficient

0.05

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.05

-0.1
*
-0.15
*
-0.2
Figure 19. Bar graphs of use of information technology items’ regression coefficients for
murder and vehicle theft offenses (***p< 0.001, **p< 0.01, *p< 0.05)

The first bar graph shows the murder model. Only the 27th item (availability of

sex offender data on computer form) coefficient is significant; however, its direction of
 113

relationship is not consistent between 2005 and 2006 data. The existence of negative

coefficients indicates that the availability of some information technologies may be

associated with the decrease in clearance rates. The largest positive coefficient is 0.29

and the largest negative coefficient is -0.24. The direction of relationship is consistent

between 2005 and 2006 coefficients in the first group which is the balanced group.

Except for one item (29th item), this pattern is the same for the second group which is the

unbalanced group. However, the direction of relationship is inconsistent for 9 items of the

third group.

The second graph shows the vehicle theft model. Coefficients are smaller than -

0.15 and +0.15. Five of 58 regression coefficients are significant. Those items are

availability of GPS (2), Web Pages / Web-based technology (7), ViCAP (12), and crime

reports (15). The crime report item is significant for both 2005 and 2006 data; however,

its direction of relationship is negative indicating that the existence of crime reports in

computer format reduces clearance rates. The direction of relationship of ViCAP

coefficients is inconsistent between 2005 and 2006 data. The pattern of direction of

relationship is similar to the murder model for the first, second and third groups. As

presented in the literature review chapter, mobile computing systems were found (5th

item) to be effective in clearing vehicle theft crimes in two studies (Meehan, 1998; Nunn,

1994). However, the regression coefficients of MDC/MDT systems are not statistically

significant in this study.

In sum, the relationship between each use of IT items and clearance rates is tested

in this supplementary section. 580 regression analyses were conducted and use of IT
 114

items were significant in 47 or 8.1% of them. Moreover, none of the items had a

consistent pattern of relationship.

 CHAPTER 5: CONCLUSION, LIMITATIONS AND POLICY IMPLICATIONS

Criminal investigation is one of the most important tasks of law enforcement

agencies. Despite of this importance it is one of the least studied topics of criminal justice

(Horvath, et al., 2001). Researchers’ interest in this field intensified during the early

1970s, after the release of the President’s Commission on Law Enforcement and

Administration of Justice Report. Early studies generally found that criminal

investigations are not effective in arresting criminals or clearing the crimes (e.g., Chaiken

et al., 1976; Greenwood 1970). Particularly, the Chaiken et al. (1976) study found that

organizational and procedural differences in criminal investigations among police

departments did not have an impact on arrest or clearance rates. Although later studies

have challenged some of its findings, the Chaiken et al. (1976) study has remained as one

of the most cited studies in this field, and its critiques of criminal investigation have been

influential in the field for a long time.

Criminal investigation studies have explored the criminal investigation process

and analyzed the impact of internal or external factors on the success of criminal

investigations. Researchers have pointed out several factors as the determinants of

successful criminal investigation. For example, the quality of preliminary investigations,

organizational structure of the investigative unit, agency size and workload of

investigators were some of the major factors analyzed in that literature. Studies have

generally reported mixed findings about most of the determinants of criminal

115
 116

investigation. However, the quality of preliminary investigation and the information

collected in this stage were consistently found as the most important factors of a

successful criminal investigation. For that reason, it was argued that the single most

important determinant of whether or not a case would be solved was the information the

victim or witness supplied to the responding patrol officer (e.g. Burrows, 1986; Chaiken

et al., 1976; Wilmer, 1970).

The importance of information for criminal investigations has led some

researchers to redefine criminal investigation as a kind of “information work” in which

investigators are continuously trying to reach to new information either by uncovering

new facts about the case, or by analyzing, interpreting, and ordering the existing

knowledge (Innes, 2007; Maguire, 2003). Since criminal investigation is a kind of

information work, it has a natural fit with information technologies. As Hogan and

Radack (1997) defined them, information technologies are blend of hardware and

software used for storage, processing, transfer, display, management, organization, and

retrieval of information. Therefore, information technologies may help investigators to

collect and store more information about a case, to analyze and interpret a large amount

of information more rapidly and to organize and present resulting information for various

purposes. To that end, information technologies may provide criminal investigators with

new tools to increase their capabilities of scanning, analysis and ordering of information.

In fact, information technologies are widely employed by the criminal justice

community especially in law enforcement services (Dunworth, 2004). An exploration of

law enforcement information technologies revealed that they are used in criminal

investigations mainly for four purposes: records management, crime analysis,

116
 117

communication and mobile computing and forensic analysis. In modern law enforcement

information systems, however, these four functions are integrated to increase their

effectiveness. For example, crime analysis systems generally work together with records

management systems. Similarly, communication and mobile computing technologies are

used mainly to access record management systems and to extend crime analysis activities

to the patrol.

On the other hand, despite their popularity among practitioners and policymakers,

law enforcement information systems, their nature, use and user characteristics and

impacts on police work constitute another understudied topic in criminal justice literature.

Moreover, researchers have used many different outcome variables to measure the

success of law enforcement information systems. For example, some studies measured

the effectiveness of a mobile computing system by the number of stolen vehicles

recovered (Braga and Pierce, 2004; Nunn, 1994), but in another study effectiveness is

measured by the number of inquiries made by officers by using the system (Marshall,

1998), and in another it is measured by information technologies’ contribution to the

objectives of problem-oriented policing perspective (Nunn & Quinet, 2002). As a result,

it is difficult to compare different studies conducted in this field and to arrive at a

conclusion about the impact of information technologies in the law enforcement world.

This study contributes to both criminal investigation and law enforcement

information technologies literature. Within the criminal investigation area, this study

investigated whether there is a relationship between modern law enforcement information

technologies and outcome of criminal investigations, namely clearance rates. Within the

law enforcement information technologies area, this study investigated the

 118

appropriateness of the clearance rate variable as an outcome measure. Besides these two

empirical contributions, this study also reviewed and presented the current status of

research in these fields. This study used the data collected from a national sample of

police departments by the Impact of Science and Technology Survey, the UCR program

data and U.S. Census data.

The first conclusion of this study is about the impact of information technologies

on police work. The relationship between clearance rates and the use of the information

technology variable was investigated both on the whole dataset and subgroups of dataset

that were suitable for multivariate analysis. As a result, 188 regression models were

estimated with 2005 and 2006 data. Analysis results showed that the use of information

technology scale was statistically significant in only 8.5% of the models. In addition to

this, each of the use of information technology scale item was used as independent

variable in multivariate analysis, and 580 models were estimated. The regression

coefficient of use of information technology items was statistically significant in only

8.1% of the models. This finding supported the argument that there may not be a

relationship between clearance rates and the use of information technology variables.

However, it must be noted again that this study is a one-group post-test only

design. Since this design lacks any pretest baseline or a comparison group, this study

does not control for temporal order and nonspuriousness; therefore, it cannot provide

much information about causal relationship between independent and dependent

variables. All one can say with the one-group post-test only design is that there is or there

is not a correlational relationship between variables, and a correlational relationship only

says that two things perform in a synchronized manner. As a result, for the majority of
 119

the analyses, this study did not find a significant correlational relationship between use of

information technologies and the clearance rates.

Despite of the design considerations, this finding argues that there may not be

really a relationship between use of information technologies and clearance rates.

Moreover, this study also controlled for a set of other variables that might be influential

on the outcome of criminal investigations. Those variables were specified after a careful

review of available literature and they were department size, proportion of officers

working at investigative functions, percent of people living under poverty threshold,

median income, percent white and crime rates. Analysis results showed that clearance

rates do not have a consistent relationship with any of the control variables either.

Another important conclusion of this study is a methodological contribution to the

criminal justice research. Data analysis results revealed that there was a strong correlation

between use of information technology, which was an index variable constructed from 29

information technology items, and the agency size variable, which was measured by the

number of sworn officers working in the department. On the other hand, further

examination of this relationship by looking at the scatter plot of two variables yielded

more concerning results. Besides the strong correlation between the two variables, the

scatter plot showed that there is a great difference between small and large agencies in

terms of information technology use. There were no large departments that scored low on

the use of information technology scale and there were almost no small departments that

scored high on the use of information technology scale. In other words, the data was

unbalanced.
 120

The imbalance problem became more visible as the sample was divided into the

deciles of the use of information technology scale. There were only small departments in

the first decile of the use of information technology scale, and there were only large

departments in the tenth decile. This meant that the regression coefficients would be the

product of only small departments on one end, and the product of only the large

departments on the other end. As a result, it would be impossible to distinguish the

impact of use of information technology from the impact of agency size. Because of the

differences between large and small law enforcement agencies, this problem can arise in

most of the datasets of cross-sectional studies about police departments, and it cannot be

remedied by just plugging the agency size variable into the equation as a control variable.

If the data is unbalanced, it is quite possible that the regression coefficient of the

independent variable would be overestimated. For this reason, researchers conducting

observational studies on police departments should further investigate the correlation

between the independent variable and the agency size variable before conducting any

multivariate analysis and arriving at some conclusions based on that analysis.

A final conclusion of this study is related to the DeLone and McLean Information

System Success model’s implementation to law enforcement agencies (1992; 2003). As

presented in the literature review chapter, the DeLone and McLean model identifies six

levels where information technology success can be measured and proposes a

relationship among them. Those success levels are information quality, system quality,

service quality, use-intention to use, user satisfaction and net benefits. Because of the

complexity of the model, previous studies have tested some parts of it and generally

reported supportive findings. This study tested the relationship between use-intention to
 121

use and net benefits and did not find a relationship between them. This study measured

the system use by looking at the availability of particular information technologies at

responding agencies because as DeLone and McLean argued managerial decision to

invest in information technologies shows intention to use the system. However, this

measure lacks the variability in the quality and intensity of use. Therefore, before arriving

at a conclusion about the relationship between use-intention to use and net benefits, it

must be tested with other measures of use-intention to use that have greater variability in

the quality and intensity of use.

In addition to this, Net Benefits level should also be reconsidered. DeLone and

McLean defined the Net Benefits success level broadly, and argued that any kind of

improvement, depending on the context of the information system, can be counted as a

benefit (DeLone & McLean, 2003). Therefore, this study assessed the adequacy of

clearance rates as a Net Benefit accrued from the use of information technologies by law

enforcement agencies particularly for investigative activities. However, as explained

above, this study found no relationship between use of information technologies and

clearance rates. Therefore, it may be argued that the clearance rate variable may not be

the correct outcome variable for measuring Net Benefit in the law enforcement context.

Nevertheless, information technologies may still be influential on some other aspects of

police departments. For example, information technologies may have an impact on the

management of organizational resources. As Nunn (2001) argued, information

technologies can make an impact on efficacy and efficiency of use of organizational

resources in police departments if they are used correctly. Therefore, information

 122

technologies’ impact on the management of organizational resources may be analyzed as

a potential net benefits variable.

In order to have a complete view of police information technology systems, other

success levels should also be analyzed. As discussed in the literature review chapter,

system quality variables measure the adoptability, availability, usability, reliability and

the response time of the systems. These and other engineering-oriented variables can be

used to test the system quality of police information technology systems. Information

quality variables measure the completeness, ease of understanding, personalization,

relevance and security of information. In addition to these variables, accuracy and

privacy of information should also be considered for the information quality of police

information technology systems. The service quality and user satisfaction can also be

measured with the variables offered by DeLone and McLean.

Limitations.

This study has several important limitations that must be acknowledged. The first

limitation of this study is about its independent variable. The use of information

technology variable measures whether the responding police departments have or do not

have particular kind of information technologies. It does not measure the variability in the

quality and intensity of use; in other words, it does not measure how well those

technologies are used by the police departments. Effective implementation and use of

information systems often require organizational change; however, police departments

are generally resistant to change, and resistance to change prevents the realization of

information technologies’ benefits (LeBeuf, 2001; Manning, 2001). In addition to this,

the performance of information technologies depends much on the performance of its

 123

user. If users were not adequately trained to use the system, then it may not be possible to

get benefit from the system (LeBeuf, 2001). In sum, there may be differences between

this study’s departments in terms of implementation and use of information technologies.

Departments that are effectively implementing and using IT may get more from their IT

systems than the departments with poorly implemented and used IT systems.

Another important limitation of this study is the imbalance problem. As presented

in the results chapter, the Impact of Science and Technology Survey dataset is

unbalanced. There are no large agencies scoring low on use of information technology

scale, and there are almost no small agencies scoring high on information technology

scale. Because of the imbalance problem, it is not possible to distinguish the impact of

agency size from the impact of use of information technology in multivariate analysis.

Overestimation of the regression coefficient of independent variable is a possible

outcome of this problem.

Finally, as presented in the methodology chapter, UCR data introduces some

important problems that should be noted also here. First of all, the UCR reports Part I

crimes that are known to the police. It does not report Part II crimes, and it is well known

that some criminal activities are not even reported to the police. Related to these

problems, participation to the UCR program is voluntary. Although the UCR program has

achieved the participation of more than 90% of all police departments, sometimes some

agencies cannot provide data due to various organizational problems. In order to

eliminate the missing data problem, the FBI imputes values for the missing pieces of

data. In addition, there is no monitoring in place to ensure accurate reporting. For that

reason, UCR data is open to manipulation by participating agencies. It is argued that in

 124

order to look better in the statistics, some participating agencies may choose not to report

all of the crimes occurring in their jurisdictions (Chambliss, 2001). The UCR’s hierarchy

rule is another problem affecting reported crime. According to the UCR Program’s

Hierarchy Rule, if more than one offense is committed in an incident, only the most

serious one is reported, and related offenses are not reported. The hierarchy rule was

implemented to prevent multiple reports of the same offense, but it also conceals related

offenses. As a result, it can be argued that the UCR does not provide a complete picture

of the crime problem in the U.S.

In addition to crime-reporting issues, the UCR’s clearance data also have some

problems. As presented in the literature review chapter, a law enforcement agency reports

that an offense is cleared by arrest, or solved for crime reporting purposes of the UCR,

when all of the following conditions have been met for at least one person: 1) arrested, 2)

charged with the commission of the offence, and 3) turned over to the court for

prosecution. However, arrest data do not provide the real number of perpetrators who are

apprehended by the police because arrested individuals might be found innocent after the

trial and are released by the court. Related with this problem, an offense that is

committed in one year can be reported as cleared in later years. Therefore, clearance rates

do not reflect how much of a crime that is committed in a particular year is cleared by the

police. It only shows the clearance achieved in a year proportional to the crimes

committed in that year. Despite all of these problems, some researchers argue that the

UCR is a valid source of criminal statistics in the U.S., and it continues to be an

important source for criminal justice research (Gove et al., 1985;Hindelang, 1974; Pare et

al., 2007)
 125

Policy implications.

This study showed that issues related to the implementation and use of

information technologies in law enforcement agencies require more attention. As outlined

above, the lack of relationship between use of information technologies and the clearance

rates may be caused by poor implementation and use of information technologies by

police departments. In order to investigate this claim and to ensure proper

implementation and use of information technologies by police departments, more process

evaluation studies should be accompanied by information technology projects. These

evaluations would also be critical for future investments in this field because they would

help to separate promising information technologies from the problematic ones and to

guide policymakers in their choices.

Another policy implication related with the first one is the need for more attention

to human factors in planning and designing information technologies. A review of related

literature showed that law enforcement information systems can be considered successful

at the technical and semantic level. This supports the argument that the problem with

most of the law enforcement information systems is related with the use of the system. In

other words, the collection of information is one thing and the use of that information is

another thing. Therefore, policymakers should ensure the acceptance of the systems by

the user community to get the most from the information technology projects. One way to

accomplish this acceptance is the provision for proper training of users on the

information technologies and related issues.

The final policy implication of this study is related with the use of information

technologies by police departments in their investigative activities. This study’s findings

 126

showed that there is not a consistent relationship with use of information technologies

and clearance rates. As a matter of fact, despite of the large investments on the

information technologies by police departments, use of information technologies may not

have a direct impact on clearance rates. The review of criminal investigation literature

showed that there are some other factors that may have an impact on the outcome of

criminal investigations. Although there are mixed evidence about the impact of most of

those factors, the preliminary investigation and the availability of information about the

suspect were consistently found as the most important determinant of criminal

investigations. Therefore, in order to be able to explain the fluctuations in the clearance

rates of police departments, preliminary investigation process should be examined in

more detail. An analysis of use of information technologies at preliminary investigation

stage may also provide better evidence about the impact of information technologies on

criminal investigations. Moreover, this kind of analysis can inform policy on how to use

information technologies more effectively in police departments.

 127

REFERENCES

Abt Associates. (2001). Police Department Information Systems Technology

Enhancement Project (ISTEP). Washington, D.C: Department of Justice, Office
of Community Oriented Policing Services.

Ackroyd, S., Harper, R., & Hughes, J. A. (1992). 6ew technology and practical police
work: The social context of technical innovation. Open Univ Pr.

Adderley, R. W., & Musgrove, P. (2001). Police crime recording and investigation
systems-a user's view. Policing: An International Journal of Police Strategies &
Management, 24(1), 100–114.

Alison, L., Snook, B., & Stein, K. L. (2001). Unobtrusive measurement: using police
information for forensic research. Qualitative Research, 1(2), 241.

Almutairi, H., & Subramanian, G. H. (2005). An empirical application of the DeLone and
McLean model in the Kuwaiti private sector. Journal of Computer Information
Systems, 45(3), 113–122.

ATF. (2009). Bureau of Alcohol, Tobacco, Firearms and Explosives - Fact Sheet. Bureau
of Alcohol, Tobacco, Firearms and Explosives. Retrieved from
http://www.nibin.gov/press/factsheets/0409-factsheet-nibin.pdf.

Ballantine, J. A., Galliers, R. D., & Straw, S. (1996). Information systems/technology

evaluation practices: Evidence from uk organizations. Journal of Information
Technology, 11(2), 129 - 141.

Barnett-Ryan, C. (2007). Introduction to the uniform crime reporting program. In J. P.

Lynch & L. Addington (Eds.), Understanding Crime Statistics: Revisiting the
Divergence of the 6CVS and UCR, Cambridge studies in criminology.
Cambridge: Cambridge University Press.

BJA. (2007). Regional Information Sharing Systems. RISS Overview. Retrieved April 13,
2009, from http://www.riss.net/overview.aspx.

Bloch, P. B., & Bell, J. (1976). Managing investigations: The Rochester system.
Washington: Police Foundation.

Blumstein, A. (1994). The task force on science and technology. In J. A. Conley (Ed.),
The 1967 President's crime commission report: Its impact 25 yearslater,
ACJS/Anderson monograph series. Highland Heights, KY: Academy of Criminal
Justice Sciences, Northern Kentucky University.

Boba, R. (2005). Crime Analysis and Crime Mapping. Thousand Oaks, Calif: Sage
Publications.

127
 128

Borg, M. J., & Parker, K. F. (2001). Mobilizing law in urban areas: The social structure
of homicide clearance rates. Law & Society Review, 35, 435.

Bozeman, B., & Bretschneider, S. (1986). Public management information systems:

Theory and prescription. Public Administration Review, 475-487.

Bradbury, S., & Feist, A. (2005). The use of forensic science in volume crime
investigations: a review of the research literature . Home Office Online Report ,
Great Britain Home Office.

Braga, A. A., & Pierce, G. (2004). Linking Crime Guns: The Impact of Ballistics
Imaging Technology on the Productivity of the Boston Police Department’s
Ballistics Unit. Journal of Forensic Sciences, 49(4).

Brandl, S. (1993). The impact of case characteristics on detectives' decision making*.

Justice Quarterly, 10(3), 395-415.

Brown, M. M. (1996). An empirical assessment of the hurdles to geographic information

system success in local government. State & Local Government Review, 193-204.

Brown, M. M. (2000). Criminal justice discovers information technology. In National

Institute of Justice (U.S.) (Ed.), Criminal justice 2000 - V.1 - The nature of crime:
Continuity and change. Washington, D.C.: U.S. Dept. of Justice, Office of Justice
Programs, National Institute of Justice.

Brown, T. M. L., McCabe, S. A., & Wellford, C. (2007). Global Positioning System
(GPS) Technology for Community Supervision: Lessons Learned. NOBLIS.

Brynjolfsson, E. (1993). The productivity paradox of information technology: Review

and assessment. Communications of the ACM, 36(12), 67-77.

Brynjolfsson, E., & Hitt, L. M. (1998). Beyond the productivity paradox.

Communications of the ACM, 41(8), 49-55.

Bureau of Justice Statistics. (2004). UCR and NIBRS participation. Bureau of Justice
Statistics UCR and 6IBRS Participation. Government, . Retrieved February 11,
2009, from http://www.ojp.usdoj.gov/bjs/nibrsstatus.htm.

Burrows, J. (1986). Investigating burglary: The measurement of police performance.

Home Office Research Study, Great Britain Home Office.

Burrows, J., & Tarling, R. (1982). Clearing up crime. Great Britain Home Office.

Burrows, J., & Tarling, R. (2004). Measuring the impact of forensic science in detecting
burglary and autocrime offences. Science & Justice, 44(4), 217-222.
 129

Burrows, J., Tarling, R., Mackie, A., Poole, H., & Hodgson, B. (2005). Forensic science
pathfinder project Evaluating increased forensic activity in two English police
forces. Home Office Online Report, Great Britain Home Office.

Bynum, T., Cordner, G., & Greene, J. R. (1982). Victim and offense characteristics:
Impact on police investigative decision-making. Criminology, 20(3-4), 301-318.

Cawley, D. F., Miron, H. J., Araujo, W. J., Wasserman, R., Mannello, T. A., & Huffman,
Y. (1977). Managing criminal investigations manual. US Dept. of Justice, Law
Enforcement Administration, Law Enforcement Assistance Administration,
National Institute of Law Enforcement and Criminal Justice: For sale by Supt. of
Docs., US Govt. Print. Off.

Chaiken, J. M., Greenwood, P. W., & Petersilia, J. (1976). The criminal investigation
process. RAND Corporation.

Chan, J. B. L. (2001). The technological game:: How information technology is

transforming police practice. Criminology and Criminal Justice, 1(2), 139.

Chen, H., Schroeder, J., Hauck, R. V., Ridgeway, L., Atabakhsh, H., Gupta, H., et al.
(2003). COPLINK Connect: Information and knowledge management for law
enforcement. Decision Support Systems, 34(3), 271-285.

Cheurprakobkit, S. (2000). Police-citizen contact and police performance Attitudinal

differences between Hispanics and non-Hispanics. Journal of Criminal Justice,
28(4), 325-336.

Chu, J. (2001). Law enforcement information technology. CRC Press.

Cohen, B., & Chaiken, J. M. (1987). Investigators who perform well. Issues and practices
in criminal justice. Washington, D.C: U.S. Dept. of Justice, National Institute of
Justice, Office of Communication and Research Utilization.

Cohen, J. (2003). Applied multiple regression/correlation analysis for the behavioral

sciences. Lawrence Erlbaum Associates.

Colton, K. W. (1972). Use of computers by police: Patterns of success and failure.

International City Management Association.

Colton, K. W. (1980). Police and computer technology: The case of the San Diego
computer-aided dispatch system. Public Productivity Review, A Symposium on
Computers and Productivity, 4(1), 21-42.

Colton, K. W., Brandeau, M. L., & Tien, J. M. (1983). A national assessment of police
command, control, and communications systems. US Dept. of Justice, National
Institute of Justice.
 130

Cordner, G. (1989). Police agency size and investigative effectiveness. Journal of

Criminal Justice, 17(1), 145-155.

Danziger, J. N., & Andersen, K. V. (2002). The impacts of information technology on

public administration: An analysis of empirical research from the “golden age” of
transformation. International Journal of Public Administration, 25(5), 591-627.

Danziger, J. N., & Kraemer, K. L. (1985). Computerized data-based systems and

productivity among professional workers: The case of detectives. Public
Administration Review, 196-209.

Davenport, T. H., & Prusak, L. (1997). Information ecology: Mastering the information
and knowledge environment. New York: Oxford University Press.

Davis, L. M., & Jackson, B. A. (2004). Acquiring, implementing and evaluating

information technology. In A. Pattavina (Ed.), Information technology and the
criminal justice system (1st ed.). Sage Publications, Inc.

Dean, G., & Gottschalk, P. (2007). Knowledge management in policing and law
enforcement: Foundations, structures, applications. Oxford: Oxford University
Press.

DeLone, W. H., & McLean, E. R. (1992). Information systems success: the quest for the
dependent variable. Information systems research, 3(1), 60-95.

DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of
information systems success: a ten-year update. Journal of Management
Information Systems, 19(4), 9-30.

DeLone, W. H., & McLean, E. R. (2004). Measuring e-commerce success: applying the
DeLone & McLean information systems success model. International Journal of
Electronic Commerce, 9(1), 31-47.

DeZee, M. (1995). Review of the book The 1967 President's Crime Commission Report:
Its impact 25 years later. Journal of Criminal Justice, 23(5), 489-492.

DuBois, J., Skogan, W. G., Hartnett, S. M., Bump, N., & Morris, D. (2007). CLEAR and
I-CLEAR: A Report on 6ew Information Technology in Chicago and Illinois.
Chicago, IL: Illinois Criminal Justice Information Authority.

Dunworth, T. (2000). Criminal justice and the IT revolution. In National Institute of

Justice (U.S.) (Ed.), Criminal justice 2000 - V.3 - Policies, processes, and
decisions of the criminal justice system. Washington, D.C.: U.S. Dept. of Justice,
Office of Justice Programs, National Institute of Justice.

Dunworth, T. (2004). Information technology and the criminal justice system: A

historical perspective. In A. Pattavina (Ed.), Information technology and the
criminal justice system (1st ed.). Sage Publications, Inc.
 131

Eastern Kentucky University. (2005). National Study of the Impact of Science and
Technology on the Process of Criminal Investigation in Law Enforcement
Agencies [Data File].

Eck, J. E. (1989). Rethinking detective management: Why investigative reforms are

seldom permanent or effective. In D. J. Kenney (Ed.), Police and policing:
Contemporary issues. New York: Praeger.

Eck, J. E. (2008). Foreword: Investigators, information, and interpretation: A summary of

criminal investigation research. In D. Rossmo (Ed.), Criminal investigative
failures. CRC.

Eck, J. E., & Maguire, E. R. (2005a). Have changes in policing reduced violent crime?
An assessment of the evidence. In A. Blumstein & J. Wallman (Eds.), The crime
drop in America. Cambridge University Press.

Eck, J. E., & Maguire, E. R. (2005b). Have changes in policing reduced violent crime?
An assessment of the evidence. In A. Blumstein & J. Wallman (Eds.), The crime
drop in America. Cambridge University Press.

Eck, J. E. (1983). Solving crimes: The investigation of burglary and robbery.

Washington, DC: Police Executive Research Forum.

Elliott, J. F. (1978). Crime control teams: An alternative to the conventional operational

procedure of investigating crimes. Journal of Criminal Justice, 6(1), 11-23.

Ericson, R. V. (1993). Making crime: A study of detective work. Toronto: University of

Toronto Press.

Faggiani, D., & Hirschel, D. (2004). The impact of information technology on crime
reporting: The NIBRS process. In A. Pattavina (Ed.), Information technology and
the criminal justice system (1st ed.). Sage Publications, Inc.

FBI. (2004a). Uniform crime reports. Uniform crime reports-Frequently asked questions.
Retrieved April 13, 2009, from http://www.fbi.gov/ucr/ucrquest.htm.

FBI. (2004b). Uniform Crime Reporting Handbook.

FBI. (2004c). A word about UCR data. Federal Bureau of Investigation - Uniform Crime
Reports. Retrieved April 13, 2009, from http://www.fbi.gov/ucr/word.htm.

FBI. (2005a). Crime in the United States. Crime in the United States, 2005. Retrieved
April 13, 2009, from http://www.fbi.gov/ucr/ucrquest.htm.

FBI. (2005b). Uniform Crime Reporting program data [United States]: Offenses known
and clearances by arrest, 2005 [Data FIle]. Inter-university Consortium for
Political and Social Research . Retrieved from http://www.icpsr.umich.edu/.
 132

FBI. (2006a). Crime in the United States. Crime in the United States, 2006. Retrieved
April 13, 2009, from http://www.fbi.gov/ucr/ucrquest.htm.

FBI. (2006b). Uniform Crime Reporting program data [United States]: Offenses known
and clearances by arrest, 2006 [Data FIle]. Inter-university Consortium for
Political and Social Research . Retrieved from http://www.icpsr.umich.edu/.

FBI. (2008a). National Crime Information Center (NCIC). 6ational Crime Information
Center (6CIC) - FBI Information Systems. Retrieved April 13, 2009, from
http://www.fas.org/irp/agency/doj/fbi/is/ncic.htm.

FBI. (2008b). Integrated Automated Fingerprint Identification System or IAFIS. FBI -

CJIS Division - Integrated Automated Fingerprint Identification System.
Retrieved April 13, 2009, from http://www.fbi.gov/ucr/faqs.htm.

FBI. (2009a). CODIS Combined DNA Index System. FBI - CODIS. Retrieved April 13,
2009, from http://www.fbi.gov/hq/lab/html/codisbrochure_text.htm.

FBI. (2009b). Violent Criminal Apprehension Program (ViCAP). Federal Bureau of

Investigations - Investigative Programs - Critical Incidence Response Group.
Retrieved April 13, 2009, from http://www.fbi.gov/hq/isd/cirg/ncavc.htm#vicap.

FBI. (2009c). National Incident-Based Reporting System. FBI 6ational Incident-Based

Reporting System - Frequently Asked Questions (FAQs). Retrieved April 13,
2009, from http://www.fbi.gov/ucr/faqs.htm.

Folk, J. (1971). Municipal detective systems-A quantitative approach. Technical Report,

Cambridge, MA: Massachusetts Institute of Technology.

Garson, G. D. (2006). Public information technology and e-governance: Managing the

virtual state (1st ed.). Jones & Bartlett Pub.

Gates, D., & Knowles, L. (1976). An evaluation of the RAND Corporation's analysis of
the criminal investigation process. Police Chief, 43(7), 20-24, 74-77.

Gelman, A., & Hill, J. (2006). Data Analysis Using Regression and
Multilevel/Hierarchical Models (1st ed.). Cambridge University Press.

Gilbert, J. N. (2004). Criminal investigation (6th ed., p. 624). Upper Saddle River, N.J:
Pearson, Prentice Hall.

Goldstein, H. (2005). Who Killed the Virtual Case File? IEEE Spectrum Online,
(September). Retrieved July 15, 2008, from
http://www.spectrum.ieee.org/sep05/1455.

Gottschalk, P. (2007). Knowledge management systems in law enforcement: technologies

and techniques. Hershey, PA: Idea Group Pub.
 133

Gove, W. R., Hughes, M., & Geerken, M. (1985). Are Uniform Crime Reports a Valid
Indicator of the Index Crimes-An Affirmative Answer with Minor Qualifications.
Criminology, 23, 451.

Greenberg, B., Elliott, C. V., Kraft, L. P., & Procter, H. S. (1977). Felony investigation
decision model: An analysis of investigative elements of information. Rockville,
MD: Stanford Research Institute, US Dept of Justice Law Enforcement
Assistance Administration.

Greenberg, B., & Lang, K. I. (1973). Enhancement of the investigative function Vol. 4,
burglary investigative checklist and handbook. Springfield, VA: National
Technical Information Service.

Greenberg, I., & Wasserman, R. (1979). Managing criminal investigations. Washington:

Dept. of Justice, Law Enforcement AssistanceAdministration, National Institute
of Law Enforcement and Criminal Justice, Office of Development, Testing and
Dissemination : for sale by the Supt. of Docs., U.S. Govt.Print. Off.

Greene, J. R. (2000). Community policing in America: Changing the nature, structure,

and function of the police. In National Institute of Justice (U.S.) (Ed.), Criminal
justice 2000 - V.3 - Policies, processes, and decisions of the criminal justice
system. Washington, D.C.: U.S. Dept. of Justice, Office of Justice Programs,
National Institute of Justice.

Greenwood, P. W. (1970). An analysis of the apprehension activities of the 6ew York city
police department. RAND.

Greenwood, P., Chaiken, J. M., & Petersilia, J. (1976). Response to - An evaluation of

the RAND Corporation's analysis of the criminal investigation process. Police
Chief, 43(12), 62-71.

Gunasekaran, A., Ngai, E., & McCaughey, R. (2008). Information technology and
systems justification. In Z. Irani & P. Love (Eds.), Evaluating information
systems: Public and private sector. Butterworth-Heinemann.

Harper, R. R. (1991). The computer game detectives, suspects, and technology. British
Journal of Criminology, 31(3), 292-307.

Harris, K., & Romesburg, W. (2002). Law Enforcement Tech Guide - How to plan,
purchase and manage technology (successfully!). Office of Community Oriented
Services. Retrieved July 15, 2009, from
http://www.cops.usdoj.gov/files/RIC/Publications/lawenforcementtechguide.pdf.

Hauk, R. V., & Chen, H. (1999). COPLINK: A case of intelligent analysis and
knowledge management (pp. 15-28). Association for Information Systems
Atlanta, GA, USA.
 134

Heintze, T., & Bretschneider, S. (2000). Information technology and restructuring in

public organizations: does adoption of information technology affect
organizational structures, communications, and decision making? Journal of
Public Administration Research and Theory, 10(4), 801-830.

Henman, P. (1996). Does computerization save government money? Information

Infrastructure & Policy, 5(4).

Hindelang, M. J. (1974). The uniform crime reports revisited. Journal of Criminal

Justice, 2(1), 1–17.

Hogan, M. D., & Radack, S. M. (1997). The quest for information technology standards
for the global information infrastructure. StandardView, 5(1), 30-35.

Holmes, M. D., Smith, B. W., Freng, A. B., & Munoz, E. A. (2008). Minority Threat,
Crime Control, and Police Resource Allocation in the Southwestern United States.
Crime & Delinquency, 54(1), 128.

Hoover, J. (2009). Obama's Budget Calls For Shifts In IT Spending. InformationWeek.

Retrieved from
http://www.informationweek.com/news/government/federal/showArticle.jhtml;jse
ssionid=THC3XYCZUQG24QSNDLOSKH0CJUNN2JVN?articleID=217400028
&pgno=1&queryText=&isPrev=.

Horvath, F., & Meesig, R. (1996). The Criminal Investigation Process And The Role Of
Forensic Evidence: A Review Of Empirical Findings. Journal of forensic
sciences, 41(6), 963-969.

Horvath, F., Meesig, R., & Lee, Y. (2001). A national survey of police policies and
practices regarding the criminal investigation process: Twenty-five years after
RA6D. U.S. Department of Justice.

Iivari, J. (2005). An empirical test of the DeLone-McLean model of information system

success. ACM SIGMIS Database, 36(2), 8-27.

Innes, M. (2003). Investigating murder: detective work and the police response to
criminal homicide. Clarendon studies in criminology. Oxford: Oxford University
Press.

Innes, M. (2007). Investigation order and major crime inquaries. In T. Newburn, T.

Williamson, & A. Wright (Eds.), Handbook of criminal investigation.
Cullompton, UK: Willan.

Institute of Policy Research. (2004). Policing Smarter Through IT: Lessons in Enterprise
Implementation. COPS.
 135

Ioimo, R. E., & Aronson, J. E. (2003). The benefits of police field mobile computing
realized by non-patrol sections of a police department. International Journal of
Police Science & Management, 5(3), 195-206.

Isaacs, H. (1967). A study of communications, crimes, and arrests in a metropolitan

police department. Appendix to the Task Force Report: Science and Technology.

Jiao, A. Y. (2007). Explaining homicide clearance: An analysis of Chicago homicide data

1965–1995. Criminal Justice Studies, 20(1), 3-14.

Jones, C., & Weatherburn, D. (2004). Evaluating police operations (1): A process and
outcome evaluation of operation vendas. New South Wales Bureau of Crime
Statistics and Research, Australia.

Keel, T. G., Jarvis, J. P., & Muirhead, Y. E. (2009). An Exploratory Analysis of Factors
Affecting Homicide Investigations: Examining the Dynamics of Murder
Clearance Rates. Homicide Studies, 13(1), 50.

Kennedy, W. (2009). The Impact of Police Agency Size on Crime Clearance Rates. UNC
Charlotte.

Kim, J., & Mueller, C. W. (1978). Introduction to Factor Analysis: What It Is and How
To Do It. Sage Publications, Inc.

Klinger, D. (1997). Negotiating order in patrol work: An ecological theory of police

response to deviance. Criminology, 35(2), 277-306.

Klug, D. J., Peterson, J. L., & Stoney, D. A. (1992). Automated Fingerprint Identification
Systems: Their Acquisition, Management, Performance, and Organizational
Impact: a Report to the 6ational Institute of Justice. National Institute of Justice.

Kraemer, K. L., Danziger, J. N., Dunkle, D. E., & King, J. L. (1993). The usefulness of
computer-based information to public managers. MIS Quarterly, 129-148.

Kuykendall, J. (1982). The criminal investigative process: Toward a conceptual

framework. Journal of Criminal Justice, 10(2), 131-145.

LeBeuf, M. E. (2001). Organized Crime and Cybercrime Criminal Investigations and

Intelligence on the Cutting Edge. Information Technology Series. Ottawa: Police
Sciences School Canadian Police College. Retrieved from
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.90.1936&rep=rep1&ty
pe=pdf.

Lee, G., & Perry, J. L. (2002). Are computers boosting productivity? A test of the
paradox in state governments. Journal of Public Administration Research and
Theory, 12(1), 77-102.
 136

Lehr, W., & Lichtenberg, F. R. (1998). Computer use and productivity growth in US
federal government agencies, 1987-92. Journal of Industrial Economics, 257-279.

Liska, A. E., & Chamlin, M. B. (1984). Social structure and crime control among
macrosocial units. American Journal of Sociology, 90(2), 383.

Losee, R. M. (1997). A discipline independent definition of information. Journal Of The

American Society For Information Science, 48(3), 254-269.

Lott, J. R., & Whitley, J. (2003). Measurement error in county-level UCR data. Journal
of Quantitative Criminology, 19(2), 185–198.

Luen, T. W., & Al-Hawamdeh, S. (2001). Knowledge management in the public sector:
principles and practices in police work. Journal of Information Science, 27(5),
311.

Maguire, M. (2003). Criminal investigation and crime control. In T. Newburn (Ed.),

Handbook of policing. Cullompton: Willan.

Maltz, M. (2007). Missing UCR data and divergence of NCVS and UCR trends. In J. P.
Lynch & L. Addington (Eds.), Understanding Crime Statistics: Revisiting the
Divergence of the 6CVS and UCR, Cambridge studies in criminology.
Cambridge: Cambridge University Press.

Manning, P. K. (1992). Information technologies and the police. Crime and Justice: A
Review of Research, 15, 349.

Manning, P. K. (1996). Information technology in the police context: The. Information

Systems Research, 7(1), 52–62.

Manning, P. K. (2001). Technology's ways: Information technology, crime analysis and

the rationalizing of policing. Criminology and Criminal Justice, 1(1), 83.

Manning, P. K. (2004). Environment, technology and organizational change: Notes from

the police world. In A. Pattavina (Ed.), Information technology and the criminal
justice system (1st ed.). Sage Publications, Inc.

Manning, P. (2008). The technology of policing: crime mapping, information technology,

and the rationality of crime cControl (illustrated edition.). NYU Press.

Manning, P. K. (2003). Policing contingencies. Chicago: University of Chicago Press.

Marche, G. E. (1994). The production of homicide solutions: An empirical analysis.

American Journal of Economics and Sociology, 53(4), 385-401.

Marshall, K. (1998). Mobile Data Access Increases Officer Efficiency. FBI Law
Enforcement Bulletin, 6.
 137

Martin Jr, R. A., & Legault, R. L. (2005). Systematic Measurement Error with State-
Level Crime Data: Evidence from the" More Guns, Less Crime" Debate. Journal
of Research in Crime and Delinquency, 42(2), 187.

Mayer, A. (2009). Geospatial Technology Helps East Orange Crack Down on Crime.
Geography & Public Safety, 1(4), 3.

McDevitt, D. S. (2005). Managing the investigative unit. Springfield, Ill: C.C. Thomas.

McEwen, T. (2002). Computer Aided Dispatch in Support of Community Policing.

National Institute of Justice .

McGill, T., Hobbs, V., & Klobas, J. (2003). User developed applications and information
systems success: A test of DeLone and McLean's model. Information Resources
Management Journal, 16(1), 24-45.

McRae, J. J., & McDavid, J. (1988). Computer-based Technology in Police Work: A

Benefit-Cost Analysis of a Mobile Digital Communications System. Journal of
Criminal Justice, 16(1), 14.

Meehan, A. J. (1998). The impact of mobile data terminal (MDT) information technology
on communication and recordkeeping in patrol work. Qualitative Sociology,
21(3), 225-254.

Morckel, K. L. (2002). GPS Vehicle Tracking Improves Auto Theft Enforcement. Police
Chief, 69(12), 39–43.

Morris, P., & Heal, K. (1981). Crime control and the police: A review of research. Home
Office Research Study, Her Majesty’s Stationery Office.

National Commission on Terrorist Attacks. (2004). The 9/11 commission report: Final
report of the national commission on terrorist attacks upon the United States. W.
W. Norton & Company.

Nogala, D. (1995). The future role of technology in policing. In J. P. Brodeur (Ed.),

Comparisons in policing: An international perspective. Aldershot, Hants,
England: Avebury.

Nolan, J. J. (2004). Establishing the statistical relationship between population size and
UCR crime rate: Its impact and implications. Journal of Criminal Justice, 32(6),
547–555.

Northrop, A., Kraemer, K. L., & King, J. L. (1995). Police use of computers. Journal of
Criminal Justice, 23(3), 259-275.

Nunn, S. (1994). How capital technologies affect municipal service outcomes: The case
of police mobile digital terminals and stolen vehicle recoveries, (13), 539- 559.
 138

Nunn, S. (2001). Police information technology: Assessing the effects of computerization

on urban police functions. Public Administration Review, 61(2), 221-234.

Nunn, S., & Quinet, K. (2002). Evaluating the effects of information technology on
problem-oriented policing: If it doesn't fit, must we quit? Evaluation Review,
26(1), 81.

O’Brien, R. M. (2003). UCR violent crime rates, 1958–2000: recorded and offender-
generated trends. Social Science Research, 32(3), 499–518.

O'Brien, J. A. (2002). Introduction to information systems: Essentials for the e-business

enterprise. McGraw-Hill.

O'Connell, P. E. (2002). An intellectual history of the Compstat model of police

management. Ph.D. Dissertation, City University of New York.

O'Shea, T., & Nicholls, K. (2003). Crime analysis In America: Findings and
recommendations. Office of Community Oriented Policing Services: Washington,
DC.

O'Shea, T. (1998). Analyzing police department data: How and how well police officers
and police departments manage the data they collect. In L. Moriarty & D. Carter
(Eds.), Criminal Justice Technology in the 21st Century. Springfield, Ill: Charles
C. Thomas.

Osterburg, J. W., & Ward, R. H. (2007). Criminal investigation: A method for

reconstructing the past (5th ed.). Newark, NJ: LexisNexis.

Ostrom, E., Parks, R. B., & Whitaker, G. P. (1978). Police agency size: Some evidence
on Its effects. Police Studies, 1(1), 34-46.

Paré, P. P., Felson, R. B., & Ouimet, M. (2007). Community variation in crime clearance:
A multilevel analysis with comments on assessing police performance. Journal of
Quantitative Criminology, 23(3), 243-258.

Paulsen, D. J. (2004). To map or not to map: Assessing the impact of crime maps on
police officer perceptions of crime. International Journal of Police Science &
Management, 6(4), 234-246.

Peak, K. J., Barthe, E. P., & Garcia, A. (2008). Campus Policing in America: A Twenty-
Year Perspective. Police Quarterly, 11(2), 239.

Peled, A. (2001). Do computers cut red tape? The American Review of Public
Administration, 31(4), 414.

Peterson, J. L., Mihajlovic, S., & Gilliland, M. (1984). Forensic evidence and the police:
The effects of scientific evidence on criminal investigations. US Dept. of Justice,
National Institute of Justice.
 139

Porter, M. E. (1998). Competitive strategy: techniques for analyzing industries and

competitors: with a new introduction. Free Press.

President's Commission on Law Enforcement and Administration of Justice. (1967). The

challenge of crime in a free society; a report. Washington: U.S. Govt. Print. Off.

Puckett, J. L., & Lundman, R. J. (2003). Factors affecting homicide clearances:

Multivariate analysis of a more complete conceptual framework. Journal of
Research in Crime and Delinquency, 40(2), 171.

Purpura, P. P. (1996). Criminal justice. Elsevier.

R Development Core Team. (2009). R: A Language and Environment for Statistical

Computing. R Foundation for Statistical Computing. Retrieved from
http://www.R-project.org.

Reaves, B. (2008). Campus Law Enforcement, 2004–2005. Washington, DC: US

Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

Reiss, A. J. (1971). The police and the public. New Haven: Yale University Press.

Rocheleau, B. A. (2005). Public management information systems (illustrated edition.).

Idea Group Publishing.

Rocheleau, B., & Wu, L. (2002). Public versus private information systems: Do they
differ in important ways? A review and empirical test. The American Review of
Public Administration, 32(4), 379.

Roldan, J., & Millian, A. (2000). Analysis of the information system success dimensions
interdependence: An adaptation of the DeLone and McLean's model in the
Spanish EIS field. Presented at the BITWorld 2000.

Rosenfeld, R., Fornango, R., & Baumer, E. (2005). Did ceasefire, compstat, and exile
reduce homicide. Criminology & Public Policy, 4(3), 419-449.

Sahai, H., & Ojeda, M. M. (2005). Analysis of Variance for Random Models:
Unbalanced data. Birkhäuser.

Sanders, W. B. (1977). Detective work: A study of criminal investigations. New York:

Free Press.

Schutt, R. (2006). Investigating the Social World - The Process and Practice of Research
- 5th (Fifth) Edition.

SEARCH. (1997). Implementing the 6ational Incident-Based Reporting System: A

Project Status Report. Washington, DC: Bureau of Justice Statistics. Retrieved
from http://www.ojp.usdoj.gov/bjs/pub/pdf/inibrs.pdf.
 140

Seaskate. (1998). The evolution and development of police technology: A technical

report. Seaskate, Inc.

Seddon, P., & Kiew, M. (1996). A partial test and development of DeLone and McLean
model of IS success. Australian Journal of Information Systems, 4(1), 90-109.

Shane, J. (2004). Compstat implementation. FBI Law Enforcement Bulletin , 73(6).

Silverman, E. B. (1999). 6YPD battles crime: Innovative strategies in policing. Boston:

Northeastern University Press.

Silverman, E. B. (2006). Compstat's innovation. In D. Weisburd & A. A. Braga (Eds.),

Police innovation: Contrasting perspectives. Cambridge, UK: Cambridge
University Press.

Skogan, W. G., & Antunes, G. E. (1979). Information, apprehension, and deterrence:

Exploring the limits of police productivity. Journal of Criminal Justice, 7(21),
217-241.

Sparrow, M. K. (1993). Information systems and the development of policing. 6ational

Institute of Justice, Perspectives on Policing, (16).

SPSS Inc. (2008). SPSS for Windows. Chicago, IL.

Steinberg, S. L. (2005). GIS. SAGE.

Stevens, J. W. (1995). Computer technology. In W. Bailey (Ed.), Garland reference

library of the humanities (2nd ed.). New York: Garland Pub.

Steverson, L. A. (2007). Policing in America. ABC-CLIO.

Swanson, C. (1978). The influence of organization and environment on arrest policies in

major US cities. Policy Studies Journal, 7, 390-398.

Tien, J. M., Cahn, M. F., Neray, R. C., Einstein, D. M., & Pei, K. (2005). Measuring the
Performance of Criminal History Records Systems: The Records Quality Index.
Structured Decisions Corporations.

Tong, D. S., & Bowling, B. (2006). Art, Craft and Science of Detective Work. The Police
Journal, 79(4), 323–330.

Traub, R. E. (1994). Reliability for the Social Sciences: Theory and Applications
(illustrated edition.). Sage Publications, Inc.

Turk, Z. (2000). What is construction information technology. Presented at the

Informacni technologie ve stavebnictvi 2000, Prague.

U.S. Census Bureau. (2000). American fact finder. Retrieved March 8, 2009, from
http://factfinder.census.gov/.
 141

Vito, G. F., Walsh, W. F., & Kunselman, J. (2005). Compstat: the manager’s perspective.
International Journal of Police Science & Management, 7(3).

Waegel, W. B. (1982). Patterns of police investigation of urban crimes. Jounal of Police

Science and Administration, 10(4), 452-465.

Waegel, W. (1981). Case routinization in investigative police work. Social Problems,

28(3), 263-275.

Walsh, W. F., & Vito, G. F. (2004). The meaning of Compstat: Analysis and response.
Journal of Contemporary Criminal Justice, 20(1), 51.

Weisburd, D., Mastrofski, S. D., McNally, A. M., & Greenspan, R. (2002). Reforming to
preserve: Compstat and strategic problem solving in American policing.
Criminology & Public Policy, 2, 421.

Weisburd, D., Mastrofski, S. D., Willis, J. J., & Greenspan, R. (2006). Changing
everything so that everything can remain the same: Compstat and American
policing. In D. Weisburd & A. A. Braga (Eds.), Police innovation: Contrasting
perspectives. Cambridge, UK: Cambridge University Press.

Weller, S., Martin, J., Price, D., & Wagenknecht-Ivey, B. (2001). Criminal justice system
project final report of evaluation findings: Critical components for successful
criminal justice system planning. Washington, D.C.: National Institute of Justice.

Wellford, C., & Cronin, J. (1999). Analysis of variables affecting the clearance of
Homicides: a multisite study. Washington, DC: Justice Research and Statistics
Association.

Wiig, K. M. (1993). Knowledge management foundations: Thinking about thinking: How

people and organizations create, represent, and use knowledge. Arlington, Tex:
Schema Press.

Willman, M. T., & Snortum, J. R. (1984). Detective work: The criminal investigation
process in a medium-size police department. Criminal Justice Review, 9(1), 33-
39.

Willmer, M. A. P. (1970). Crime and information theory. Edinburgh: University Press.

WordNet. (n.d.). information Technology. Word6et. Online Dictionary, . Retrieved July

25, 2008, from
http://wordnet.princeton.edu/perl/webwn?s=information+technology&sub=Search
+WordNet&o2=&o0=1&o7=&o5=&o1=1&o6=&o4=&o3=&h=.

Zaid, M. N. (1967). Urban differentiation, characteristics of boards of directors, and

organizational effectiveness. American Journal of Sociology, 261–272.
 142

Zavala, A., & Mullen, T. H. (1970). Use Of computer-based modus operandi data
systems. US Dept of Justice Law Enforcement Assistance Admin.
 143

APPENDIX A: CRIMINAL INVESTIGATION SURVEY

General

1. What is your title/rank? _______________________________________________

1a. How long have/did you serve as an investigator? (If not applicable, please
write 0.) _______________________________________________________

2. Name of Department: __________________________________________________

3. Population served: _____________________________________

4. Area of agency’s jurisdiction (square miles): _______________________________

5. Total number of full-time sworn officers: ___________________________________

6. Total number of personnel in the department assigned primarily to investigative duties

Sworn: _________________
Civilian: _________________

7. Does your department have a special title for officers assigned to investigative duties,
whether or not they have a special official rank?
0 = No 1 = Yes

7a. If yes, what is this title? ______________________________________________

8. How many officers in the department have this title? _________________________

Investigative Training

9. How many hours of formal investigative training are provided to recruits in the basic
academy? _____________ Hours

10. How many hours of additional formal investigative training are provided to newly
appointed investigators? _____________ Hours

11. If routine refresher training is provided to investigators, please specify how often.
__________________ If not, go to Question 12.

11a. Number of hours: __________________

11b. What type(s) of routine refresher training is provided? (Circle all that apply)

143
 144

a. Crime type training (homicide, crimes against property, drugs, etc.)

b. Investigative techniques (interviews/interrogations, crime scene
management, etc.)
c. Legal issues (arrest, search, court testimony, etc.)
d. Management/administration (report writing, case management, data
systems, etc.)
e. Other – Specify: __________________________________________

11c. What portion of the required training is documented for liability purposes?
0 = None
1 = Some
2 = Most
3 = All

12. For classroom instruction on investigations provided for investigators and/or

uniformed officers, who does the training? (Circle all that apply)
a. Educational institutions
b. Federal agencies
c. In-house personnel
d. Other local agencies
e. Private organizations
f. State agencies
g. Other – Specify: __________________

13. To what degree has each of the factors listed below been a problem regarding
training of investigators?

Factor None Slight Moderate Large

a. Excessive length of training

b. Ineffectiveness of training

c. Lack of funding

d. Lack of management support

e. Lack of quality of training

f. Low individual motivation

g. Manpower shortage

h. Availability of desired training

i. Technology for training

 145

j. Other – Specify:
_________________

Investigative Roles

14. For each of the items listed below, place an X in the column that most closely
describes the frequency for which investigators in your agency participate when
investigating serious crimes. If your agency does not have an investigator(s), please
skip to question 17.

a. Tasks Never Sometimes Usually Always

(1) Conduct undercover investigations

(2) Perform community problem solving

(3) Process crime scenes for physical

evidence

(4) Prioritize cases based on local area

problems

(5) Self-assign cases based on local

problems

(6) Work in pairs

b. Work with Uniformed Officers

(1) In teams

(2) On decoy units, stakeouts, etc.

(3) To analyze crime patterns

c. Community-related Activities

(1) Provide crime information to the

public

(2) Receive at least 8 hours of

community policing training

(3) Regularly participate in community

 146

meetings

(4) Use citizen volunteers to assist in

investigations

(5) Work in teams with citizen groups

(6) Work with citizens on community

outreach

15. Listed below are a number of criteria and processes that can be used to select
investigators. For each one, please indicate the frequency it is used in your agency.

a. Criteria: Never Sometimes Usually Always

(1) Arrest productivity

(2) Education requirements

specifically for investigators

(3) Investigation skills

(4) Minimum number of years of

experience

(5) Personnel records

(commendations, complaints,
etc)

(6) Promotion to a certain grade

level

(7) Supervisor/staff ratings or

evaluations

(8) Competitive examination

(9) Other – Specify:______________

Never Sometimes Usually Always

b. Processes:

(1) Civil service exam

 147

(2) Oral board interview

(3) Peer evaluation

(4) Personal interview

(5) Written tests

(6) Verbal tests

(7) Other –
Specify:_______________

16. Is a probationary period required for newly selected investigators? If no, go to

Question 17.
0 = No 1 = Yes

16a. If yes, number of weeks of probation? _________weeks

Case Assignments

17. Many agencies “screen out” burglaries with low solvability. Approximately what
percentage of your agency’s burglary cases are screened out, i.e., not given any
investigation following completion of the initial crime report? _________________

18. How does your agency "screen out" or "skim" cases? (Circle all that apply)
a. Seriousness of offense
b. Presence of physical or forensic evidence
c. Witness testimony
d. Victim characteristics
e. Offender history
f. Formal solvability factors
g. Supervisor judgment
h. Investigator judgment
i. By type of crime (i.e. Cyber crime) – Specify: ___________________________
j. We don’t screen out cases
k. Others – Specify:

19. Many agencies use alternative methods of taking crime reports, such as telephone
reporting. Does your agency utilize any alternative methods for Part 1 crimes? If no,
go to Question 20.
0 = No 1 = Yes
 148

19a. If yes, please specify the alternative method(s) and for which crimes methods
are used.
_____________________________________________________________
_____________________________________________________________
_____________________________________________________________

20. Once a decision is made to investigate a case, how is it assigned to an investigator?

If your agency does not have an investigator(s), please skip to question 21.

Never Sometimes Usually Always

a. By rotation

b. By the size of investigator’s

caseload

c. By the experience of the

investigator

d. By the personal characteristics of

the investigator

e. By the specialty of the

investigator

f. Other – Specify: _______________

21. Which of the following investigative functions do uniformed officers perform in your
agency?

Never Sometimes Usually Always

a. Canvass areas for witnesses

b. Collect physical evidence from crime

scene

c. Collect physical evidence from

suspect

d. Conduct drug field tests

e. Conduct records checks

f. Conduct surveillance
 149

g. Conduct undercover activities

h. Coordinate investigations with

prosecutors

i. Interrogate suspects

j. Interview suspects

k. Interview victims

l. Interview witnesses

m. Secure crime scene

n. Submit evidence for forensic analysis

o. Testify in court

Job Enhancement

22. Within the past five years, has your agency attempted to enhance the role of
uniformed officers in investigating crimes? If no, go to Question 23.
0 = No 1 = Yes

22a. If yes, in what ways? (Circle all that apply)

a. Investigators can refer cases back to officers for follow-up
investigation
b. Officers conduct complete follow-up investigation as part of a team
c. Officers conduct more investigation at scene prior to handing case to
investigator
d. Officers are temporarily assigned to an investigation unit as part of
career development
e. Other – Specify: ___________________________________________

22b. Why did your agency try to enhance the uniformed officer’s role in
investigating crime? (Circle all that apply)
a. To assist in evaluating the work performance of uniformed officers
b. To clear more crimes
c. To free investigators for major crime investigation
d. To improve the morale of uniformed officers
e. To improve the quality of reports passed to investigators
f. To improve the relationship between uniformed officers and
investigators
g. To improve uniformed officer awareness of the investigative process
 150

h. To meet budgetary constraints

i. To shorten case closure time
j. Other – Specify: ___________________________________________

Investigative Functions and Effectiveness

23. Listed below are a number of different goals that may be associated with the criminal
investigation function. For each goal, indicate how important your agency considers
it to be with regard to criminal investigations.

Crime-related Goals Not Slightly Moderately Largely

Important
Important Important Important

(1) Clear cases

(2) Collect intelligence about

other crimes

(3) Convict suspects

(4) Investigate all serious

crimes

(5) Prevent crime

(6) Prosecute suspects

(7) Protect victim and

witnesses

(8) Reduce crime

(9) Solve problems

(10) Other – Specify:

_____________________
__

Other Goals

(1) Citizen
satisfaction/support

Not Slightly Moderately Largely

 151

Important Important Important Important

(2) Keep the community

informed

(3) Plan/implement crime

prevention strategies

(4) Protect the public

(5) Provide support/feedback

to victims

(6) Recover/return property

(7) Insure justice in the

community

(8) Other – Specify:

____________________

Evidence and Crime Labs

24. Does your department use evidence technicians who respond to the crime scene? If
no, go to Question 26.
0 = No 1 = Yes

24a. If yes, are they sworn personnel?

0 = No 1 = Yes

25. Are people who are designated as evidence technicians in your agency required to
have any specialized experience or training? If no, go to Question 26.
0 = No 1 = Yes

25a. If yes, what type(s)? (Circle all that apply)

a. A college degree
b. Investigative experience
c. Some college education
d. Specialized in-house training
e. Specialized training outside of your agency
f. Sworn officer experience
g. Other-Specify: ____________________________________________
 152

26. Please estimate how frequently the following physical evidence checks (i.e.
systematic efforts to determine if such evidence is present) are made at the crime
scene.

Enter 0 = Never, 1 = Rarely, 2 = Sometimes, 3 = Usually, 4 = Always

Physical Evidence Check

Crime Type Finger Tool Chemical Shoe Print or Digital Other:

Prints Marks Analysis Tire Casing Evidence ______

Homicide

Residential
Burglaries

Commercial
Burglaries

Robberies

27. In your experience, what proportion/percentage of the following types of solved

cases are solved today based primarily on fingerprint evidence? (i.e., 50% would
indicate that half of the cases that are solved are solved primarily due to fingerprint
evidence).
a. Homicide _________________
b. Rape _____________________
c. Burglary __________________
d. Auto theft _________________
e. Robbery __________________

28. When your investigators make use of routine crime laboratory services, what type of
laboratory is most often used? (Circle one)
a. Your agency’s own crime laboratory
b. A crime laboratory that is part of another local/county police agency
c. A crime laboratory that is part of another state/federal police agency
d. A state laboratory that is not part of a police organization (e.g., public health)
e. A laboratory that is privately owned

29. How would personnel in your agency describe their access to routine crime
laboratory services?
a. Readily available in all cases
b. Readily available but only in serious cases
c. Available but difficult to get timely access
d. Access is limited, hindering some investigations
 153

30. When personnel make use of routine crime laboratory services, how would they
describe the average turn-around time for analysis other than for drug/alcohol cases?
1 = Timely
2 = Somewhat slow
3 = Very slow
4 = Completely inadequate

31. Does your agency own a mobile crime scene/laboratory vehicle? If no, go to
Question 32.
0 = No 1 = Yes

31a. If yes, please describe the vehicle.

___________________________________________________________________
___________________________________________________________________
___________________________________________________________________

32. What is the approximate number of cases that your agency has cleared as a result of
DNA analysis? _____________

32a. How many cases in the past year would not have been cleared otherwise?
____________

33. In your experience, what proportion/percentage of the following types of solved

cases are solved today based primarily on DNA evidence?
a. Homicide ______________
b. Rape _______________
c. Burglary _______________
d. Robbery _______________
e. Auto theft ______________

34. When your agency uses the Automated Fingerprint Identification System (AFIS),
who provides the service? (Circle all that apply)
a. Your agency’s own AFIS
b. A neighboring agency’s AFIS
c. A state administered AFIS
d. A federally administered AFIS
e. Other – Specify: _________________________________________________

Technology

35. Does your agency have electronic crime investigation capability, i.e. the ability to
analyze digital information located on a variety of storage mediums? If no, go to
Question 36.
0 = No 1 = Yes
 154

35a. If yes, how many investigators are assigned to investigate electronic crimes?
Full-time: _________________
Part-time: _________________

36. Does your department use any of the following technologies? (Circle all that apply)
a. Cellular or mobile phones
b. Global Positioning System (GPS)
c. Personal Digital Assistant (PDA) or handheld computer
d. Blackberry
e. Mobile Data Computer/Mobile Data Terminal (MDC/MDT)
f. Intranet
g. Web pages or web-based
h. Computer Based Training (CBT)
i. Laptop computer

37. Does your department use any of the following services for investigative purposes?
(Circle all that apply)
a. National Integrated Ballistics Information Network (NIBIN)
b. Integrated Ballistics Identification System (IBIS)
c. Violent Criminal Apprehension Program (ViCap)
d. Regional Information Sharing Systems (RISS)
e. National Crime Information Center (NCIC)

Agency Policies

38. In some jurisdictions recording of police-witness and/or victim interviewing is legally

required. Is this true in your agency’s jurisdiction? If no, go to Question 39.
0 = No 1 = Yes

38a. If yes, how are you required to record interrogations?

a. Only written recording (by stenographer, court reporter) is required
b. Only audio is required
c. Both audio and visual recording is required

38b. Have you had cases that were denied prosecution or which did not go to trial
because the required interrogation recording was not available?
0 = No 1 = Yes

39. Even if not legally required, are interrogations of suspects routinely recorded by
audio or audio/visual means?
0 = No 1 = Yes
 155

Agency Files and Records

40. Do personnel conducting investigations complete any kind of formal activity log to
account for how their time is spent?
0 = No 1 = Yes

41. Are the types of records listed below available to personnel conducting investigation
in your agency in manual or computer form?

Records Not Available Available Available

Available Manually on Both
Computer Manually and
on Computer

a. Crime reports

b. Arrest reports

c. Case disposition

d. Prosecution disposition

e. Court dispositions

f. Summary crime statistics

42. Please identify the availability of the following files for criminal investigations
purposes.

Files Not Available Available Available

Available Manually on Both
Computer Manually and
on Computer

a. Fingerprints

b. Known offender

c. M.O. file

d. Mug shot

e. Organized crime
intelligence
 156

f. Narcotics intelligence

g. Sex offender

h. Stolen property

i. Stolen vehicles

j. Other – Specify:
____________________

43. Within the next year, does your agency plan to upgrade or enhance any of the
following? (Circle all that apply)
a. Computers in police vehicles
b. Crime analysis capabilities
c. Crime report forms
d. Records Management Systems
e. Case disposition files
f. Investigative support files
g. Personal communication devices
h. Dispatch communication systems
i. Other – Specify: _________________________________________________

Agency "eeds

44. What is the extent of your agency’s need for additional funding in the areas listed
below in order to improve investigative effectiveness?

None Slight Moderate Large

a. Equipment (e.g., vehicles,

surveillance)

b. Evidence collection issues

c. Evidence processing (e.g., crime labs,

DNA analysis)

d. Funding for informants

e. Investigative operations (e.g., task

forces, stings)

f. Personnel
 157

g. Communication technology

h. Information technology

i. Investigative technology

j. Training

k. Crime mapping

l. Other – Specify: _____________

Other

45. Please respond to each of the following statements as they apply to criminal
investigation in your agency today.

Strongly Agree Don’t Disagree Strongly N/A

Agree Know Disagree

a. A lot of fingerprint evidence at

crime scenes is not collected.

b. A lot of latent fingerprints

collected at crime scenes are
never processed through AFIS.

c. Our department could solve a

lot more cases with fingerprint
evidence if we had more crime
scene technicians.

d. Our department could solve a

lot more cases with better
crime scene technology.

e. Our department could solve a

lot more cases with better
access to AFIS.

f. A lot of DNA evidence at crime

scenes is not collected.

g. A lot of DNA evidence collected

at crime scenes is never
 158

processed for analysis.

h. Our department could solve a

lot more cases with DNA
evidence if we had more crime
scene technicians.

i. Our department could solve a

lot more cases with better
access to the crime lab.

j. Our department could solve a

lot more cases if there were
more convicted offenders in the
DNA database.

k. We are currently maximizing

the potential of fingerprints for
solving crimes.

l. We are currently maximizing the

potential for solving crimes
using DNA.

46. If additional research on the criminal investigation process were carried out, what
priority would you give each of the following areas?

Research None Low Moderate High

a. Case screening

b. Clearance rates

c. Crime intelligence

d. Crime mapping

e. Information systems

f. Decentralization/centralization of
investigators

g. Generalization/specialization of
investigator roles
 159

h. Integration of community policing

and investigations

i. Interagency cooperation

j. Investigator relationship with

communities

k. Investigative role of patrol officers

l. Investigator selection

m. Investigator training

n. Management of continuing
investigations

o. Performance evaluation of
investigators

p. Police/prosecutor relations

q. Prosecution and conviction rates

r. Role of technology in criminal

investigations

s. Technological improvements in
investigations management

t. Technological improvements in
investigative techniques

u. Other – Specify:
______________________________
 APPENDIX B: CORRELATION COEFFICIENTS OF USE OF IT ITEMS

Table 1
Correlation Coefficients
Cellular / PDA / Web Pages
Mobile Handheld MDC / / Web- Laptop
Phones GPS Computer Blackberry MDT Intranet based CBT Computer NIBIN
Cellular / Mobile Phones 1 .193** .112 .072 .118* .231** .197** .128* .139* .073
**
GPS .193** 1 .353** .315 .253** .270** .269** .228** .384** .253**
PDA / Handheld Computer .112 .353** 1 .429** .311** .210** .357** .325** .366** .268**
Blackberry .072 .315** .429** 1 .336** .271** .293** .286** .266** .224**
MDC / MDT .118* .253** .311** .336** 1 .355** .330** .229** .451** .347**
**
Intranet .231** .270** .210** .271 .355** 1 .370** .289** .308** .247**
Web Pages / Web-based .197** .269** .357** .293** .330** .370** 1 .342** .391** .209**
**
CBT .128* .228** .325** .286 .229** .289** .342** 1 .245** .214**
Laptop Computer .139* .384** .366** .266** .451** .308** .391** .245** 1 .242**
**
NIBIN .073 .253** .268** .224 .347** .247** .209** .214** .242** 1
IBIS .064 .214** .266** .208** .300** .253** .267** .138* .219** .566**
**
ViCap .176** .353** .425** .310 .305** .236** .352** .252** .374** .402**
RISS .045 .175** .288** .173** .147* .159** .177** .213** .195** .207**
NCIC .339** .148* .130* .087 .161** .191** .160** .120* .196** .168**
Crime Reports -.005 .037 .087 .047 .169** .058 .071 .083 -.019 .067
Arrest Reports .004 .054 .030 -.059 .133* .012 .048 .060 -.055 .081
Case Disposition -.029 .138* .065 .079 .181** .143* .076 .064 .034 .065
Prosecution Disposition .059 .164** .117 .016 .148* -.006 .041 -.003 .010 .129*
Court Dispositions .019 .122* .163** .042 .202** .085 .057 .026 .030 .099
Summary Crime Statistics .009 .241** .160** .179** .237** .128* .098 .081 .060 .235**
160
Table 1 (Continued)
Summary
Crime Arrest Case Prosecution Court Crime
IBIS ViCap RISS NCIC Reports Reports Disposition Disposition Dispositions Statistics
Cellular / Mobile Phones .064 .176** .045 .339** -.005 .004 -.029 .059 .019 .009
GPS .214** .353** .175** .148* .037 .054 .138* .164** .122* .241**
PDA / Handheld Computer .266** .425** .288** .130* .087 .030 .065 .117 .163** .160**
Blackberry .208** .310** .173** .087 .047 -.059 .079 .016 .042 .179**
MDC / MDT .300** .305** .147* .161** .169** .133* .181** .148* .202** .237**
Intranet .253** .236** .159** .191** .058 .012 .143* -.006 .085 .128*
Web Pages / Web-based .267** .352** .177** .160** .071 .048 .076 .041 .057 .098
CBT .138* .252** .213** .120* .083 .060 .064 -.003 .026 .081
Laptop Computer .219** .374** .195** .196** -.019 -.055 .034 .010 .030 .060
NIBIN .566** .402** .207** .168** .067 .081 .065 .129* .099 .235**
IBIS 1 .405** .219** .086 .094 .089 .042 .146* .190** .215**
ViCap .405** 1 .298** .200** .209** .147* .123* .119 .150* .234**
RISS .219** .298** 1 .149* .050 .098 .109 .026 .091 .129*
NCIC .086 .200** .149* 1 .039 .111 .055 .086 .117 .036
Crime Reports .094 .209** .050 .039 1 .748** .347** .169** .146* .332**
Arrest Reports .089 .147* .098 .111 .748** 1 .390** .270** .269** .285**
Case Disposition .042 .123* .109 .055 .347** .390** 1 .542** .472** .278**
Prosecution Disposition .146* .119 .026 .086 .169** .270** .542** 1 .713** .338**
Court Dispositions .190** .150* .091 .117 .146* .269** .472** .713** 1 .251**
Summary Crime Statistics .215** .234** .129* .036 .332** .285** .278** .338** .251** 1
161
Table 1 (Continued)
Known M.O. Mug Organized Crime Narcotics Sex Stolen Stolen
Fingerprints Offender File Shot Intelligence Intelligence Offender Property Vehicles
Cellular / Mobile Phones -.004 .028 -.017 .109 .083 .054 .006 .010 .034
GPS .207** .116 .252** .189** .336** .243** .189** .113 .141*
PDA / Handheld Computer .268** .129* .217** .215** .266** .302** .195** .155* .149*
** ** *
Blackberry .244** .037 .124* .204** .195 .219 .123 .085 .094
** ** ** **
MDC / MDT .273** .219** .223** .433** .161 .187 .250 .247 .306**
*
Intranet .219** .088 .100 .226** .150 .062 .100 -.026 .008
** * ** **
Web Pages / Web-based .203** .185** .200** .216** .239 .143 .249 .192 .160**
** ** *
CBT .206** .144* .105 .189** .243 .205 .125 .078 .111
** ** ** *
Laptop Computer .208** .103 .194** .242** .187 .229 .162 .149 .051
NIBIN .214** .106 .154* .200** .110 .169** .098 .004 .123*
IBIS .229** .128* .204** .176** .141* .157* .107 .013 .079
** ** **
ViCap .329** .252** .282** .222** .294 .292 .186 .113 .181**
RISS .054 .106 .205** .115 .229** .240** .139* .066 .101
NCIC .074 -.037 .024 .080 .079 .055 .115 .061 .096
Crime Reports .124* .235** .210** .242** .207** .135* .245** .269** .316**
Arrest Reports .103 .310** .197** .281** .190** .161** .276** .342** .445**
Case Disposition .259** .217** .204** .272** .274** .293** .348** .314** .426**
Prosecution Disposition .251** .202** .212** .222** .222** .351** .211** .220** .261**

Court Dispositions .272** .209** .192** .318** .213** .324** .250** .218** .254**
Summary Crime Statistics .253** .383** .321** .284** .330** .375** .213** .177** .269**
162
Table 1 (Continued)
Cellular PDA / Web Pages
/ Mobile Handheld MDC / / Web- Laptop
Phones GPS Computer Blackberry MDT Intranet based CBT Computer NIBIN
** ** ** ** ** **
Fingerprints -.004 .207 .268 .244 .273 .219 .203 .206** .208** .214**

Known Offender .028 .116 .129* .037 .219** .088 .185** .144* .103 .106
M.O. File -.017 .252** .217** .124* .223** .100 .200** .105 .194** .154*
Mug Shot .109 .189** .215** .204** .433** .226** .216** .189** .242** .200**
Organized Crime
.083 .336** .266** .195** .161** .150* .239** .243** .187** .110
Intelligence
Narcotics Intelligence .054 .243** .302** .219** .187** .062 .143* .205** .229** .169**

Sex Offender .006 .189** .195** .123* .250** .100 .249** .125* .162** .098

Stolen Property .010 .113 .155* .085 .247** -.026 .192** .078 .149* .004

Stolen Vehicles .034 .141* .149* .094 .306** .008 .160** .111 .051 .123*
163
Table 1 (Continued)
Summary
Crime Arrest Case Prosecution Court Crime
IBIS ViCap RISS NCIC Reports Reports Disposition Disposition Dispositions Statistics
** **
Fingerprints .229 .329 .054 .074 .124* .103 .259** .251** .272** .253**

Known Offender .128* .252** .106 -.037 .235** .310** .217** .202** .209** .383**
M.O. File .204** .282** .205** .024 .210** .197** .204** .212** .192** .321**
Mug Shot .176** .222** .115 .080 .242** .281** .272** .222** .318** .284**
Organized Crime
.141* .294** .229** .079 .207** .190** .274** .222** .213** .330**
Intelligence
Narcotics Intelligence .157* .292** .240** .055 .135* .161** .293** .351** .324** .375**
Sex Offender .107 .186** .139* .115 .245** .276** .348** .211** .250** .213**

Stolen Property .013 .113 .066 .061 .269** .342** .314** .220** .218** .177**

Stolen Vehicles .079 .181** .101 .096 .316** .445** .426** .261** .254** .269**
164
Table 1 (Continued)
Known M.O. Mug Organized Crime Narcotics Sex Stolen Stolen
Fingerprints Offender File Shot Intelligence Intelligence Offender Property Vehicles
Fingerprints 1 .288** .369** .310** .287** .318** .215** .092 .162**
Known Offender .288** 1 .432** .327** .242** .250** .202** .267** .356**
M.O. File .369** .432** 1 .251** .248** .328** .194** .246** .240**
Mug Shot .310** .327** .251** 1 .232** .227** .292** .330** .413**
Organized Crime
.287** .242** .248** .232** 1 .613** .327** .158* .207**
Intelligence
Narcotics Intelligence .318** .250** .328** .227** .613** 1 .358** .229** .235**
Sex Offender .215** .202** .194** .292** .327** .358** 1 .394** .505**
Stolen Property .092 .267** .246** .330** .158* .229** .394** 1 .769**
Stolen Vehicles .162** .356** .240** .413** .207** .235** .505** .769** 1
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
165
 166

APPENDIX C. FACTOR ANALYSIS RESULT

Table 3
The Third Factor Analysis
Factors
1 2 3
Cellular or mobile phones .492 .002 .151
Global Positioning System (GPS) .577 .084 .101
Personal Digital Assistant (PDA) or handheld computer .525 .024 .021
Blackberry .504 .270 .104
Mobile Data Computer/Mobile Data Terminal (MDC/MDT) .503 .009 .005
Intranet .579 .148 -.053
Web pages or web-based .463 .132 -.063
Computer Based Training (CBT) .607 .032 -.028
Laptop computer .496 .000 .129
National Integrated Ballistics Information Network (NIBIN) .479 .013 .200
Integrated Ballistics Identification System (IBIS) .618 .115 .159
Does your department use "ViCap" for investigative purposes? .339 .070 .065
Crime reports .050 .473 .156
Arrest reports -.046 .574 .227
Case disposition .039 .404 .517
Prosecution disposition -.008 .198 .820
Court dispositions .045 .182 .724
Summary crime statistics .200 .291 .358
Fingerprints .402 .136 .312
Known offender .191 .460 .220
M.O. file .325 .279 .235
Mug shot .290 .459 .157
Organized crime intelligence .370 .226 .307
Narcotics intelligence .351 .222 .425
Sex offender. .209 .410 .160
Stolen property -.010 .763 .029
Stolen vehicles .030 .856 .074
 167

APPENDIX D: DESCRIPTIVE STATISTICS

Table 1
2005 Clearance Rates
Zero Valid SE of Std
"=233 Missing Min Max Mean Median
Crime " Mean Dev
Murder 93 13 127 0 2 0.6 0.032 0.6 0.358
Rape 65 13 155 0 2 0.375 0.025 0.333 0.314
Robbery 53 13 167 0 1 0.311 0.019 0.276 0.246
Assault 23 13 197 0 1 0.538 0.018 0.525 0.246
Burglary 24 13 196 0 1.5 0.14 0.011 0.111 0.149
Larceny 21 13 199 0 0.79 0.156 0.008 0.146 0.112
Vehicle
30 13 190 0 1 0.196 0.015 0.143 0.209
Theft
Violent
23 13 197 0 1 0.514 0.017 0.501 0.241
Crimes
Property
21 13 199 0 0.752 0.156 0.008 0.139 0.111
Crime
All
21 13 199 0 0.788 0.261 0.01 0.247 0.134
Crimes
 168

Table 2
2006 Clearance Rates
Zero Valid SE of Std
"=233 Missing Min Max Mean Median
Crime " Mean Dev
Murder 94 17 122 0 1.5 0.619 0.026 0.626 0.29
Rape 64 17 152 0 1 0.365 0.023 0.34 0.281
Robbery 51 17 165 0 2 0.274 0.019 0.25 0.247
Assault 25 17 191 0 1.044 0.54 0.018 0.533 0.251
Burglary 25 17 191 0 0.5 0.122 0.006 0.104 0.086
Larceny 24 17 192 0 0.549 0.145 0.007 0.134 0.091
Vehicle Theft 31 17 185 0 1 0.177 0.013 0.117 0.177
Violent
23 17 193 0 1 0.513 0.018 0.498 0.249
Crimes
Property
23 17 193 0 1 0.145 0.008 0.128 0.104
Crime
All Crimes 22 17 194 0 1 0.253 0.01 0.232 0.134

Table 3
2005 Crime Rates
"=233 Missing Valid " Min Max Mean SE of Mean Median Std Dev
Murder 13 220 0 0.431 0.054 0.006 0.014 0.085
Rape 13 220 0 1.22 0.227 0.018 0.143 0.267
Robbery 13 220 0 9.354 1.429 0.131 0.47 1.946
Assault 13 220 0 62.182 12.023 0.832 9.006 12.346
Burglary 13 220 0 35.545 6.805 0.463 4.971 6.862
Larceny 13 220 0 85.636 20.094 1.206 16.584 17.885
Vehicle Theft 13 220 0 25.691 3.999 0.328 1.924 4.871
Violent Crimes 13 220 0 65.636 13.733 0.942 10.279 13.974
Property Crime 13 220 0 143.089 30.897 1.852 24.883 27.469
All Crimes 13 220 0 203.489 44.663 2.698 33.812 40.025
 169

Table 4
2006 Crime Rates
"=233 Missing Valid " Min Max Mean SE of Mean Median Std Dev
Murder 17 216 0 0.545 0.057 0.006 0.016 0.091
Rape 17 216 0 1.409 0.24 0.019 0.153 0.279
Robbery 17 216 0 9.283 1.545 0.14 0.597 2.057
Assault 17 216 0 56.636 12.307 0.841 9.194 12.362
Burglary 17 216 0 34.273 6.923 0.452 5.762 6.642
Larceny 17 216 0 84.545 20.117 1.22 16.78 17.93
Vehicle Theft 17 216 0 25.531 3.882 0.326 1.79 4.794
Violent Crimes 17 216 0 60 14.15 0.955 10.349 14.031
Property Crime 17 216 0 141.711 30.922 1.852 26.701 27.213
All Crimes 17 216 0 199.8 45.104 2.719 39.385 39.965

Table 5
Use of IT and Control Variables
Valid SE of
"=233 Missing Min Max Mean Median Std Dev
" Mean
Use of IT 34 199 2 29 20.186 0.391 21 5.516
# of Officer 4 229 1 37038 814.681 184.634 210 2794.016
Log of # of
4 229 0 10.52 4.804 0.149 5.347 2.255
Officer
Population 0 233 465 8e+06 398987.399 52438.69 152000 800441.869
Log of
0 233 6.142 15.895 11.145 0.153 11.932 2.339
Population
Median
0 233 19063 148173 41949.232 1086.673 37954 16587.337
Income
% Below
0 233 0.5 40 13.786 0.484 12.7 7.382
Poverty
% White 0 233 1.1 99.2 72.995 1.338 75.1 20.427
Proportion of
7 226 0 3 0.252 0.022 0.154 0.332
Investigators
 APPENDIX E: HISTOGRAMS OF ALL VARIABLES

05 Murder Clr Rate 05 Rape Clr Rate 05 Robbery Clr Rate

50

30
40

40

25
30

20
30

15
20

20

Frequency
Frequency
Frequency

10
10

10

5
0
0
0

0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 0.0 0.2 0.4 0.6 0.8 1.0

05 Assault Clr Rate 05 Burglary Clr Rate 05 Larceny Clr Rate

30
80

150

25
60

20
100

15
40

Frequency
Frequency
Frequency

10
50
20

5
0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5 0.0 0.2 0.4 0.6 0.8

Figure 1. 2005 Clearance Rate Variables

170
 05 Vehicle Theft Clr Rate 05 Violent Crimes Clr Rate 05 Property Crimes Clr Rate

40

70
60
80

30

50
60

40
20

30
40

Frequency
Frequency
Frequency

20
10
20

10
0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8

05 All Crimes Clr Rate

60
50
40
30
Frequency
20
10
0
0.0 0.2 0.4 0.6 0.8

Figure 2.2005 Clearance Rate Variables

171
 05 Sqrt Murder Clr Rate 05 Sqrt Rape Clr Rate 05 Sqrt Robbery Clr Rate

50

40

40
40

30

30
30

20

20
Frequency
Frequency
Frequency
20

10
10
10

0
0
0

0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 0.0 0.2 0.4 0.6 0.8 1.0

05 Sqrt Assault Clr Rate 05 Sqrt Burglary Clr Rate 05 Sqrt Larceny Clr Rate
60

60

40
50

50

30
40

40
30

30

20
Frequency
Frequency
Frequency
20

20

10
10

10

0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8

Figure 3. Square Rooted 2005 Clearance Rate Variables

172
 05 Sqrt Vehicle Theft Clr Rate 05 Sqrt Violent Crimes Clr Rate 05 Sqrt Property Crimes Clr Rate

80

50
40

40
60

30

30
40

20

Frequency
Frequency
Frequency

20
20

10

10
0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8

05 Sqrt All Crimes Clr Rate

30 60
5040
Frequency
20
10
0
0.0 0.2 0.4 0.6 0.8

Figure 4. Square Rooted 2005 Clearance Rate Variables

173
 06 Murder Clr Rate 06 Rape Clr Rate 06 Robbery Clr Rate

80

35
35

30
30
60

25
25

20
20
40

15
15

Frequency
Frequency
Frequency

10
10
20

5
5

0
0
0

0.0 0.5 1.0 1.5 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5 2.0

06 Assault Clr Rate 06 Burglary Clr Rate 06 Larceny Clr Rate

60

30
50
40

25
40
30

20
30

15
20

Frequency
Frequency
Frequency

20

10
10

5
10

0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5

Figure 5. 2006 Clearance Rate Variables

174
 06 Vehicle Theft Clr Rate 06 Violent Crimes Clr Rate 06 Property Crimes Clr Rate

35

80
80

30

60
60

25
20

40
40

15

Frequency
Frequency
Frequency

10

20
20

0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

06 All Crimes Clr Rate

60
50
40
30
Frequency
20
10
0
0.0 0.2 0.4 0.6 0.8 1.0

Figure 6. 2006 Clearance Rate Variables

175
 06 Sqrt Murder Clr Rate 06 Sqrt Rape Clr Rate 06 Sqrt Robbery Clr Rate

30

50
80

25

40
60

20

30
15
40

Frequency
Frequency
Frequency

20
10
20

10
5

0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5

06 Sqrt Assault Clr Rate 06 Sqrt Burglary Clr Rate 06 Sqrt Larceny Clr Rate

70
60

40
60
50

50

30
40

40
30

20
30

Frequency
Frequency
Frequency
20

20

10
10

10

0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 0.0 0.2 0.4 0.6 0.8

Figure 7. Square Rooted 2006 Clearance Rate Variables

176
 06 Sqrt Vehicle Theft Clr Rate 06 Sqrt Violent Crimes Clr Rate 06 Sqrt Property Crimes Clr Rate

40
70

40
60

30
50

30
40

20
20
30

Frequency
Frequency
Frequency
20

10
10
10

0
0
0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

06 Sqrt All Crimes Clr Rate

60
50
40
30
Frequency
20
10
0
0.0 0.2 0.4 0.6 0.8 1.0

Figure 8. Square Rooted 2006 Clearance Rate Variables

177
 05 Murder 05 Rape 05 Robbery

140

100

150
120

80
100

100
60
80
60

40

Frequency
Frequency
Frequency

50
40

20
20

0
0
0

0.0 0.1 0.2 0.3 0.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 2 4 6 8 10

05 Assault 05 Burglary 05 larceny

80
80

100

60
60

80
60

40
40

Frequency
Frequency
Frequency

40

20
20

20

0
0
0

0 10 20 30 40 50 60 0 10 20 30 40 0 20 40 60 80

Figure 9. 2005 Crime Rate Variables

178
 05 Vehicle Theft 05 Violent Crimes 05 Property Crimes

100

150
100
80

80

100
60

60
40

Frequency
Frequency
Frequency

40

50
20

20

0
0
0

0 5 10 15 20 25 30 0 10 20 30 40 50 60 70 0 50 100 150

05 All Crimes

60
40
Frequency
20
0
0 50 100 150 200

Figure 10. 2005 Crime Rate Variables

179
 06 Murder 06 Rape 06 Robbery

120
120

100
100

100 120 140

80
80

80
60
60

60
Frequency
Frequency
Frequency

40
40

40
20
20

20
0
0
0

0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.5 1.0 1.5 0 2 4 6 8 10

06 Assault 06 Burglary 06 larceny

80
80

100
80

60
60

60

40
40

Frequency
Frequency
Frequency

40

20
20

20

0
0
0

0 10 20 30 40 50 60 0 5 10 15 20 25 30 35 0 20 40 60 80

Figure 11. 2006 Crime Rate Variables

180
 06 Vehicle Theft 06 Violent Crimes 06 Property Crimes

150
80

60

100
60

40
40

Frequency
Frequency
Frequency

50
20
20

0
0
0

0 5 10 15 20 25 30 0 10 20 30 40 50 60 0 50 100 150

06 All Crimes

60
40
Frequency
20
0
0 50 100 150 200

Figure 12. 2006 Crime Rate Variables

181
 Population Served # of Sworn Officers Proportion of Investigators

200

200
200
150

150
150
100

100
100

Frequency
Frequency
Frequency
50

50
50

0
0
0

0e+00 2e+06 4e+06 6e+06 8e+06 0 10000 20000 30000 40000 0.0 0.5 1.0 1.5 2.0 2.5 3.0

Median Income % of Population Below Poverty % of Population White

120
60
60

50
50

100
80
40
40

60
30
30

Frequency
Frequency
Frequency

40
20
20

20
10
10

0
0
0

0 50000 100000 150000 0 10 20 30 40 0 20 40 60 80 100

Figure 13. Control Variables

182
 Log of Population Served Log of # of Sworn Officers

60

40
50

30
40
30

20
Frequency
Frequency
20

10
10

0
0
6 8 10 12 14 16 0 2 4 6 8 10

Figure 14. 2005 Logged Control Variables

183
 APPENDIX F: CORRELATION ANALYSIS RESULTS

Table 1.
Correlation Coefficients of Use of IT Scale and 2005 Square Rooted Clearance Rates
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crimes Crimes
Use of IT .013 .176* .017 -.009 .075 .072 -.090 -.031 .009 -.009
N 111 137 147 171 170 173 166 171 173 173
6ote. ***p< 0.001, **p< 0.01, *p< 0.05

Table 2.
Correlation Coefficients of Use of IT Scale and 2006 Square Rooted Clearance Rates
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crimes Crimes
Use of IT .031 .252** .202* -.029 .122 .042 -.078 -.037 .025 -.031
N 109 135 147 167 168 168 162 168 169 169
6ote. ***p< 0.001, **p< 0.01, *p< 0.05

Table 3.
Correlation Coefficients of Use of IT Scale and Six Control Variables
Logged # of Logged Median % Below Prop. Of
Officer Population Income Poverty % White Invest.
Use of IT 0.627** 0.62** 0.178* -0.164* -0.272** -0.08
N 195 199 199 199 199 193
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
184
Table 4.
Correlation Coefficients of 2005 Square Rooted Clearance Rates and Six Control Variables
Logged # of Logged Median % Below Prop. Of
Officer Population Income Poverty % White Invest.
Murder -.028 .019 -.074 -.072 .188* .021
N 126 127 127 127 127 124
Rape .195* .132 .059 -.019 -.055 -.041
N 153 155 155 155 155 151
Robbery -.235** -.212** .109 -.151 .314** .076
N 165 167 167 167 167 163
Assault -.137 -.108 .215** -.185** .244** .083
N 195 197 197 197 197 193
Burglary -.188** -.179* .124 -.129 .184** .324**
N 194 196 196 196 196 192
Larceny -.040 -.032 .117 -.137 .161* .123
N 197 199 199 199 199 194
Vehicle Theft -.245** -.200** .145* -.211** .294** .150*
N 188 190 190 190 190 186
Violent Crimes -.178* -.149* .216** -.198** .279** .083
N 195 197 197 197 197 193
Property Crimes -.207** -.194** .140* -.189** .270** .181*
N 197 199 199 199 199 194
All Crimes -.161* -.143* .101 -.155* .243** .138
N 197 199 199 199 199 194
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
185
Table 5.
Correlation Coefficients of 2006 Square Rooted Clearance Rates and Six Control Variables
Logged # of Logged Median % Below Prop. Of
Officer Population Income Poverty % White Invest.
Murder ** *
-.286 -.182 -.035 -.124 .338** -.008
N 121 122 122 122 122 119
Rape ** **
.220 .220 .009 .056 -.086 .108
N 150 152 152 152 152 147
Robbery .007 .011 .156* -.148 .267** -.154
N 163 165 165 165 165 161
Assault -.189** -.158* .199** -.216** .273** -.001
N 189 191 191 191 191 187
Burglary *
-.025 -.005 .182 -.115 .100 .012
N 189 191 191 191 191 187
Larceny -.057 -.044 .054 -.082 .153* .048
N 190 192 192 192 192 187
Vehicle Theft ** *
-.228 -.172 -.001 -.118 .277** .091
N 183 185 185 185 185 181
Violent Crimes -.159* -.130 .223** -.254** .320** .015
N 191 193 193 193 193 188
Property Crimes -.087 -.071 .077 -.119 .212** .036
N 191 193 193 193 193 188
All Crimes -.054 -.036 .054 -.132 .239** .007
N 192 194 194 194 194 189
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
186
Table 6.
Correlation Coefficients of 2005-2006 Crime Rates and use of IT scale
2005 2006
Use of IT Use of IT
Murder .271** .264**
N 189 186
**
Rape .205 .114
N 189 186
**
Robbery .251 .261**
N 189 186
Assault .130 .135
N 189 186
Burglary .108 .135
N 189 186
Larceny .139 .138
N 189 186
**
Vehicle Theft .297 .289**
N 189 186
*
Violent .155 .161*
N 189 186
*
Property .169 .175*
N 189 186
All Crimes .170* .176*
N 189 186
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
187
 188

APPENDIX G: MODERATE AND HIGH IT DEPARTMENT COMPARISON

Murder Rape
H

H
M

M
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Robbery Assault
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Burglary Larceny
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Vehicle Theft Violent Crimes
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Property Crimes All Crimes
H

H
M

0.0 0.2 0.4 0.6 0.8 0.0 0.2 0.4 0.6 0.8
Figure 1. Box plot comparison of 2005 square rooted clearance rates between low and
high IT departments
 189

H Murder Rape

H
M

M
0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0
Robbery Assault
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 M 0.0 0.2 0.4 0.6 0.8 1.0
Burglary Larceny
H

H
M

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.0 0.2 0.4 0.6
Vehicle Theft Violent Crimes
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Property Crimes All Crimes
H

H
M

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Figure 2. Box plot comparison of 2006 square rooted clearance rates between low and
high IT departments
 190

H Murder Rape

H
M

M
0.0 0.1 0.2 0.3 0.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2
Robbery Assault
H

H
M

M
0 2 4 6 8 0 10 20 30 40 50 60
Burglary Larceny
H

H
M

0 5 10 15 20 25 30 35 0 20 40 60 80
Vehicle Theft Violent Crimes
H

H
M

0 5 10 15 20 25 0 10 20 30 40 50 60
Property Crimes All Crimes
H

H
M

0 20 40 60 80 100 120 140 0 50 100 150 200

Figure 3. Box plot comparison of 2005 crime rates between low and high IT departments
 191

Murder Rape
H

H
M

M
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Robbery Assault
H

H
M

M
0 2 4 6 8 0 10 20 30 40 50
Burglary Larceny
H

H
M

0 5 10 15 20 25 30 35 0 20 40 60 80
Vehicle Theft Violent Crimes
H

H
M

0 5 10 15 20 25 0 10 20 30 40 50 60
Property Crimes All Crimes
H

H
M

0 20 40 60 80 100 120 140 0 50 100 150 200

Figure 4. Box plot comparison of 2005 crime rates between low and high IT departments
 192

H
M Logged # of Officer

2 4 6 8 10
Logged Population
H
M

6 8 10 12 14 16
Median Income
H
M

20000 40000 60000 80000 100000 120000 140000

% Below Poverty
H
M

0 10 20 30
% White
H
M

20 40 60 80 100
Proportion of Investigators
H
M

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Figure 5. Box plot comparison of six control variables between low and high IT
departments
 193

APPENDIX H: DECILES OF USE OF IT SCALE

10 1st Decile 2nd Decile

L o g g ed # o f O fficer

L o g g ed # o f O fficer
8
8

6
6

4
4
2

2
0

2 4 6 8 10 12 13.0 13.5 14.0 14.5 15.0 15.5 16.0

Use of IT Use of IT

3rd Decile 4th Decile

7
7
L o g g ed # o f O fficer

L o g g ed # o f O fficer
6
6
5

5
4

4
3

3
2
1

16.0 16.5 17.0 17.5 18.0 18.0 18.5 19.0 19.5 20.0
Use of IT Use of IT

5th Decile 6th Decile

7
L o g g ed # o f O fficer

L o g g ed # o f O fficer
6

6
5

5
4

4
3

3
2
2

20.0 20.2 20.4 20.6 20.8 21.0 21.0 21.2 21.4 21.6 21.8 22.0
Use of IT Use of IT

Figure 1. Scatter plots of logged number of officer and use of IT scale variable (1-6)
 194

7th Decile 8th Decile

9
8

8
7
Logged # of Officer

Logged # of Officer
6

7
5

6
4

5
3

4
2

3
1

22.0 22.5 23.0 23.5 24.0 24.0 24.2 24.4 24.6 24.8 25.0
Use of IT Use of IT

9th Decile 10th Decile

7

9
Logged # of Officer

Logged # of Officer
6

8
7
5

6
4

5
3

15 20 25 30 35 27.0 27.5 28.0 28.5 29.0

Use of IT Use of IT

Figure 2. Scatter plots of logged number of officer and use of IT scale variable (7-10)
 195

APPENDIX I: TEN GROUPS OF LOGGED NUMBER OF OFFICERS

x<1 N=3 1>=x<2 r=-0.016 N=22

0.6

1.8
1.6
0.4

1.4
0.2

1.2
0.0

2 4 6 8 10 12 14 16 5 10 15 20
2>=x<3 r=0.42* N=30 3>=x<4 r=0.32* N=23
3.0

3.8
2.8

3.6
2.6

3.4
2.4

3.2
2.2

5 10 15 20 25 10 15 20 25
4>=x<5 r=0.55* N=14 5>=x<6 r=-0.57 N=16
5.0

6.0
4.8

5.8
4.6

5.6
4.4

5.4
4.2

5.2

16 18 20 22 24 26 28 16 18 20 22 24 26 28
Figure 1. Correlations and scatterplots of number of officer and use of IT variables (1-6)
 196

6>=x<7 r=0.073 N=57 7>=x<8 r=-0.027 N=23

8.0
6.8

7.8
6.6

7.6
6.4

7.4
7.2
6.2

7.0
6.0

15 20 25 16 18 20 22 24 26 28
8>=x<9 N=3 x>9 N=4

10.5
8.8
8.6

10.0
8.4

9.5
8.2

9.0

22 23 24 25 26 27 28 29 15 20 25
Figure 2. Correlations and scatterplots of number of officer and use of IT variables (7-
10)
 APPENDIX J: REGRESSION ANALYSIS RESULTS

TABLE 1
Reduced Model Fitted to 2005 Data
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Thef Crimes Crime Crimes
Intercept 0.769 0.245 0.527 0.715 0.293 0.324 0.484 0.719 0.368 0.497
Use of IT 0.002 0.007 0.008 0.005 0.008** 0.005 0.004 0.005 0.006** 0.005
Log of #
-0.015 0.025 -0.035* -0.024* -0.024*** -0.012 -0.035* -0.028** -0.025 -0.021*
Officer

F 0.252 3.119* 3.686* 3.133* 6.602** 2.230 7.397*** 4.313 9.634*** 5.132**
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 107 132 142 3 165 168 161 166 168 168
R -0.014 0.031 0.036 0.025 0.063 0.014 0.073 0.038 0.092 0.046
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
197
TABLE 2
Complex Model Fitted to 2005 Data
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.008 -0.177 -0.054 0.239 0.036 0 0.134 0.244 0.06 0.202
Use of IT 0.004 0.009 0.008 0.003 0.006* 0.003 0.003 0.003 0.005* 0.003
Log of #
0.011 0.026 -0.016 -0.006 -0.008 0.005 -0.013 -0.01 -0.008 -0.003
Officer
Median
0 0 0 0 0 0 0 0 0 0
Income
% Below
0.007 0.007 0.007 0.002 0.002 0.001 0.003 0.002 0.002 0
Poverty
% White 0.007* 0.002 0.004* 0.003* 0.002** 0.002* 0.002 0.003* 0.002** 0.002*
Prop. of
0.076 -0.004 0.12 0.063 0.184 0.086* 0.096 0.06 0.09** 0.091*
Invest.
Crime Rate 0.126 0.009 -0.004 0.003* 0.001 0.001 -0.006 0.002 0 0

F 1.429 1.040 2.352* 3.457** 6.607 3.34** 4.306*** 3.876*** 5.696*** 3.644**
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 101 126 136 160 159 161 155 160 161 161
R 0.027 0.002 0.062 0.093 0.191 0.089 0.125 0.108 0.164 0.099
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
198
TABLE 3
Reduced Model Fitted to 2006 Data
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.962 0.159 0.259 0.748 0.257 0.339 0.455 0.737 0.354 0.514
Use of IT 0.007 0.011 0.015** 0.008* 0.006* 0.004 0.005 0.007 0.004 0.003
Log of #
-0.06** 0.023 -0.023 -0.038*** -0.01 -0.012 -0.036*** -0.038*** -0.014* -0.016*
Officer

F 5.641** 5.369** 4.667* 8.302*** 2.489 2.048 7.371 9.03*** 2.670 2.860
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 105 130 142 162 163 163 157 163 164 164
R 0.080 0.062 0.048 0.082 0.018 0.013 0.074 0.089 0.020 0.022
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
199
TABLE 4
Reduced Model Fitted to 2006 Data
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.603 0.379 -0.555 0.461 0.008 0.09 0.416 0.422 0.161 0.313
Use of IT 0.006 0.012 0.018*** 0.006 0.005 0.003 0.002 0.006 0.003 0.002
Log of #
-0.032 0.025 -0.02 -0.025* -0.003 0.001 -0.01 -0.028* -0.006 -0.006
Officer
Median
0 0 0 0 0 0 0 0 0 0
Income
% Beci
0 -0.006 0.01 -0.001 0.003 0 -0.005 -0.001 0 -0.001
Poverty
% White 0.004 0 0.006*** 0.002 0.001 0.002* 0.002 0.003* 0.002* 0.002*
Prop. of
-0.023 0.164 -0.19* 0.002 0.007 0.04 0.119* 0 0.027 0.016
Invest.
Crime
-0.045 0.007 0.02 0.004** 0 0.002** -0.006 0.003* 0.001* 0.001**
Rate

F 2.726* 2.081 4.558*** 4.65*** 1.889 2.505* 4.371*** 5.552*** 3.161** 3.098**
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 99 123 136 156 157 156 151 156 157 157
R 0.102 0.055 0.148 0.136 0.037 0.061 0.130 0.164 0.084 0.082
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
200
 APPENDIX K REGRESSION ANALYSIS OF MODERATE/HIGH IT GROUPS

Table 1
2005 Moderate IT Reduced Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.272 -0.050 0.303 0.477 0.249 0.251 0.267 0.478 0.287 0.367
Use of IT 0.035 0.022 0.025 0.019 0.010 0.010 0.014 0.019 0.010 0.012
Log of #
-0.021 0.040 -0.035 -0.019 -0.016 -0.008 -0.019 -0.022 -0.015 -0.013
Officer

F 0.704 1.931 1.687 1.709 1.539 0.823 1.016 1.968 2.088 1.651
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 16 32 38 51 50 51 48 51 51 51
R -0.034 0.052 0.033 0.026 0.020 -0.007 0.001 0.035 0.039 0.024
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
201
Table 2
2005 Moderate IT Complex Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept -1.911 1.160 0.877 -0.044 0.093 0.056 -1.008 0.007 0.013 0.116

Use of IT 0.049 0.006 0.027 0.021 0.019 0.014 0.032* 0.021 0.016 0.017

Log of #
0.047 0.043 -0.012 -0.010 -0.024 -0.002 -0.007 -0.016 -0.013 -0.010
Officer
Median
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Income
% Below
0.009 -0.016 -0.009 0.006 0.005 0.000 0.031** 0.005 0.004 0.003
Poverty
% White 0.014 0.000 -0.001 0.002 -0.001 0.000 0.004 0.002 0.000 0.001
Prop. of
0.438 -0.433 0.174 0.146 0.110 0.203 0.017 0.133 0.114 0.136
Invest.
Crime Rate 1.994 -0.120 -0.043 0.002 -0.003 0.000 -0.036* 0.002 -0.001 0.000

F 0.453 1.188 0.940 0.998 0.941 1.039 2.162 0.987 1.197 0.775
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 10 26 32 45 44 45 42 45 45 45
R -0.291 0.038 -0.011 0.000 -0.008 0.005 0.142 -0.002 0.026 -0.031
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
202
Table 3
2006 Moderate IT Reduced Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept -0.304 -0.140 -0.260 0.548 0.378 0.160 0.512 0.494 0.284 0.369
Use of IT 0.077** 0.029 0.049* 0.021 -0.001 0.015 -0.001 0.023 0.009 0.013
Log of #
-0.036 0.037 -0.014 -0.036** -0.007 -0.009 -0.027 -0.038** -0.013 -0.017
Officer

F 7.44** 1.933 3.417* 4.797* 0.254 2.031 1.402 5.763** 1.583 2.437
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 19 35 40 51 51 51 49 51 51 51
R 0.380 0.048 0.103 0.125 -0.029 0.037 0.016 0.152 0.022 0.051
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
203
Table 4
2006 Moderate IT Complex Model
Vehicle Violent Property
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime All Crimes
Intercept -0.712 -0.239 -0.536 0.609 0.329 0.074 0.284 0.511 0.243 0.339

Use of IT 0.070 0.030 0.041 0.017 -0.001 0.013 0.001 0.020 0.007 0.011

Log of #
0.016 0.062 -0.014 -0.036 -0.004 -0.003 -0.011 -0.038 -0.007 -0.012
Officer
Median
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Income
% Below
-0.003 -0.006 0.002 -0.003 0.000 -0.002 0.006 -0.003 -0.002 -0.002
Poverty
% White 0.008 0.002 0.003 -0.001 -0.001 0.001 0.003 0.000 0.000 0.000
Prop. of
0.373 0.605 -0.276 0.060 0.055 0.070 -0.043 0.059 0.082 0.074
Invest.
Crime Rate -0.012 0.099 0.016 0.002 -0.003 0.001 -0.014 0.001 0.000 0.000

F 2.226 1.665 1.535 1.399 1.719 0.827 1.010 1.705 0.827 0.724
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 13 29 34 45 45 45 43 45 45 45
R 0.300 0.115 0.084 0.051 0.088 -0.024 0.001 0.087 -0.024 -0.039
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
204
Table 5
2005 High IT Reduced Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.819 0.361 0.705 0.782 0.298 0.376 0.465 0.805 0.427 0.549
Use of IT -0.003 0.005 -0.001 0.003 0.009 0.003 0.009 0.002 0.005 0.004

Log of # - -
-0.004 0.014 -0.032* -0.026 -0.031** -0.016 -0.03* -0.027**
Officer 0.049*** 0.032***

F 0.051 0.440 2.616 2.036 5.141** 1.652 9.14*** 2.921 8.896*** 4.450*
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 88 96 99 108 108 110 107 108 110 110
R -0.022 -0.012 0.031 0.018 0.070 0.012 0.130 0.034 0.124 0.058
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
205
Table 6
2005 High IT Complex Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept -0.023 -0.499 -0.032 0.419 0.137 0.168 0.611 0.442 0.286 0.398

Use of IT -0.007 0.007 -0.003 0.000 0.004 -0.002 0.002 -0.002 0.000 -0.002
Log of #
0.024 0.020 -0.015 -0.003 -0.006 0.004 -0.021 -0.007 -0.010 -0.003
Officer
Median
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Income
% Below
0.015 0.016 0.009 -0.002 -0.002 0.000 -0.012* -0.002 -0.002 -0.002
Poverty

% White 0.008 0.003 0.005*** 0.003* 0.002* 0.003** 0.001 0.003* 0.002* 0.003**

Prop. of
0.057 0.162 0.118 0.057 0.215*** 0.085* 0.122* 0.057 0.099** 0.091*
Invest.
Crime
-0.026 -0.035 0.008 0.003 0.001 0.001 -0.003 0.003* 0.000 0.000
Rate

F 1.477 0.870 3.911*** 2.928** 7.716*** 3.096** 6.873*** 3.495** 6.129*** 3.675**
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 83 91 94 103 103 104 102 103 104 104
R 0.036 -0.009 0.168 0.109 0.299 0.117 0.274 0.137 0.244 0.144
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
206
Table 7
2006 High IT Reduced Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.860 0.109 0.354 0.786 0.214 0.354 0.588 0.813 0.451 0.577
Use of IT 0.011 0.015 0.011 0.006 0.008 0.003 0.003 0.003 0.000 -0.001
Log of #
-0.063* 0.013 -0.024 -0.037** -0.015 -0.013 -0.048*** -0.036** -0.013 -0.013
Officer

F 3.246* 1.484 1.476 3.848* 1.920 1.136 9.659*** 4.012* 1.513 1.203
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 83 90 97 104 105 105 102 105 106 106
R 0.050 0.010 0.010 0.051 0.017 0.003 0.143 0.053 0.009 0.004
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
207
Table 8
2006 High IT Complex Model
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.264 0.407 -0.457 0.472 0.097 0.162 0.573 0.457 0.285 0.406

Use of IT 0.009 0.017 0.011 0.003 0.007 -0.001 -0.004 0.003 0.001 0.000
Log of #
-0.041 0.003 -0.021 -0.023 -0.008 0.002 -0.017 -0.027 -0.009 -0.007
Officer
Median
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Income
% Below
0.009 -0.006 0.013* -0.001 0.001 0.000 -0.006 -0.001 -0.002 -0.002
Poverty
% White 0.005 -0.001 0.006*** 0.003* 0.001 0.002* 0.002 0.003* 0.002* 0.002*
Prop. of
-0.060 -0.033 -0.107 -0.005 0.029 0.050 0.115* -0.005 0.031 0.013
Invest.
Crime Rate -0.013 -0.006 0.016 0.004* 0.001 0.002* -0.005 0.004** 0.001* 0.001**

F 1.643 0.553 3.652* 3.061** 0.885 2.044 6.369*** 3.632** 2.782* 2.551*
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 78 84 92 99 100 99 97 99 100 100
R 0.050 -0.036 0.158 0.120 -0.008 0.064 0.265 0.148 0.104 0.092
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
208
 APPENDIX L: REGRESSION ANALYSIS OF TEN GROUPS

Table 1
2005 Third Group Reduced Models

Assault Burglary Larceny Vehicle Theft Violent Crimes Property Crime All Crimes
Intercept 0.55 0.381 0.357 0.136 0.585 0.423 0.419

Use of IT 0.016 0.009 0.013* 0.005 0.015 0.011* 0.009

Log of # Officer -0.029 -0.073 -0.078 0.088 -0.042 -0.083 -0.022

F 1.398 0.682 2.919 0.350 1.285 3.331 1.315

DF Reg 2 2 2 2 2 2 2
DF Res 23 22 23 21 23 23 23
R 0.031 -0.027 0.133 -0.060 0.022 0.157 0.025
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
209
Table 2
2005 Third Group Complex Models

3 Assault Burglary Larceny Vehicle Theft Violent Crimes Property Crime All Crimes
Intercept -0.865 0.419 -0.135 -0.655 -0.857 0.022 -0.183

Use of IT 0.028* 0.006 0.018* 0.02 0.028* 0.014* 0.015*

Log of # Officer -0.045 -0.011 -0.06 -0.059 -0.059 -0.053 -0.024

Median Income 0 0 0 0 0 0 0

% Below Poverty 0.033 -0.012 0.009 0.033 0.034 0.004 0.013

% White 0.008 -0.001 0.001 0.006 0.008 0.001 0.003

Prop. of Invest. 0.251 -0.018 -0.055 0.251 0.252 -0.009 0.064

Crime Rate -0.004 0.013 0.003 -0.087 -0.004 0.002 0

F 1.115 0.479 1.782 0.783 1.118 1.553 0.781

DF Reg 7 7 7 7 7 7 7
DF Res 18 17 18 16 18 18 18
R 0.031 -0.179 0.180 -0.071 0.032 0.134 -0.065
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
210
Table 3
2005 Fourth Group Reduced Models
Vehicle Violent Property All
4 Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 1.611 0.294 0.44 0.008 -0.156 0.313 0.082 0.022
Use of IT 0.019 0.012 0.012 0.01 0.026 0.013 0.016 0.012
Log of #
-0.379 0.073 -0.089 0.063 0.052 0.058 0.011 0.081
Officer

F 0.7387807 1.19679 0.99096 0.95033 2.28151 1.17717 1.81029 1.99788

DF Reg 2 2 2 2 2 2 2 2
DF Res 13 16 16 16 15 16 16 16
R -0.036086 0.0214 -0.001 -0.0055 0.13101 0.01931 0.0826 0.09981
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
211
Table 4
2005 Fourth Group Complex Models
Violent All
Robbery Assault Burglary Larceny Vehicle Theft Crimes Property Crime Crimes
Intercept 0.267 0.217 0.149 -0.471 -1.04 0.264 -0.232 -0.24

Use of IT 0.039 0.012 0.014 0.009 0.027 0.013 0.015 0.012

Log of # Officer -0.442 0.152 -0.1 0.023 0.191 0.16 -0.002 0.115

Median Income 0 0 0 0 0 0 0 0
% Below
0.024 -0.011 0.004 0.002 -0.006 -0.013 -0.003 -0.007
Poverty
% White 0.015 0.001 0.003 0.002 0.008 0.001 0.001 0.002

Prop. of Invest. -1.587 0.425 -0.132 1.689 0.063 0.266 1.42 0.937

Crime Rate -0.11 -0.001 -0.003 0.001 0.003 -0.001 0.001 0

F 0.7995851 0.61149 0.40097 0.77726 1.02231 0.70865 0.77901 0.9027

DF Reg 7 7 7 7 7 7 7 7
DF Res 8 11 11 11 10 11 11 11
R -0.103177 -0.178 -0.3037 -0.0948 0.0091 -0.1278 -0.094 -0.0393
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
212
Table 5
2005 Seventh Group Reduced Models
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 1.258 0.983 0.694 0.907 0.353 0.45 0.766 0.938 0.456 0.731
Use of IT -0.005 0 -0.001 0 0.004 0 0.005 0 0.001 -0.001
Log of #
-0.063 -0.07 -0.029 -0.031 -0.017 -0.011 -0.081 -0.038 -0.02 -0.032
Officer

F 0.2892159 0.15912 0.07937 0.05192 0.48524 0.02503 1.24907 0.08411 0.1342 0.14362
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 51 50 53 53 53 53 53 53 53 53
R -0.027561 -0.0334 -0.0346 -0.0357 -0.0191 -0.0368 0.00898 -0.0345 -0.0325 -0.0321
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
213
Table 6
2005 Seventh Group Complex Models
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.559 0.902 0.278 0.165 0.196 0.453 0.202 0.254 0.427 0.549

Use of IT -0.004 -0.012 -0.007 -0.005 0.001 -0.005 0.003 -0.007 -0.003 -0.006
Log of #
-0.04 -0.001 0.008 0.03 -0.013 0.027 -0.051 0.022 0.008 -0.007
Officer
Median
0 0 0 0 0 0 0 0 0 0
Income
% Below
0.009 -0.015 -0.001 -0.003 0.001 -0.009 0.008 -0.004 -0.006 -0.005
Poverty
% White 0.007* -0.002 0.004 0.004 0.002 0.002 0.004 0.004 0.002 0.003
Prop. of
0.062 0.282 0.098 0.043 -0.027 0.048 0.016 0.042 0.028 -0.007
Invest.
Crime
0.032 0.11 -0.002 0.004* 0 0.001 -0.012* 0.004* 0 0.001
Rate

F 1.1013946 1.60272 2.13397 2.519* 1.28004 1.73095 3.474** 2.855* 1.80434 1.55946
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 46 45 48 48 48 48 48 48 48 48
R 0.0132148 0.07505 0.12612 0.16203 0.03441 0.08511 0.2395 0.19102 0.09286 0.06647
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
214
Table 7
2005 Eighth Group Reduced Models
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 2.459 1.732 0.791 0.715 0.83 0.378 0.669 0.706 0.55 0.714
Use of IT 0.009 0.018 0.013 0.026 0.012 0.012 0.002 0.024 0.008 0.015
Log of #
-0.257 -0.214 -0.089 -0.087 -0.106 -0.037 -0.051 -0.085 -0.053 -0.082
Officer

F 1.4921966 1.16693 0.8674 2.20292 2.4016 1.34015 0.12064 2.10157 0.99213 1.75856
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 19 18 19 19 19 19 19 19 19 19
R 0.0447769 0.01642 -0.0128 0.10279 0.11777 0.03138 -0.0914 0.09495 -0.0008 0.06738
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
215
Table 8
2005 Eighth Group Complex Models
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 2.193 -1.182 -0.831 -0.715 0.712 0.215 1.031 -0.678 0.429 0.329

Use of IT 0.011 0.06 0.027 0.034 0.01 0.011 -0.014 0.035 0.005 0.018

Log of #
-0.337 -0.109 -0.046 -0.052 -0.136 -0.072 -0.087 -0.052 -0.087 -0.097
Officer
Median
0 0 0 0 0 0 0 0 0 0
Income
% Below
0.009 0.025 0.018 0.017 0.007 0.007 -0.002 0.016 0.006 0.007
Poverty
% White 0.007 0.004 0.007* 0.008* 0.003 0.004* 0.004 0.007* 0.004* 0.005
Prop. of
0.104 -0.218 -0.132 -0.199 0.222 0.097 -0.04 -0.192 0.071 -0.014
Invest.
Crime
0.099 0.384 0.028 0.005 0.001 0 -0.003 0.004 0 0
Rate

F 1.2785664 1.0579 1.37125 1.83144 2.12269 1.94026 1.05699 1.87618 1.70383 1.19337
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 13 12 13 13 13 13 13 13 13 13
R 0.0888368 0.02089 0.115 0.22541 0.2821 0.24761 0.01956 0.23469 0.19765 0.06339
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
216
Table 9
2006 Third Group Reduced and Complex Models
Vehicle Vehicle Violent Violent Property Property All All
LarcenyR Larceny TheftR Theft CrimesR Crimes CrimeR Crime CrimesR Crimes
Intercept 0.614 0.238 0.465 -0.288 0.636 -0.609 0.576 0.261 0.581 -0.045
Use of IT 0.009 0.011 -0.011 -0.013 0.016 0.025 0.006 0.007 0.006 0.011

Log of #
-0.149 -0.084 0.05 0.176 -0.056 -0.037 -0.108 -0.065 -0.061 -0.037
Officer
Median
0 0 0 0 0
Income
% Below
0.002 -0.01 0.022 0.001 0.009
Poverty
% White 0 0.006 0.007 0.001 0.003
Prop. of
-0.04 0.366 0.207 0.002 0.067
Invest.
Crime
0.004* 0.156 0 0.002 0.001
Rate

F 1.420 1.199 0.239 1.497 1.316 0.984 0.859 0.638 0.516 0.522
DF Reg 2 7 2 7 2 7 2 7 2 7
DF Res 23 18 21 16 23 18 23 18 23 18
R 0.033 0.053 -0.071 0.131 0.025 -0.004 -0.011 -0.113 -0.040 -0.155
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
217
Table 10
2006 Fourth Group Reduced Models
Vehicle Violent Property All
Rape Robbery Burglary Larceny Theft Crimes Crime Crimes
Intercept 1.07 -1.068 0.094 -0.175 -0.484 0.728 -0.068 0.168
Use of IT 0.039 0.025 0.009 0.005 0.002 0.016 0.007 0.008
Log of #
-0.381 0.297 0.032 0.133 0.25 -0.083 0.092 0.052
Officer

F 1.903 1.652 1.395 1.324 0.829 1.328 1.434 0.946

DF Reg 2 2 2 2 2 2 2 2
DF Res 13 13 17 17 16 17 17 17
R 0.107 0.080 0.040 0.033 -0.019 0.033 0.044 -0.006
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
218
Table 11
2006 Fourth Group Complex Models
Vehicle Violent Property All
Rape Robbery Burglary Larceny Theft Crimes Crime Crimes
Intercept -0.147 -1.166 -0.007 -0.657 -0.975 0.657 -0.464 -0.255
Use of IT 0.046 0.027 0.012 0.004 0.015 0.012 0.008 0.007
Log of # Officer -0.12 0.262 -0.003 0.203 0.215 0.038 0.135 0.143
Median Income 0 0 0 0 0 0 0 0
% Bedort Poverty -0.009 -0.018 0.003 0 0.005 -0.012 0.002 -0.005
% White 0.005 0.006 0.001 0.003 0.006 0 0.002 0.003
Prop. of Invest. 0.438 -0.447 0.019 0.211 -0.558 0.279 0.137 0.184
Crime Rate 0.369 0.114 -0.01 0 -0.019 -0.002 -0.001 0

F 1.053 0.902 2.402 0.875 1.408 1.159 1.106 1.098

DF Reg 7 7 7 7 7 7 7 7
DF Res 8 8 12 12 11 12 12 12
R 0.024 -0.048 0.341 -0.048 0.137 0.055 0.037 0.035
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
219
Table 12
2006 Seventh Group Reduced Models
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.688 0.842 0.459 0.704 0.117 0.342 0.289 0.716 0.295 0.554
Use of
-0.001 0.001 0.003 -0.001 0.004 -0.001 0.006 0 0.001 -0.001
IT
Log of #
0.013 -0.055 -0.01 0.001 0.02 0.007 -0.013 -0.005 0.004 -0.006
Officer

F 0.018 0.096 0.137 0.003 0.440 0.023 0.710 0.002 0.059 0.022
DF Reg 2 2 2 2 2 2 2 2 2 2
DF Res 50 51 52 53 53 52 52 53 53 53
R -0.039 -0.035 -0.033 -0.038 -0.021 -0.038 -0.011 -0.038 -0.035 -0.037
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
220
Table 13
2006 Seventh Group Complex Models
Vehicle Violent Property All
Murder Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.811 1.725 0.698 0.651 0.397 0.557 0.15 0.745 0.587 0.799
Use of IT -0.005 -0.012 0.001 -0.003 0.002 -0.005 0.004 -0.004 -0.002 -0.005
Log of #
0.022 0.013 -0.031 0.016 0.013 0.028 -0.004 0.004 0.01 -0.004
Officer
Median
0 0 0 0 0 0 0 0 0 0
Income
% Below
-0.004 -0.031 -0.003 -0.007 -0.005 -0.009 0.003 -0.008 -0.008 -0.008
Poverty
% White 0.003 -0.004 0.004 0.002 0.001 0.001 0.003 0.002 0.001 0.002
Prop. of
-0.025 0.079 -0.197 -0.196 -0.092 -0.053 -0.083 -0.199 -0.067 -0.132
Invest.
Crime Rate -0.266 0.171 -0.002 0.005* 0 0.001 -0.009* 0.004* 0 0

F 0.613 0.839 2.079 1.312 0.685 1.113 1.511 1.361 0.866 0.967
DF Reg 7 7 7 7 7 7 7 7 7 7
DF Res 45 46 47 48 48 47 47 48 48 48
R -0.055 -0.022 0.123 0.038 -0.042 0.014 0.062 0.044 -0.017 -0.004
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
221
Table 14
2006 Eighth Group Reduced Models
Vehicle Violent Property All
Rape Robbery Assault Burglary Larcen Theft Crimes Crime Crimes
Intercept 1.995 1.151 0.534 1.241 0.351 0.719 0.535 0.607 0.635
Use of IT 0.015 0.016 0.026 0.009 0.01 0 0.025 0.007 0.014
Log of #
-0.237 -0.143 -0.066 -0.156 -0.031 -0.057 -0.066 -0.058 -0.07
Officer

F 0.790 2.381 1.863 3.197 0.920 0.123 1.798 0.683 1.304

DF Reg 2 2 2 2 2 2 2 2 2
DF Res 51 52 53 53 52 52 53 53 53
R -0.025 0.127 0.083 0.188 -0.008 -0.102 0.078 -0.035 0.031
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
222
Table 15
2006 Eighth Group Complex Models
Vehicle Violent Property All
Rape Robbery Assault Burglary Larceny Theft Crimes Crime Crimes
Intercept 0.737 -0.64 -1.35 0.485 -0.167 0.679 -1.495 0.038 -0.436
Use of IT 0.032 0.037* 0.04 0.013 0.014 -0.013 0.043* 0.007 0.023
Log of #
-0.235 -0.136 -0.034 -0.145 -0.064 -0.082 -0.036 -0.077 -0.074
Officer
Median
0 0 0 0 0 0 0 0 0
Income
% Below
0.021 0.023* 0.023 0.012 0.013 0.003 0.023 0.012 0.016
Poverty
% White 0.005 0.007* 0.008* 0.004* 0.004 0.005 0.008* 0.005* 0.005
Prop. of
-0.118 -0.088 -0.171 -0.022 0.113 -0.059 -0.174 0.022 -0.018
Invest.
Crime Rate -0.186 0.031* 0.008 0.002 0.001 0.001 0.008 0.001 0.001

F 0.482 5.041** 1.628 1.747 1.446 1.430 1.977 1.389 1.259

DF Reg 7 7 7 7 7 7 7 7 7
DF Res 46 47 48 48 47 47 48 48 48
R -0.293 0.611 0.196 0.225 0.148 0.143 0.275 0.131 0.092
6ote. ***p< 0.001, **p< 0.01, *p< 0.05
223
 APPENDIX M: ITEMS WITH OVERLAPPING AGENCY SIZE

log of # of officer log of population median income

16

10
14

8
12

6
4
10
60000 100000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
3.0

100
80

30
2.0

60

20
40
1.0

10
20

0
0
0.0

Figure 1. Does your department use "Cellular / Mobile Phones" technology? No=9, Yes=224
224
 log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

4
10
60000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
3.0

100
2.5

80

30
2.0

60

20
1.5

40
1.0

10
20
0.5

0
0.0

Figure 2. Does your department use "RISS" for investigative purposes? No=114, Yes=119
225
 log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

10

4
60000

2
8

0
6
20000
% below poverty % white % proportion of investigators

40
100
3.0
2.5

80

30
2.0

60
1.5

20
40
1.0

10
20
0.5

0
0
0.0

Figure 3. Does your department use "NCIC" for investigative purposes? No=9, Yes=224
226
 APPENDIX N: ITEMS WITH NO OR POOR OVERLAPPING AGENCY SIZE

log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

10

4
60000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
100
3.0
2.5

80

30
2.0

60
1.5

20
40
1.0

10
20
0.5

0
0
0.0

Figure 1. Does your department use "Blackberry" technology? No=175, Yes=58

227
 log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

10

4
60000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
100
3.0
2.5

80

30
2.0

60
1.5

20
40
1.0

10
20
0.5

0
0
0.0

Figure 2. Does your department use "MDC / MDT" technology? No=65, Yes=168
228
 log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

10

4
60000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
100
3.0
2.5

80

30
2.0

60
1.5

20
40
1.0

10
20
0.5

0
0
0.0

Figure 3. Please identify the availability of "Fingerprints" for criminal investigations purposes. No=108, Yes=120
229
 log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

10

4
60000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
100
3.0
2.5

80

30
2.0

60
1.5

20
40
1.0

10
20
0.5

0
0
0.0

Figure 4. Please identify the availability of "Mug Shot" for criminal investigations purposes. No=36, Yes=192
230
 log of # of officer log of population median income

16

10
140000

14

8
12

6
100000

10

4
60000

2
8

0
6
20000

% below poverty % white % proportion of investigators

40
100
3.0
2.5

80

30
2.0

60
1.5

20
40
1.0

10
20
0.5

0
0
0.0

Figure 5. Please identify the availability of "Stolen Vehicles" for criminal investigations purposes. No=13, Yes=215
231
 232

APPENDIX O: REGRESSION COEFFICIENTS OF USE OF IT ITEMS

05 Murder 06 Murder
0.4

0.3
*
0.2

0.1

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.1

-0.2

-0.3
Figure 1. 2005-2006 murder clearance and use of IT items (***p< 0.001, **p< 0.01, *p<
0.05)

05 Rape 06 Rape
0.6
*
*
0.5

0.4

0.3
*
0.2 *
*
0.1

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.1

-0.2 *

Figure 2. 2005-2006 rape clearance and use of IT items(***p< 0.001, **p< 0.01, *p<
0.05)
 233

05 Robbery 06 Robbery
0.6

0.5 *

0.4
*
0.3
*
0.2
**
*
0.1 * * *
*
0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.1
Figure 3. 2005-2006 robbery clearance and use of IT items (***p< 0.001, **p< 0.01,
*p< 0.05)

05 Assault 06 Assault
0.3
*
0.25 *
0.2
0.15

0.1
*
0.05

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.05 *
*
-0.1 *
-0.15
Figure 4. 2005-2006 assault clearance and use of IT items(***p< 0.001, **p< 0.01, *p<
0.05)
 234

05 Burglary 06 Burglary
0.12
*
0.1

0.08 *

0.06

0.04
*
0.02
*
0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.02
-0.04

-0.06
Figure 5. 2005-2006 burglary clearance and use of IT items (***p< 0.001, **p< 0.01,
*p< 0.05)

05 Larceny 06 Larceny
0.1

0.08 *

0.06

0.04

0.02

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.02

-0.04

-0.06
Figure 6. 2005-2006 larceny clearance and use of IT items (***p< 0.001, **p< 0.01, *p<
0.05)
 235

05 Vehicle Theft 06 Vehicle Theft

0.2

0.15

0.1 *
*
*
0.05

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.05

-0.1
*
-0.15
*
-0.2
Figure 7. 2005-2006 vehicle theft clearance and use of IT items (***p< 0.001, **p< 0.01,
*p< 0.05)

05 Violent Crimes 06 Violent Crimes

0.3
*
0.25
*
0.2
*
0.15

0.1 *
**
*
0.05

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.05 *
-0.1

-0.15
Figure 8. 2005-2006 violent crimes clearance and use of IT items (***p< 0.001, **p<
0.01, *p< 0.05)
 236

05 Property Crimes 06 Property Crimes

0.08
*
0.06
*
* *
0.04

0.02

0
1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28
-0.02

-0.04
Figure 9. 2005-2006 property crimes clearance and use of IT items (***p< 0.001, **p<
0.01, *p< 0.05)

05 All Crimes 06 All Crimes

0.14
0.12
0.1
0.08
0.06
* *
0.04
0.02
0
-0.02 1 13 14 4 5 21 24 29 2 3 6 7 8 9 10 11 12 15 16 17 18 19 20 22 23 25 26 27 28

-0.04
-0.06
Figure 10. 2005-2006 all crimes clearance and use of IT items (***p< 0.001, **p< 0.01,
*p< 0.05)