SITY LIBRARI |$E

WW\IIHHWNW\IHIUWHl‘HHIM L

 

 

 

 

 

 

 

3 1293 00789
LIBRARY
Michigan State
University

 

This is to certify that the

dissertation entitled

The Outcomes and Processes
of Detective Decision Making
in Burglary and Robbery Investigations

presented by

Steven G. Brandl

has been accepted towards fulfillment
of the requirements for

Ph.D. mgmem Social Science

 

 

 

 

Major professor

nae July 1, 1991

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ififi—

MSU I. An Affirmative Action/Equal Opportunity lnflituticn
cWDIna-ox

 

 

 

 

 

7,,

THE OUTCOMES AND PROCESSES
OF DETECTIVE DECISION MAKING
IN BURGLARY AND ROBBERY INVESTIGATIONS

by

Steven G. Brandl

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

School of Criminal Justice
College of Social Science

1991

65¢-584l

ABSTRACT
THE OUTCOMES AND PROCESSES
0F DETECTIVE DECISION MAKING
IN BURGLARY AND ROBBERY INVESTIGATIONS
By

Steven G. Brandl

This study described the cognitive processes associated
with detective decision making and examined the influence of
case (victim and offense) characteristics on detectives’
decisions to (1) select a case for a follow-up investigation
and (2) allocate varying amounts of time to a follow-up
investigation. The data were gathered from a medium sized
Midwestern police department. Three methodologies were
used. First, case characteristics were coded from burglary
(N = 857) and robbery (N = 305) investigative reports and
the resulting data were used in OLS regression analyses to
determine the relationship between case characteristics and
decision outcomes. Second, an information board was used to
collect verbal protocol data from burglary and robbery
detectives (N = 10). These data provided insight into the
depth, content, and linearity of search. Third,
observations (370 hours) of detectives allowed for
additional insight into the factors which influence
decisions and the cognitive processes associated with

decision making. The OLS multiple regression analyses

showed that victim age, race, sex, income, employment
status, and identifiability of the stolen property did not
affect decision making. Dollar value of the stolen
property, strength of suspect information, and presence of
physical evidence did have a significant impact on decision
making. Victim type, victim desire for effort, victim-
offender relationship, presence of a suspect vehicle
description and license number displayed inconsistent
effects across decisions. Observations and verbal protocol
analyses showed interaction and dependency effects among
many of the variables and illustrated the extensive use of
linear decision making strategies by investigators. These.
findings are discussed in relation to their theoretical
contribution to the detective, police, and criminal justice

decision making literature.

To Kara
who has made countless sacrifices
and provides so much support

iii

ACKNOWLEDGMENTS

I am indebted to many people. I would like to thank the
administration of the Landau Police Department which allowed
me access to their organization and to Dr. Robert
Trojanowicz whose support was instrumental in gaining entry
to the department. Several faculty members -- Drs. Robert
Worden, Peter Manning, Micheal Lindell, Kevin Ford, and
Chairman Frank Horvath -- all helped shape and improve this
dissertation. My thanks also go to Jim Frank who assisted
me on this and many other projects over the past several
years. Finally, I would like to thank the subjects of this
study -- the detectives of the Landau Police Department.
They expanded my reality by letting me study theirs. For

this I am deeply grateful.

iv

TABLE OF CONTENTS
Page
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . ix
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . xi

CHAPTER ONE INTRODUCTION . . . . . . . . . . . . . . . 1

Problem . . . . . . . . . . . . . . . . . . . . . . . . 1
Purpose of the Study . . . . . . . . . . . . . . . . . 4
A Definitional Model of the Investigative Process . . . 5
Initial Discovery and Notification . . . . . . . . . 7
Initial Investigation . . . . . . . . . . . . . . . . 7
Follow-up Investigation . . . . . . . . . . . . . . . 8
Closure . . . . . . .1. . . . . . . . . . . . . . . . 8
Definition of Terms . . . . . . . . . . . . . . . . . . 9

Overview of Study . . . . . . . . . . . . . . . . . . . 11
Footnotes . . . . . . . . . . . . . . . . . . . . . . . 13
CHAPTER TWO LITERATURE REVIEW . . . . . . . . . . . . . 14
The Decision Tasks of Detectives . . . . . . . . . . . 14
Case Selection Decision . . . . . . . . . . . . . . . 15
Time Allocation Decision . . . . . . . . . . . . . . 19
Analytical Foundations for Research on Decision Making 23
Decisions as Outcomes . . . . . . . . . . . . . . . . 25
Victim Characteristics and Decision Making . . . . 26
Offense Characteristics and Decision Making . . . . 36

V

Decisions as Processes . . . . . . .
Summary . . . . . . . . . . . . . . . .
Footnotes . . . . . . . . . . . . . . .

CHAPTER THREE RESEARCH SITE . . . . . .
The City . . . . . . . . . . . . . . .
The Landau Police Department . . . . .
The Investigative Process at the Landau
Summary . . . . . . . . . . . . . . . .

CHAPTER FOUR METHODOLOGY . . . . . . .
Decisions as Outcomes . . . . . . . . .

Case Selection Procedure . . . . . .

Variables . . . . . . . . . . . . . .

Analysis . . . . . . . . . . . . . .

Hypotheses . . . . . . . . . . . . .

Decisions as Processes . . . . . . . .

The Information Board / Verbal Protocol

subjects 0 O O C O O O O O O O O O 0

Information Board Structure and Content

Procedure . . . . . . . . . . . . .
Research Questions . . . . . . . . .
Analysis . . . . . . . . . . . . . .

The Observation Method . . . . . . .
Footnotes . . . . . . . . . . . . . . .
CHAPTER FIVE RESULTS . . . . . . . . .

Decisions as Outcomes . . . . . . . . .

vi

Analysis

page
41
48
49
50
50
51
55
62
63
64
66
67
74
75
77
77
77
78
80
82
83
85
90
91
91

The Selection of Burglaries
Time Allocation in Burglary Investigations

Time Allocation in Robbery Investigations

Summary

Decisions as Processes .

The Selection of Burglaries .

Prioritization of Burglaries

Prioritization of Robberies

Summary

Observations

Footnotes

CHAPTER SIX DISCUSSION .

Decision Outcomes in Context
Case Characteristics and Decision

The Processes of Decision Making

Limitations of the Study

Directions for Future Research

Footnotes

APPENDIX A

APPENDIX B

APPENDIX C

APPENDIX D

APPENDIX E

APPENDIX F

Initial Investigation Report Form

Supplemental Report Form . . .

Modus Operandi Descriptor Form

Personal Descriptor Form . . .

Vehicle Descriptor Form . . . .

Property Form

vii

page
91
100
109
119
120
120
127
136
142
143
154
155
155
158
174
178
183
186

187
189
190
191
192
193

page
APPENDIX G - Follow-up Investigation Case Log . . . . . 194
APPENDIX H - Case Data Coding Form . . . . . . . . . . . 195
APPENDIX I - Information Board: Selection of Burglaries 196

APPENDIX J - Information Board: Prioritization of
Burglaries O O O O O O O O O O O O O O O 198

APPENDIX K - Information Board: Prioritization of

Robberies . . . . . . . . . . . . . . . . 200
APPENDIX L - Information Board Instructions for the

Selection of Burglaries . . . . . . . . . 202
APPENDIX M - Depth of Search Formula and Examples . . . 203
APPENDIX N - Content of Search Formula and Examples . . 204
APPENDIX 0 - Linearity of Search Formula and Examples . 205
LIST OF REFERENCES . . . . . . . . . . . . . . . . . . . 207

viii

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

10

11

12

LIST OF TABLES

Independent and Dependent Variables: Values
and Descriptive Statistics by all
Burglaries . . . . . . . . . . . . . . . . .

Correlation Coefficients Among Variables
for all Burglaries . . . . . . . . . . . . .

Probit (and Multiple Regression) Analysis of
Burglary Case Selection as a Function of
Victim and Offense Characteristics . . . . .

Independent and Dependent Variables: Values
and Descriptive Statistics by "Selected"
Burglaries . . . . . . . . . . . . . . . . .

Frequency of Detective Activities in Burglary
Follow-up Investigations . . . . . . . . . .

Correlation Coefficients Among Variables
for "Selected" Burglaries . . . . . . . . . .

Multiple Regression of Time Spent on Burglary
Follow-up Investigations as a Function of
Victim and Offense Characteristics . . . . .

Independent and Dependent Variables: Values
and Descriptive Statistics by "Selected"
Rabberies C O O O O O O O O O O O O C O C O 0

Frequency of Detective Activities in Robbery
F0110w-UP Investigations 0 o o o c e c e o 0

Correlation Coefficients Among Variables
for "Selected" Robberies . . . . . . . . . .

Multiple Regression of Time Spent on Robbery
Follow-up Investigations as a Function of
Victim and Offense Characteristics . . . . .

Proportion of Available Information Searched
in the Burglary Selection Decision . . . . .

ix

Page

92

95

97

101

103

105

107

110

113

114

117

121

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

13

14

15

16

17

18

19

20

21

Page

The Importance of Case Information in the
Burglary Selection Decision . . . . . . . . . 122

Linearity of Search in the Burglary
Selection Decision . . . . . . . . . . . . . 125

Proportion of Available Information Searched
in the Prioritization of Burglary Cases . . . 128

The Importance of Case Information in the
Prioritization of Burglary Cases . . . . . . 130

Linearity of Search in the Prioritization
of Burglary Cases . . . . . . . . . . . . . . 132

Proportion of Available Information Searched
in the Prioritization of Robbery Cases . . . 137

The Importance of Case Information in the
Prioritization of Robbery Cases . . . . . . . 138

Linearity of Search in the Prioritization
of Robbery Cases . . . . . . . . . . . . . . 140

Decision Outcomes for Burglary and Robbery
Cases in Eck (1983) and Present Study . . . . 157

LIST OF FIGURES

Page

FIGURE 1 Landau Police Department Organizational
Chart 0 O O O O O I O O O O O O O O O O O O O 52

FIGURE 2 A Decision Making Model of Burglary and
Robbery Investigations . . . . . . . . . . . 144

xi

CHAPTER ONE

INTRODUCTION

Chapter One contains an introduction to the research.
The problem and purpose of the study are discussed and the
definitions used in the study are presented. Chapter One

concludes with an overview of the dissertation.

Problem

The criminal justice system, that mechanism of society
created to deal with crime and criminals, can be
conceptualized as a sequential series of decision stages.
Research attention has been directed at examining the
critical decisions of participants within each of the
stages. For example, the victim’s decision to report a
crime has been analyzed (Hindelang & Gottfredson, 1976;
Laub, 1981), along with the police decision to arrest
(Black, 1971; Smith & Visher, 1981; Visher, 1983; Worden &
Pollitz, 1984; Smith, 1987) and investigate (Bynum, Cordner,

& Greene, 1982), judicial decision to grant pretrial release

(Frazier, Bock, & Henretta, 1980; Nagel, 1983), prosecutor
decision to charge (Adams & Cutshall, 1987; Albonetti, 1986;
Schmidt & Steury, 1989), and plea bargain (Holmes,
Duadistel, & Farrell, 1987), juridic decision to convict
(Brooks a Doob, 1975), judicial decision to sentence
(Baldus, Pulanski, & Woodworth, 1983; Platt-Jendrek, 1984;
Welch & Spohn, 1986), and parole board decision to grant
release (Von Hirsch & Hanrahan, 1979).

While most of these stages and participants have been
the objects of extensive research attention, little research
has focused specifically on investigative decisions by
detectives. In fact, only one study in the literature has
taken this as its primary focus (Bynum, Cordner, & Greene,
1982). Other studies with a broader focus on the
investigative process (e.g., Eck, 1983; Greenwood, Chaiken &
Petersilia, 1977; Sanders, 1977) have contributed only
indirectly to our understanding of detective decision
making.

There appear to be at least two reasons for the lack of
research on this topic. First, unlike many other decision
stages in the justice process (e.g., arrest, plea bargain,
convict, sentence) the decision to investigate is
characterized by a relatively low degree of visibility
(Ericson, 1981). The decision occurs ”backstage" and
therefore, is not often open to public scrutiny. As a

result, the topic may be simply overlooked by researchers.

Second, detective decision making has been widely portrayed
as being "routine" (Eck, 1983; Greenwood et al., 1977;
Sanders, 1977) where the strength of the evidence is assumed
to automatically determine the disposition of the case.
Given this widely shared (but untested) reality, few
researchers have deemed this topic as particularly worthy of
research attention.

The lack of research on detective decision making is
troubling. A substantial amount of police resources and
activities are allocated to follow-up investigations yet we
know very little about this decision stage. From the
perspective of developing an adequate understanding of
police decision making in total, this is inherently
unsettling. Relatedly, students of investigative management
have long called for strategies to increase the capacity of
the police to apprehend offenders. However, a prerequisite
for improving the effectiveness of the criminal
investigation process is a sound understanding of the
process. As Ericson (1981) explains, ”perhaps most of the
proposals for reform have little impact because reformers
know too little about what it is they are trying to reform"
(p. x). One dimension of a more complete understanding of
the investigative process is the identification of the
premises and cognitive processes associated with detective

decision making.

Purpose of the Study

The purpose of this study is to analyze detective
decision making. To do so, two broad research questions are
addressed. First, what case (victim and offense)
characteristics influence the decisions to (1) select a case
for a follow-up investigation and (2) allocate varying
amounts of time to a follow-up investigation? As research
has illustrated, the selection of a case for a follow-up
investigation does not necessarily mean that attention is
given to the case (Bynum et al., 1982; Greenwood et al.,
1977; Sanders, 1977). Thus, to provide a thorough inquiry,
both decisions within the follow-up investigation are
examined.

Second, how do detectives treat case information in
making decisions? Whereas the first question is most
concerned with specifying the relationship between the input
(information) and the outcome of the decision process (the
decision), the second question is concerned primarily with
describing the cognitive processes involved in decision
making. As such, the two questions emerge from different
theoretical perspectives on decision making and require the
use of different methodological approaches in order to be
addressed. By studying the decision behavior of detectives
through the "outcome" and "process" perspectives, it is
possible to attain a better understanding of investigative
decision making and ultimately, the complexities of the

criminal investigation process.

4

1
A Definitional Model of the Investigative Process

The municipal police organization provides three valued
outputs -- service, order maintenance, and law enforcement
(i.e. ”crime control") (Wilson, 1968). These outputs also
comprise the major categories of work activities within the
police organization. "Service" refers to the provision of
assistance to the public in regard to non-crime related
matters. "Order maintenance" involves activities oriented
around maintaining the public peace. "Crime control"
activities involve intervening in situations where a law has
been violated and the identity of the perpetrator needs to
be determined. Conceptually, the criminal investigation
process can be placed within the crime control aspect of the
police mission.

”reactive"

Typically, criminal investigations are of a
nature, where the police respond to the report of a criminal
offense. Some investigations however, especially those
associated with vice offenses, are proactive or police
initiated (see Manning, 1980; Wilson, 1978). The focus of
this study is on the more typical "reactive" type
investigation.

At the simplest level, the criminal investigation
process involves activities oriented around the collection
of crime related information in order to: (1) determine if a

crime has been committed; (2) identify the perpetrator(s);

(3) apprehend the perpetrator(s); and (4) provide evidence

to support a conviction in court (Eck, 1983; Greenwood et

al., 1977; Kuykendall, 1982). With the arrest rate as the
primary measure of investigative effectiveness, arresting

offenders (attaining the second and third objective above)
has been most often portrayed as the overriding concern of
investigators (Greenwood et al., 1977; Waegel, 1981).

According to Willmer (1970), the criminal investigation
process can be viewed as a battle between the police and the
perpetrator over crime related information. That is, the
perpetrator, in committing a crime, emits signals
(information) which the police attempt to collect through
investigative activities. If the perpetrator is able to
minimize the amount of information available for the police
to collect, or if the police are unable to recognize the
information left behind, then the perpetrator will not be
apprehended and therefore, the perpetrator wins the battle.
Conversely, if the police are able to collect a significant
amount of signals from the perpetrator, then the perpetrator
will be apprehended and the police win.

For definitional purposes, the (reactive) criminal
investigation process can be organized into several stages:
initial discovery and response, preliminary investigation,
follow-up investigation, and closure. Each of these stages

is discussed below.

Initial Discovery and Notification

In order for the criminal investigation process to be
invoked, the police must discover that a crime has taken
place and then notify the victim, or the victim (or witness)
must discover that a crime has occurred and notify the
police. In the vast majority of cases it is the victim who
first discovers that a crime occurred and who contacts the
police (Greenwood et al., 1977). Then, in most cases, a

patrol officer is dispatched to the crime scene.

Initial Investigation

If, upon arrival, the officer actually defines the
matter as a crime (see Black, 1971), then an initial (or
preliminary) investigation is conducted. The initial
investigation consists of the immediate post-crime
activities of the patrol officer who arrives at the crime
scene. The officer may proceed to gather information
("signals”) concerning the crime by questioning the victim
and/or witness(es), searching the crime scene, etc. The
specific activities engaged in by the officer may be a
function of the particular case at hand. All of the
information relating to the crime would then be recorded in

an initial investigation report.

Follow-up Investigation

If a perpetrator is not arrested during the initial
investigation, the case may be selected for a follow-up
investigation -- the second stage whereby ”signals” may be
collected. Typically, detective supervisors take the
initial investigation reports from the case pool which
appear relevant to their unit (e.g., ”Homicide," "Crimes
Against Persons,” etc.) and then decide which of the cases
should receive a follow-up investigation. If a case is
selected for a follow-up, then the detective assigned to the
case must decide what activities to perform in the
investigation. Depending on the particular case, the
follow-up investigation may involve a variety of activities
ranging from recontacting and re-interviewing the victim, to
submitting evidence to the crime laboratory, to seeking out
informants (Eck, 1983). The information which is cultivated
as a result of these activities would be recorded in a
follow-up report. It is the complexities of the follow-up
investigation as well as the case transition process, from
the initial investigation to the follow-up investigation,

that is of direct interest in this study.

Closure
At any time during the investigative process the case
may be closed and investigative activities terminated. For

instance, the case could be closed due to a lack of leads or

as a result of the offender being apprehended. In the
latter situation, the crime would be considered ”cleared by
an arrest” and primary responsibility for the case would
shift from the police department to the prosecutor’s
office. However, the detective(s) assigned to the case
would still have the responsibility of assisting the

prosecutor in preparing the case for prosecution.

Definition of Terms

The following are definitions of terms used in this
study.

Crime: A crime is the commission of an act prohibited by
criminal law or the failure to act as required by criminal
law for which punishment is prescribed (Reid, 1989).
Specifically, the present study focuses on the crimes of
burglary and robbery. As defined by the F.B.I. Uniform
Crime Report:

”Burglary" refers to the unlawful entry of a structure

to commit a felony or theft. The use of force to gain

entry is not required to classify an offense as a

burglary. (For purposes of this study attempted

burglary is not included.)

”Robbery" refers to taking (or the attempt to take)

anything of value from the care, custody, or control of

a person or persons by force or threat of force or

violence and/or by putting the victim in fear.

Detective: A detective holds a specialized position
within the police organization being concerned primarily

with the "law enforcement" function of the police mission.

Typically, a detective becomes involved in a criminal

investigation only after the initial investigation has been
completed by a patrol officer. Normally, a detective has
the sole responsibility of conducting a follow-up
investigation.

Detective Sergeant: A detective sergeant is the first
line supervisor of detectives within an investigative unit.
Detective sergeants have the primary task of deciding which
initial investigation reports to assign to detectives for
follow-up investigations.

Follow-up investigation report: Follow-up reports are
produced by detectives and identify the information
cultivated as a result of follow-up investigation
activities.

Investigator: For purposes of this study, an
investigator refers to either a detective or a detective
sergeant.

Official complainant records: Official complainant
records are reports completed by patrol officers which
detail the nature of the police-citizen contact. In
criminal incidents, these reports contain information on the
victim and the offense which is obtained through the initial
investigation activities. These reports are also known as
initial investigation reports. Complainant records are
maintained within the police department.

Patrol Officer: A patrol officer has broad and diverse
responsibilities within the municipal police organization.
A patrol officer is concerned with the order maintenance,

10

service, and law enforcement functions of the police
mission. In the case of a criminal incident, a patrol
officer typically responds to the scene of the crime (and/or
the location of the victim) and is responsible for
conducting the initial investigation.

Personal crime: A personal crime involves the victim
directly -- the crime is an attack on the individual. If
the crime is directed toward an individual who is a
representative of a business establishment, then the crime
would be considered a crime against a business. The
personal crime of interest in this study is robbery.

Property crime: A property crime is directed toward a
victim’s property and hence, is an indirect attack on the
individual. Again, the property of a business establishment
may be the focus of the crime and, in such an instance, the
crime would be considered a crime against a business.
Burglary is the property crime of interest in this study.

Victim: For purposes of this study, a victim is an
individual (either a representative of a business or not)
that is the object of a criminal act (burglary or robbery)
and suffers injuries and/or material losses as a result of

the act.

Overview of the Study
In Chapter Two, the decision tasks of detectives are
discussed, the analytical foundations for research on
decision making are outlined, and previous research is

11

reviewed. In Chapter Three the research site is described.
In Chapter Four, the methodologies used in this study are
outlined. The results of the study are presented in Chapter

Five. Chapter Six contains the discussion and conclusions.

12

Footnotes

This discussion represents a general and generic
definitional overview of the criminal investigation
process. The mechanics of the investigative process, as
found in the present study site, are discussed in Chapter
Three.

13

CHAPTER TWO

LITERATURE REVIEW

Chapter Two begins by discussing the decision tasks of
detectives. Attention then turns to the analytical
approaches used in studying decision making. Studies which
have adopted these approaches in analyzing decisions are
reviewed. Through the review, the propositions and research

questions addressed in the study are developed.

The Decision Tasks of Detectives

Myriad studies have highlighted the discretionary nature
of police work. Discretion, in this context, refers to
”autonomy of decision making” (Black, 1968, p. 25). As
stated by Davis (1969), "a public officer has discretion
whenever the effective limits on his power leave him free to
make a choice among possible courses of action or inaction"
(p. 4). Simply put, discretion exists when one has the
freedom and authority to make decisions. A decision, at the

most basic level, is a choice among alternatives based on

14

information and guided by preferences.

Just as the criminal justice system can be
conceptualized as a sequential series of discretionary
decisions, so too can the criminal investigation process.
While a model of the criminal investigation process was
presented earlier (Chapter One), the purpose of the
following discussion is to describe the two critical
decision tasks of investigators which correspond to the two
stages of the follow-up investigation process: the case

selection decision and the "time allocation" decision.

Case Selection Decision

Case selection, or screening, the first stage of the
follow-up investigation process, typically involves a
detective sergeant deciding whether or not the initial
investigation report should be assigned to a detective for a
follow-up investigation. Depending on the department (or
the decision maker) the case screening decision may reflect
an aided-analytic strategy, an unaided-analytic strategy, or
a non-analytic strategy (Beach and Mitchell, 1978). It is
also possible that in some departments, the screening
decision does not technically exist.

According to Beach and Mitchell (1978), an
aided-analytic strategy requires the decision maker ”to
apply a prescribed procedure utilizing tools such as pencil

and paper, mathematics, calculator, or computer, etc. in a

15

guided, systematic attempt to analyze the decision and
evaluate its components" (p. 441). Specifically, if a
"screening decision model” is used, the decision may
resemble the aided analytic strategy. As defined by Eck
(1983), the use of a screening decision model...
involves simply making a decision to assign or not to
assign investigative resources to cases by applying a
fixed set of criteria to information contained in
preliminary investigative reports (p. 274).
Gaines, Lewis, and Swanagin (1983) add...
A case screening process identifies those cases which
have the potential for being solved and allows
investigators to spend more time on them by eliminating
from officers' caseloads cases which probably cannot be
solved due to absence of substantive evidence (p. 22).
If an investigative unit used a case screening model,
each initial investigation report would be examined in light
of the case screening assessment criteria (”solvability
factors") and then the utility of a follow-up investigation
would be mathematically determined. For example, in the
decision model presented by Eck (1983), various information
elements (e.g., presence of suspect identification,
fingerprints, etc.) are combined in a weighted sum and those
cases with a score higher than a certain predetermined score
are selected for a follow-up investigation.
While the advantages of a case screening system have
been noted (Hastings, 1980), there is often much resistance
given to the formal use of this type of device in

investigative decision making (Eck, 1983). As a result,

some departments have instituted policies which identify

16

certain solvability factors to sensitize decision makers to
the information on which the screening decision should be
based; but weights are not assigned to these elements. As
such, the decision to select a case for a follow-up
investigation often resembles an "unaided-analytic
strategy." With an unaided-analytic strategy, "an attempt
is made to explore the dimensions of the problem but... no
tools are used, and the decision maker restricts processing
to the confines of his or her mind” (Beach & Mitchell, 1978,
p. 441). They continue, "unaided analytic strategies have
the advantage of reducing information processing by
restricting attention to only part of the available
information about the alternatives, but they have the
disadvantage of introducing possible irrationalities" (p.
442).

The screening decision may also reflect a mpg-analytic
strategy where "little information is procured or processed,
little time is needed, and the rules do not require that the
decision be decomposed nor that its multiple aspects be
considered" (p. 442). Examples include flipping a coin or
such conventions as "eeny, meeny, miney, mo..." Decisions
made by habit, an "extreme example of rote application of a

" (p. 442) are also non-analytic in nature.

rule,

When the screening decision does not technically exist
in a department, all initial investigation reports are given
to the detectives and the detectives determine not

necessarily which ones to select for an investigation, but

17

which cases should receive the most attention. This is a
subtle distinction in practice but one important for
analytical purposes.

The literature on detective decision making has tended
to consider the follow-up investigation process as a whole
and as a result, the selection decision as a distinct stage
in the process has not received much attention. When the
screening decision does receive comment, it is usually only
in passing. For example, Sanders (1977) notes, "the
sergeant would give one of us researchers the batch of
reports to go over and determine which ones would be worked
and which ones would not. The selection process was so
routine that we rarely made mistakes" (p. 77). No other
discussions in the text are devoted to the selection
decision. Even Bynum et a1. (1982), in an empirical study
which focused specifically on detective decision making, did
not discuss the screening decision.

There exists at least two reasons for this lack of
attention. First, the structure and organization of the
investigative process in the departments previously studied
may not have provided for a ”screening decision." Or,
second, the decision may have been simply ignored,
overshadowed by the other decision stage within the
follow-up investigation -- the "decision" as to how much
time to devote to an investigation. It is to this stage of

the investigative process that attention now turns.

18

Time Allocation Decision

To be accurate, the decision concerning how much time to
spend on a given follow-up investigation is not really a
single decision, as is the screening decision, but rather a
series of interrelated decisions. That is, upon receiving
and reviewing an initial investigation report, a detective
does not decide how much time to spend on the
investigation. Rather, after receiving the report, the
detective may (or may not) decide to contact the victim,
then he may (or may not) decide to query departmental
records, then he may (or may not) decide to interview
witnesses, etc. It is only after the detective decides to
discontinue any further investigative activities that one
can identify the total amount of time the detective spent on
the investigation. With a slightly broader perspective, one
form of decision making at this stage is case prioritization
(Eck, 1979; Ericson, 1981). Detectives are likely to be
more willing to spend time on certain cases than on others.
Conceptually, case prioritization represents the aggregate
outcome of many decisions, decisions which are likely to be
at least partially based on victim and offense
characteristics.

The selection decision is relatively well defined and
structured with much (but perhaps not all) of the
information considered in making the decision likely to be
contained within the initial investigation report (Eck,
1983). On the other hand, decisions concerning the amount

19

of time to devote to a case may be based on not only
information contained within the initial investigation
report but also on other information gathered directly by
the detective through investigative activities. While the
search for information in a screening decision could be
completed within a matter of minutes, the search and
collection of information in the ”time allocation decision"
could take place over a period of hours, days, or even
weeks, thus increasing the complexity of this stage of the
follow-up investigation.

Previous research on detective decision making has
focused most directly on this decision stage. In the
previous studies, the time spent on a follow-up
investigation, or investigative effort, has been measured in
several different ways. For example, Bynum et a1. (1982),
in an analysis of detective decision making in a "medium
sized midwestern police department,” examined the extent to
which follow-up investigations were conducted on a sample of
1,124 personal and property crimes reported during a five
week time period in 1978 and 1979. The ”extent of follow-up
investigation" variable had the values of (1) reviewing
report only; (2) making a few phone calls; and (3)
conducting a more extensive investigation including
examining the crime scene, searching for additional
witnesses, interviewing suspects, etc. Data on the variable
were obtained from a review of follow-up investigation
reports completed by detectives. In describing

20

investigative effort, the researchers found that "82 percent
of all cases that come to detective attention receive little
or no investigative effort" (i.e., fall into categories one
and two; p. 315).

Eck (1983), in his analysis of preliminary and follow-up
investigations and their relative impact on the solution of
burglaries and robberies, collected data on 320 robberies
and 3,360 burglaries from three police departments -- DeKalb
County (GA), St. Petersburg (FL), and Wichita (KS). These
departments ranged in size from 374 officers to 445
officers. Data on the time spent by detectives on case
investigations came from "activity-time logs" completed by
detectives for every case worked. Three measures of
investigative time were used: (1) the number of days the
case remained open; (2) the number of days on which the case
was worked; and (3) the number of minutes actually spent
investigating a case. In regard to the last measure of
time, Eck found that a mean of 167 minutes (2.8 hours) were
spent on robbery follow-up investigations across all study
sites and 77 minutes (1.3 hours) were spent on burglary
investigations.

In Ericson’s (1981) qualitative (observation) study of
detective work in ”a jurisdiction in the Canadian province
of Ontario" (p. 24), it was found that approximately 30
percent of all cases that came to the attention of the
detective bureau received one or more hours of investigative
time. In this study, time spent on investigations was

21

estimated on the basis of detailed field observations.

According to the study by Greenwood et al. (1977), less
than half of all crimes reported to the Kansas City (MO)
Police Department "received serious consideration by an
investigator” (p. 109). Specifically, based on their review
of the department’s "computer-readable case assignment
file,” it was found that 63 percent of robberies, 36 percent
of non-residential burglaries, and 30 percent of residential
burglaries received ”at least half an hour of a detective’s
time" (p. 130).

In regard to how investigative time was actually spent,
Ericson (1981) noted that detectives interviewed one or more
victims, complainants, and/or informants in 31.5 percent of
the cases and suspects in 27.5 percent of the cases. Eck
(1983) found that as investigations progressed, they became
less routine. The activities performed later in the
investigation were more uncommon than those performed
earlier in the investigation. For example, the frequency
with which victim interviews and crime scene checks were
conducted declined over each investigative day while the
frequency of suspect interviews increased.

With a basic understanding of the decision stages within
the follow-up investigation process, attention turns to a
discussion of the approaches used to study decision making

and a review of the related literature.

22

Analytical Foundations for Research on Decision Making

Studies of human decision making have taken one of two
analytic approaches (Hogarth, 1974; Payne, 1976). In the
first approach, referred to here as ”outcome oriented,” the
focus is on specifying the relationship between information
stimuli and the decision outcome. Typically, information
”input" is represented in terms of cues to which the
decision maker responds. Through the use of statistical
procedures (e.g. linear regression), the strength of the
stimulus-response relationships can be measured and decision
choices can be predicted. While implicit assumptions about
cognitive processes are made in such models, the actual
nature of the processes remain hidden within an inaccessible
”black box" (Hogarth, 1974).

The "process oriented” approach, the other stance
adopted by decision making research, attempts to cast light
into the ”black box” of outcome oriented studies by
examining the thought processes involved in decision
making. Accordingly, the intent is to describe how
decisions are made (Abelson & Levi, 1985). Studies which
have attempted to illuminate (or "trace") the processes of
decision making have employed several methods. One of the
more formal and rigorous methods involves the collection and
analysis of decision maker’s verbal protocols. Usually with
the aid of an information board (see Payne, 1976), decision
makers are asked to ”think aloud” while performing a
decision task. The resulting verbal statements are

23

recorded, broken into short task assessment phrases, and
then content analyzed for evidence of different decision
strategies (Payne & Ragsdale, 1978; Payne, 1976). Through
an analysis of the verbal data, the processes by which
information input is transformed into decision outcomes can
be described.

Observationally based studies can also be used to gain
insight into the cognitive processes of decision makers. In
such studies, observers ask decision makers, during their
normal course of work activities, to describe the cognitive
processes which were involved in resolving a given decision
situation (Mastrofski & Parks, 1990). While this method of
collecting process data is often viewed as being less
rigorous than verbal protocol analysis, it offers the
potential advantage of keeping the study of decision making
it its natural environment. It also appears most
appropriate when the decision tasks are not naturally well
defined or structured. Regardless of the approach however,
both ”methods” can provide at least preliminary insight into
how decision makers go about making decisions.

Decision making studies which have adopted the "outcome"
and "process" approaches are discussed below. The research
highlighted in the outcome oriented section focuses
specifically on decision making within the criminal justice
system. The review of process oriented studies focuses on
the psychological literature of process tracing and
highlights the contributions a process oriented approach can

24

make to the more traditional outcome oriented inquiry. In
each of the sections, the general research propositions and

questions which are addressed in this study are developed.

Decisions as Outcomes
Myriad studies have attempted to identify the case
characteristics (stimuli) which influence the decision
responses of participants within the justice system. In
defining case characteristics, one can distinguish between
”legal" and "extra-legal” factors. Legal factors include
characteristics of the offense such as the amount and type
of evidence and the amount of harm done (Nagel, 1983).
Extra-legal factors include most often characteristics of
the victim and/or offender such as sex, age, race, etc.
Since an ideal in the administration of justice is the
fair and equal application of the law, it is considered
irrational to base decisions on "irrelevant" extra-legal
considerations. As explained by Karmen (1984)...
The Fourteenth Amendment to the Constitution promises
”equal protection of the law” for all citizens. The
standard interpretation of this pledge is that [the
criminal justice system] ought to regard factors like
social class, race, nationality, religion, and sex as
irrelevant to the administration of the law (p. 164).
Because of the complexity and uncertainty inherent in many
decision making tasks within the criminal justice system
however, it is difficult to structure or ”program" decision

situations with rules, guidelines, and instructions (Lipsky,

1980; Thompson, 1967). Therefore, since there is often not

25

any method by which extra-legal factors can be absolutely
ruled out of decision making, research to understand the
stimuli which affect decision making must examine the

relative importance of both sets of variables.

Victim characteristics and decision making

A great deal of previous criminal justice decision
making research has demonstrated how characteristics of
effendege affect decisions of legal actors (e.g.
Platt-Jendrek, 1984; Nagel, 1983; Spohn, Gruhl, & Welch,
1987). However, at the criminal investigation stage, the
offender is often unknown to the police and the victim (U.S.
Department of Justice, 1988). Therefore, it is more
appropriate to examine how the characteristics of gigging
influence decisions.

Black’s (1976) theory of the behavior of law provides
the foundation for much of the research which has attempted
to predict the relationship between victim characteristics
and criminal justice decision making. Black presents a
series of propositions which attempt to predict the way law
behaves, or in specific reference to criminal law, the
zeepeneee of decision makers within the criminal justice
system (e.g., victims, police, judges, etc.).1 Law, as
defined by Black, is "governmental social control" which can
vary quantitatively (as well as qualitatively)2 across
time, space, and individuals. By the quantity of law, Black
refers to the extent to which legal action is initiated,

26

invoked, or applied. For example, an arrest by the police
is more law than no arrest, a long prison sentence is more
law than a short prison sentence and, in specific reference
to this study, an investigation is more law than no
investigation.

According to Black, the amount of law invoked as a
result of a particular incident varies with the perceived
seriousness of the incident; with more seriousness
corresponding to ”more law." However, unlike other
conceptions of seriousness where seriousness is attributable
to the objective nature of the act itself (e.g., amount of
harm done), seriousness is viewed by Black as a function of
the contextual (or social structural) factors of the victim
and/or offender. For example, if two homicides occurred and
only the characteristics of the victims varied (i.e.,
offender characteristics were unknown or controlled for),
then the perceived seriousness of the incidents (and the
amount of law invoked) would vary in the way predicted by
the theory. The social structural variables of importance
in Black’s theory are: stratification, morphology, culture,
organization, and social control.

In the discussion which follows, the propositions
associated with these variables, as derived by Black, are
presented. Previous criminal justice, police (patrol
officer), and detective decision making research which has
examined these propositions is then reviewed.

Stgetifieejjen. Stratification, as discussed by Black,

27

is "the vertical aspect of social life... the uneven
distribution of material conditions of existence” or
”inequality of wealth" (p. 11). According to Black, one’s
wealth is equatable to one’s position or rank in society.
Holding the offender’s rank constant, ”law varies directly
with the rank of the victim" (p. 26). Because crimes
against the wealthy are viewed as more serious than crimes
against the poor, according to Black, ”the lower ranks have
(or get) less law than the higher ranks” (p. 17). Black
adds that ”it is even possible to rank entire
neighborhoods. This may be done either according to the
distribution of wealth among residents or according to the
wealth of the... area itself” (p. 20). Black explicitly
states, "the wealthier the victim of a crime, the more
likely is an investigation by the police" (p. 27).
Accordingly:

Victim income influences the amount of law invoked as a
result of a criminal incident.

Specifically, wealthy crime victims get more law than

poor victims.

In addition, according to Black, sex is associated with
one’s rank (p. 17) -- with females receiving less law than
males. Therefore:

Victim sex influences the amount of law invoked as a
result of a criminal incident.

Specifically, male crime victims get more law than
female crime victims.
In examining the relationship between victim income and

28

decision making, Smith and Klein (1984) found that ”the
police respond differently to interpersonal disputes
depending on the socioeconomic status of the neighborhood in
which it occurs" (p. 475). Specifically, the police were
more likely to make arrests in low status areas (which were
presumably populated with lower status victims). Smith
(1987) reported that "economic status of the neighborhood,"
measured by the percent of households with an annual income
below $5,000, had a significant impact on police decision
making in violent disputes with an arrest more likely to
occur in lower status neighborhoods. Black (1970) however,
found that the police were less likely to file a felony
complaint report when the complainant was a "blue collar”
individual as opposed to a ”white collar" individual.

In reference to the affect of victim income on detective
decision making, Ericson (1981) found that "cases with high
status or otherwise special victims were sometimes given
immediate priority because of orders that could ultimately
be traceable to the upper echelons of the police
organization" (p. 79). Waegel (1981) explained that in
burglary cases the victim’s class position had a ”decisive
impact” on the amount of attention given to the case.
According to Bynum et al. (1982), burglary cases with
victims who lived in census tracts with higher median
incomes were more likely than victims who lived in tracts
with lower incomes to receive extensive investigative
attention. This variable was not found to affect decision

29

making in personal offenses however.

Concerning the influence of victim gender on decision
making, Williams (1978), in a comprehensive examination of
the role of the victim in the prosecution of violent
offenses, found that cases with female victims were more
likely to be rejected by the prosecutor. Similarly, Smith
and Klein (1984) and Smith (1987) found that the police were
significantly less likely to arrest in situations involving
female complainants. With detective decision making
however, Bynum et al. (1982) found that the victim’s gender
did not have an effect on decision making in either property
or personal offenses.

Megphelegy. Morphology is defined as ”the horizontal
aspect of social life, the distribution of people in
relation to one another, including their division of labor,
networks of interaction, intimacy, and integration” (p.

37). Individuals "may participate more or less in social
life itself. Some participate fully and usefully; others
stay at the margin, hardly involved at all" (p. 48). ”Some
people work; others idle or loiter” (p. 48). According to
Black, employment status serves as an indicator of
integration. Hold constant the offender, "and law varies
directly with the integration of the victim” (p. 53). Black
claims that crimes against integrated (employed) victims are
viewed as more serious than crimes against non-integrated
(not employed) victims, and as a result, those who are
non-integrated get less law than those who are integrated.

30

”The closer to the center he is... the more extensive is an
investigation of his problem" (p. 53). Accordingly:

Victim employment status influences the amount of law
invoked as a result of a criminal incident.

Specifically, victims who are employed get more law than
victims who are not employed.

In addition, "people vary in the degree to which they
participate in one another’s lives" (p. 40). ”It is
possible to measure relational distance in many ways,
including the scope, frequency, and length of interaction
between people, the age of their relationship, and the
nature and number of links between them in a social network”
(p. 41). According to Black, ”law is inactive among
intimates, increasing as the distance between people
increases" (p. 41). Therefore:

The victim-suspect relationship influences the amount of
law invoked as a result of a criminal incident.

Specifically, victims who are not acquainted with the
suspect get more law than victims who are acquainted
with the suspect.

Previous research, to one degree or another, has
addressed both of these predictions. Only one study in the
literature, Bynum et al. (1982), examined the impact of
victim employment status on decision making. They found
that employment status did not affect the amount of effort
devoted to follow-up investigations.

The victim-suspect relationship has been found to

influence decision making. According to Schmidt and Steury

(1989), domestic assault cases in which victims shared a

31

dwelling with the offender and were sexually intimate with
the offender prior to the assault were less likely to be
continued. Similarly, Albonetti (1986) discovered that
victims who were acquainted with the suspect were more
likely to have their cases discontinued than were victims
who were strangers to the suspect.

According to Black (1971), the police were more likely
to arrest an offender when he was a stranger to the victim
as Opposed to when the offender and the victim were friends,
acquaintances, or relatives (also see Smith & Visher, 1981;
LaFave, 1965; Friedrich, 1977). Similarly, Worden and
Pollitz (1984) found that in domestic disturbance
situations, the offender was more likely to be arrested if
the disputing parties were not married. No studies have
examined the impact of the victim-offender relationship on
detective decision making.

Culture. Black defines culture as ”the symbolic aspect
of social life including expressions of what is true, good,
and beautiful" (p. 61). Because of variance in ideas,
information, languages, and customs, some societies and
”individuals have more culture than do others" (p. 64).
While education is presented as the primary indicator of
culture, Black also suggests that certain groups in a
society are closer to the mainstream of culture or are more
"conventional." Holding the offender’s characteristics
constant, "law varies directly with the conventionality of
the victim" (p. 70). For example, a crime against a

32

cultural minority is claimed to be viewed as less serious
than a crime committed against a cultural majority and
consequently, it is predicted that cultural minorities
receive less law than cultural majorities. Therefore:

Victim race influences the amount of law invoked as a
result of a criminal incident.

Specifically, white crime victims get more law than
non-white victims.

Concerning the influence of racial identification, Smith
(1987) found that the police were significantly less likely
to make an arrest when the situation involved a black
complainant. However, Smith and Klein (1984) did not find
such a relationship. As for detective decision making,
Waegel’s (1981) observations led him to believe that the
victim’s race had an impact on the amount of attention given
to the case. However, Bynum et a1. (1982) found that victim
race did not affect the extent of effort spent in a
follow-up investigation.

ngegizeeien. Black refers to organization as "the
corporate aspect of social life, the capacity for collective
action" (p. 85). ”Measures of organization include the
presence and number of administrative officers, the
centralization and continuity of decision making, and the
quantity of collective action itself" (p. 85). "Any group
is, by definition, more organized than an individual on his
own” (p. 86). According to Black, "law varies directly with
organization” (p. 86) and ”the more organized the victim of
a crime, the more serious is the offense" (p. 95).

33

Therefore:

The type of victim influences the amount of law invoked
as a result of a criminal incident.

Specifically, businesses get more law than non-
businesses.

Only one study in the literature, Albonetti (1986),
assessed the impact that the type of victim (individual or
organized collective) had on decision making. In this
study, the type of victim did not affect the decision of
whether or not to continue prosecution at the
post-indictment stage.

Sociel Qeggg 1. Black describes social control as ”the
normative aspect of social life. It defines and responds to
deviant behavior specifying what ought to be... It divides
people into those who are respectable and those who are not"
(p. 105). Respectability refers to one’s normative status,
the degree to which a person has been subject to law and
other forms of social control. According to Black, "law
varies directly with respectability” (p. 112) with more
respectable victims receiving more law than non-respectable
victims. Since an indicator of victim respectability was
not available in the data collected here, a hypothesis
relating to this component of Black’s theory could not be

formulated or tested.

In addition to the victim characteristics identified by
Black, previous research has also suggested that the

preferences, or wishes, of the victim are important in

34

predicting decision responses of criminal justice actors.

In reference to police decision making, when the victim
prefers no arrest, the police will likely comply with the
request. In fact, in a study by Smith and Klein (1984), the
strongest determinant of an arrest was the complainant’s
request to have the offender arrested (also see Berk &
Loseke, 1981; Black, 1980; Friedrich, 1977; Lundman, Sykes,
& Clark, 1978; LaFave, 1965). Hence, on the basis of
previous research, one could expect that:

Victim desire for formal action influences the amount of
law invoked as a result of a criminal incident.

Specifically, victims who desire formal action get more

law than victims who do not desire formal action.

Finally, it is of interest in this study to explore the
impact of victim age on decision making. Only Bynum et al.
(1982) included age as a predictor of decision outcomes.
They found that victim age did not influence the amount of
effort devoted to follow-up investigations. Due to the lack
of ”theory" and previous research on this issue, a
hypothesis on this issue is not specified.

In sum, while the relationship between victim income,
victim-suspect relationship, and victim preferences and
criminal justice decision making is generally well
established and congruent with Black’s theory, the impact of
gender, racial identification, employment status, and victim
type is not. Furthermore, when focusing exclusively on the

research which has examined the impact of victim

35

characteristics on detective decision making in follow-up
investigations, it becomes apparent that the empirical
evidence is scant and often contradictory. As a result,
theoretical benefits could be realized from the provision of

additional evidence on these issues.

Offense characteristics and decision making

Along with victim characteristics, a host of studies
have examined the impact of offense ("legal")
characteristics on criminal justice decision making. These
studies, either implicitly or explicitly, have tested an
alternative conception of seriousness from that proposed by
Black (1976); specifically, that seriousness of the offense
is reflected not through the social structural
characteristics of the participants but through the nature
of the offense -- most cOmmonly, the extent of injury, the
amount of property loss, and the involvement of a weapon.
Also considered an offense characteristic, but not
reflective of ”seriousness,” is the strength of the
evidence. In the discussion which follows, research
findings concerning the influence of offense characteristics
on decision making are reviewed and, on the basis of this
review, hypotheses are developed.

Deggee_ef_lnjezz. Several studies have addressed the
expectation that more injury is reflective of a more serious
offense, and therefore deserving of "more law." According
to Schmidt and Steury (1989), victims who suffered moderate

36

or severe injury were more likely to see the case result in
a charge against the defendant. However, in relation to
police decision making, numerous studies (Berk & Loseke,
1980; Smith & Klein, 1984; Worden & Pollitz, 1984; Smith,
1987) found that the likelihood of arrest did not increase
if one of the disputing parties was injured (but see Waaland
a Keeley, 1985). In regard to detective decision making,
Bynum et al. (1982) found that the degree of injury did not
have a significant impact on the extent of the follow-up
investigation in personal offenses ("injury" was not
included in the analysis of property offenses).

In accord with the original expectation concerning the
relationship between degree of injury and decision making,
the following is hypothesized:

Degree of injury influences the amount of law invoked as
a result of the criminal incident.

Specifically, cases which involve more injury will get
more law than cases which involve less injury.

Velge gt Egepeggy Leee. Similar to the degree of

injury, one might expect cases with much property loss to be
viewed as more serious, and more deserving of attention,
than cases which involve little property loss. Adams and
Cutshall (1987) and Bynum et al. (1982) are the only
available studies which have examined the effect of property
loss on decision making. Adams and Cutshall (1987) found
the value of property loss to be of "marginal significance"

in the decision to prosecute. Bynum et al. (1982) found

37

that the extent of property loss did not have a significant
impact on the extent to which property offenses were
investigated ("loss" was not included in the analysis of
personal offenses). On the basis of the original
expectation, it is hypothesized that:

The value of property loss influences the amount of law
invoked as a result of a criminal incident.

Specifically, cases which involve much property loss get

more law than cases which involve little property loss.

Weapon Use. Several studies have examined the impact of
weapon use on decision making. The rationale for this
examination is that crimes which involve a weapon have a
potential for greater personal harm and are therefore "more
serious" and deserving of increased attention. In a study
conducted by Schmidt and Steury (1987), it was found that
cases which involved the use of a weapon in the commission
of the crime were more likely to proceed to court. However,
Nagel (1983) found that the commission of a crime with a
weapon did not affect the pre-trial release decision. In
reference to police decision making, Smith and Klein (1984)
and Smith (1987) found that the presence of a weapon at a
dispute did not significantly increase the probability of
arrest. No studies of detective decision making have
examined this relationship. In accord with the underlying
rationale for this examination, the following is suggested:

Weapon use influences the amount of law invoked as a
result of a criminal incident.

Specifically, crimes committed with a weapon get more
law than crimes not committed with a weapon.
38

Exifienee. Adams and Cutshall (1987), Albonetti (1986),
Burnstein, Kelly, and Doyle (1977), Schmidt and Steury
(1989), and Forst and Brosi (1977) found that the strength
of the evidence was an important factor in the prosecutor’s
determination of whether or not to issue charges or continue
prosecution of an offender; the stronger the evidence, the
more likely charges would be pursued. A similar
relationship between strength of the evidence and
disposition has been found concerning the decisions to
release on recognizance or bail (Frazier, Bock, & Henretta,
1980), sentence (Platt-Jendrek, 1984), and release on parole
(Heinz, Heinz, Senderowitz, a Vance, 1976).

Previous research also indicates that evidence plays a
major role in the police decision to arrest -- "the stronger
the evidence in the field situation, the more likely is an
arrest” (Black, 1971). Specifically, Black (1971) found
that police were more likely to make an arrest when they
actually witnessed a criminal incident as opposed to merely
learning of the incident through a citizen report.

Prior research on the criminal investigation process
suggests that detective decision making is also affected by
the strength of evidence. For example, in the seminal
research by Greenwood et al. (1977), it is stated that
”investigators choose the [cases] they will work by
considering... whether sufficient leads are present to
indicate that the chances of clearing the crime are high"
(p. 110). The observational studies conducted by Sanders

39

(1977) and Ericson (1981) also come to the same general
conclusion. Eck (1983) provides additional support to this
conclusion by finding that robberies receive more
investigative attention than burglaries because first,
robberies are viewed as more serious than burglaries and
second, robberies have a greater potential for better
evidence. Additionally, Bynum et al. (1982) found that the
amount of evidence was strongly related to the extent of
investigative effort in property offenses (i.e., more
evidence led to a more extensive investigation) but not in
personal offenses. In the Bynum et al. (1982) study,
"amount of evidence" was measured as an interval level index
through the presence of ten solvability factors: was there a
witness? can suspect(s) be named? can suspect(s) be located?
can suspect(s) be described? can suspect(s) be identified?
is the stolen property traceable? is there a significant
M.O. present? is there physical evidence present? has
evidence technician work been performed?

In accord with the previous research which has found a
relationship between strength of evidence and decision
making responses, it expected that:

Evidence influences the amount of law invoked as a
result of a criminal incident.

Specifically, the stronger the evidence, the more likely
the case will receive more law (where ”more law" equals
”selected for an investigation").

In regard to the amount of time spent on the follow-up

investigation, one might expect that when there is weak

4O

evidence in the case there will be little time spent on the
investigation because the detective does not expect much
chance of solution regardless of the activities performed
(Eck, 1979). Similarly, when the evidence associated with a
crime is very strong, the detective may not need to spend
much time on the investigation because the suspect can
easily be identified and arrested. However, cases in which
the evidence is of moderate strength may lead to much time
being spent on the investigation because the investigation
has a reasonable chance of solution if further information
becomes available. Therefore:
Crimes with evidence of moderate strength get more law
than than crimes with weak or strong evidence (where
”more law" equals "more time spent on an investigation”).
In sum, similar to the research on the relationship
between victim characteristics and decision making, there is
empirical support for the claim that criminal justice
decision making is influenced by offense characteristics.
While one might infer the same to be true in regard to
detective decision making, the relationship here is
generally not well established. As a consequence, there is
a need for additional research to assess the impact of

offense characteristics on detective decision making.

Decisions as Processes
The process tracing approach to decision making allows

for the examination of the actual cognitive processes

41

invoked to perform a decision task. Inferences are made as
to the nature of these processes not on the basis of
mathematical computations as with statistical models, but
rather on subjects’ search patterns and/or verbal reports of
the cognitive steps taken in order to make a decision.
Currently, in the criminal justice decision making
literature, there are no studies which have attempted to
cultivate such data. This is unfortunate because process
data would contribute additional insight into the
complexities of detective decision making and ultimately
further our understanding of the investigative process.

Process tracing data have been collected from decision
makers performing various decision tasks including consumer
product selections (Payne 5 Ragsdale, 1978; Olshavsky,
1979), clinical diagnosis (Einhorn, Kleinmuntz, a
Kleinmuntz, 1979; Hogarth, 1974), securities selection
(Clarkson, 1962), and problem solving type tasks such as
verbal analogies (Grundin, 1980), geometry theorems (Greeno,
1976) and playing chess (DeGroot, 1975).

Many of these studies used an information board to
present the stimuli for the decision task. An information
board consists of a matrix of alternatives (the thing to
choose; e.g., an apartment) and dimensions (information
about the thing to choose; e.g., cost of rent) for a
particular decision situation. Information boards are
either mechanically operated (Payne, 1976) or computerized
(Gilliland, 1990). With mechanically operated information

42

boards, cards with piecesof information are placed face
down to form the matrix of information and then subjects are
asked to manually turn over the cards on which information
is desired. Computerized information boards provide for the
display and search of information by pressing computer
terminal command keys. Regardless of the type of
information board used, information search patterns can be
observed through manifested actions and supplemented with
verbal reports of thoughts and actions.

The majority of process oriented studies have been
conducted in laboratory settings with student subjects.3
As a result, this methodology has not been well tested in
field settings. Given this factor along with the
observation that detectives (and the police in general) are
protective of their work, suspicious of outsiders, and
generally non-cooperative (cf. Ericson, 1981), an issue of
concern in this study is whether it is feasible to collect
process data through the use of an information board from
detectives in the field setting (this issue is discussed in

4
the final chapter).

The process tracing literature has identified two
decision making strategies -- linear (compensatory) and
non-linear (non-compensatory) (Payne, 1976). An individual
who uses a linear strategy in making a decision examines a
constant number of dimensions across alternatives, mentally
assigns weights to each of the dimensions, sums the negative

43

and positive weights for each alternative, then chooses the
alternative with the highest overall score (Payne, 1976). A
linear strategy of decision making is evident in the
following verbal protocol obtained from a student selecting
a hypothetical apartment:

O.K., we have an A and a B. First look at the rent for

both of them. The rent for A is $170 and the rent for B

is $140. $170 is a little steep, but it might have a

low noise level. So we’ll check A’s noise level. A’s

noise level is low. We’ll go back to 8’3 noise level.

It’s high. Gee, I can’t really very well study with a

lot of noise. So I’ll ask myself the question, is it

worth spending that extra $30 a month for, to be able to

study in my apartment (Payne, 1976, p. 378).

Apparently for this individual less expensive rent could
compensate for a higher noise level in selecting an
apartment.

Conversely, with a non-linear strategy, a variable
number of dimensions across alternatives are examined
(Payne, 1976) and ”a low score on one dimension cannot be
compensated for by a high score on another dimension" (Ford
et al., 1989, p. 77). Within the non-linear strategy,
several "substrategies" of decision making have been
identified -- conjunctive, disjunctive, lexographic, and
elimination by aspect (Svenson, 1979; Olshavsky, 1979;
Payne, 1976). A eenjeneeixe strategy is used when the
decision maker assigns an acceptable value to each important
dimension, and if the acceptable value is not obtained for
each dimension, then the alternative is eliminated. With
the diejnneeige strategy, the decision maker once again

establishes acceptable values for each important dimension.

44

However, the alternative which is chosen exceeds the minimum
value on at least one dimension while all of the other
alternatives would be equal or fall below the minimum

value. With the lexicographic strategy, dimensions are
rank-ordered in terms of importance. An alternative is then
selected based on the ranking of the most important
dimension. Finally, a decision maker who uses the
elimipation by espect strategy once again rank-orders
dimensions within each alternative but the alternatives in
which a dimension does not meet a minimum value are
eliminated from consideration. A non-linear (elimination by
aspect) decision strategy is apparent in the protocol below:

Since we have a whole bunch here, I’m going to go across

the top and see which noise levels are high. If there

are any high ones, I’ll reject them immediately (Payne,

1976, p. 375).

Apparently, an attractive dimension of an alternative, such
as inexpensive rent, could not compensate for a high noise
level.

Statistical models of decision making assume that
decision makers use linear decision strategies. Research
has shown, however, that under certain conditions this
assumption is inaccurate. For example, increased task
complexity, generally defined in terms of the amount of
information available to the decision maker, has been
associated with the use of non-linear, non-compensatory

decision strategies (Payne, 1976; Onken, Hastie, & Revelle,

1985; Olshavsky, 1979). Non-linear strategies serve to

45

limit the amount of information to be processed by the
decision maker thus simplifying the decision task (Onken et
al., 1985). These simplifying strategies may be used early
in the task and then, when some of the alternatives have
been eliminated from consideration, the decision maker may
switch to linear strategies (Payne, 1976).

Given the amount of information available to detectives
when making decisions, and therefore the seemingly complex
nature of the decision tasks, one might expect that
detectives employ, to a large extent, non-linear decision
making strategies. Previous research on investigative
decision making has not addressed this expectation.
Therefore, another question addressed in this study is the
extent to which detectives use linear (vs. non-linear)
strategies in making decisions.

Process tracing research has also illustrated that
decision making involves search processes -- processes which
vary in depth, sequence, content, and latency (Ford et al.,
1989). Of direct concern in this study are depth of search
and content of search. erph_ef_eeepeh refers to the
proportion of total information examined prior to rendering
a decision. Through an examination of a decision maker’s
depth of search, it is possible to infer the existence of
linear or non-linear decision making strategies (Payne,
1976). For example, searching a large proportion of
information implies a linear strategy while the search of a

small proportion indicates the use of a non-linear

46

strategy. In addition, as explained by Payne (1976),
searching a constant number of dimensions (information)
across alternatives implies the use of a linear strategy
while searching a variable number of dimensions across
alternatives suggests that the decision maker was using
non-linear strategies (also see Ford et al., 1989).
Therefore, it is useful to examine the proportion of case
information searched by detectives in making decisions.
Qentepp_e£_eeepeh refers to the specific elements of
information examined by a subject in making a decision.
From an examination of search content, one can specify the
dimensions upon which decisions are (at least partially)
based. For example, Payne and Ragsdale (1978) attempted to
describe the extent to which certain consumer product
attributes (e.g., price) were mentioned (and presumably
considered) in making decisions to purchase grocery items.
Similarly, through an analysis of detectives’ search
patterns and verbal reports, insight into the case factors
most often considered in decision making can be obtained and
the influence of other factors on decision making can be
illuminated. An examination of search content offers an
alternative means by which the hypotheses concerning case
characteristics and decision making can be examined.
Accordingly, the process data will be used to identify the
information elements (case characteristics) which are most

important to detectives in making decisions.

47

Summary

Chapter Two has provided the theoretical and analytic
foundation for this study. The case selection decision and
the time allocation decision were presented as the two major
decisions of detectives. The outcome and process oriented
approaches were identified as the two approaches used in the
study of decision making. Propositions which predict the
decision responses of detectives were derived from Black’s
theory of the behavior of law and previous research, and
research questions concerning the cognitive processes of
detectives were developed on the basis of the process

tracing literature.

48

Footnotes

As Hembroff (1987) illustrates, the theory is not

limited to predicting the actions of individuals within
the criminal justice system. Rather, the theory predicts
when and how much law will be invoked by any individual in
all types of situations. While the decisions may differ
by the actor, all can be equated with "more or less law"
as discussed by Black.

Qualitatively, law can take several forms: penal,
compensatory, therapeutic, or conciliatory.

Clarkson, 1962; Hogarth, 1974; and Payne and Ragsdale,
1978, are notable exceptions.

As discussed earlier in this chapter and in detail in
Chapter Four, along with the information board as a method
of collecting process data, the less rigorous method of
field observations and interviews was also used to collect
data on how detectives make decisions. The use of this
method in this manner is also a move into unchartered
territory (Mastrofski & Parks, 1990).

49

CHAPTER THREE

RESEARCH SITE

Chapter Three contains a description of the study site.
The city in which the police department is located is
briefly described, the features of the police department are
discussed, and the mechanics of the criminal investigation

process within the department are outlined.

The City

The City of Landau (a pseudonym) is a medium sized
midwestern city with a (1990) population of approximately
130,000 people, approximately 33 percent of which are
non-white. The greater metropolitan area has a population
of approximately 450,000 people. The city is located on 34
square miles of land. In 1980, the city contained 49,516
households. Landau is administered by a council-mayor form
of government. The major employers in the City of Landau
are manufacturing and assembly plants, state government,
retail, and health care. The unemployment rate in 1986 was
7.0%.

50

The Landau Police Department

At the time of this study, the Landau Police Department
employed 245 sworn officers, eleven non-sworn officers, and
91 civilians. In 1989, the department responded to 128,442
calls for service. The operating budget for 1989 was
$12,388,532.

In Figure 1, the organizational chart of the L.P.D. is
illustrated. As seen, the department is managed by a chief
of police. A deputy chief and assistant chief are
responsible for the two major operating components of the
department -- staff services and field services,
respectively. The staff services component consists of the
administrative support division (administrative services,
personnel and training) and the operations support division
(records, radio maintenance, and communications center).
The field services component consists of the uniform
division (special services and patrol) and the
investigations division. Each division within the
department is supervised by a captain.

During this study, 30 of the 245 sworn officers in the
department were assigned to the investigations division.
Twenty-two of these personnel were the rank of detective,
one was a patrol officer temporarily assigned to the
division, four were sergeants, two were lieutenants, and
there was one captain. All of the detectives worked

primarily a fixed shift of 8 a.m. to 4 p.m., Monday through

51

 

 

 

 

 

til...

BEESSL éimijg

 

 

_SZ>5§_
23:58

g

 

 

L
m

 

 

 

 

 

 

 

%

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E

 

 

 

mmmmmmmmmmmm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

see some
g5
53 use

gas
SE.
$833.9: :5:

3:: bag
gain
$88.33
see.
85238.5
0 .223: 30a 23

gig

th

Friday. One detective was on-duty every Saturday from 8
a.m. to 4 p.m. During off-duty hours, detectives (and any
of the supervisors) could have been called in to investigate
(or supervise) major crimes such as homicides, bank
robberies, serious assaults, etc.

The investigations division contained three squads:
"youth," "crimes against property," and ”crimes against
persons." Each squad was supervised by a sergeant. The
youth squad had the responsibility for investigating crimes
in which a juvenile was the accused. However, they also
investigated child abuse and neglect cases, as well as all
sex-related crimes. Six detectives were assigned to this
squad.

The crimes against property squad investigated
burglaries, arsons, larcenies, auto thefts, fraud, malicious
destruction, and worthless document cases. Ten detectives
were assigned to this squad. Within this squad, there was
an informal division of labor among the detectives --
detectives specialized in the investigation of one or, in
some cases, two types of crime. For example, one detective
specialized exclusively in the investigation of burglaries
and as a result, this detective got assigned only burglary
cases. Four other detectives (including the patrol officer)
investigated burglaries along with either larcenies, auto
thefts, or arsons.

The crimes against person squad was responsible for

53

investigating homicides, assaults, robberies, and other
lesser crimes against the person. Eight detectives were
assigned to this squad. Like the crimes against property
squad, there was an informal division of labor among the
detectives. For example, two detectives specialized in
robbery investigations and as a result, these two detectives
got assigned the vast majority of robbery cases. Homicide
investigations were assigned, on a rotating basis, to teams
of typically two detectives from the squad. Six of the
eight detectives were routinely used in homicide
investigations.

Prior to their assignments as detectives, all but one of
the detectives in the investigations division were assigned
to the patrol unit as patrol officers. All of the
detectives were assigned on a permanent basis to their
respective squads. When a detective position became vacant
in another squad, detectives were able to apply for the
position. If an inter-squad transfer occurred, it was
usually from the youth squad to either the crimes against
person or crimes against property squad.

Detectives were evaluated on their performance
annually. Detectives were judged on the basis of "job
knowledge," "cooperation," "acceptance of supervision,”
etc. They were not formally evaluated on the disposition of
the cases they were assigned. The same performance
evaluation form was used for all municipal employees in
Landau.

54

The investigations division occupied one of six floors
in the police department building. Each squad had its own

partitioned area on the floor.

The Investigative Process at the Landau Police Department

The criminal investigation process of the Landau Police
Department was similar to the process described in Chapter
One. However, in order to understand the context and
complexities of the L.P.D. criminal investigation process,
it is helpful to describe the details of the process.

During the time of this study, the L.P.D. operated a
Differential Police Response (DPR) strategy of call
management (McEwen, Connors, & Cohen, 1986). As part of
this strategy, certain types of citizen reports are
identified as not requiring mobile police responses.
Instead, these designated complaints are taken over the
telephone by police telephone operators. At the L.P.D.,
complaints handled in this manner were assigned a ”No Report
Forthcoming" (NRF) status and were very rarely brought to
the attention of the investigations division.

The decision of whether or not to dispatch a mobile
police unit for the complaint was the discretion of the
telephone operator but was guided by departmental policy.
According to departmental policy, for a report to be taken
over the telephone, none of the following circumstances

could exist:

55

1) there are known suspects, or a description of an
accused subject or vehicle is available
2) there is a witness to the crime
3) there is evidence to be tagged into the Quartermaster
(evidence room) or processed at the scene
4) the incident involved the use of weapons or resulted
in serious injury
5) property loss or damage exceeds $1,000
6) there is damage to public or police property
7) the offense is related to another offense in which a
report is required
8) important or potentially useful information regarding
the incident exists that should be included in a
written report
The absence of these circumstances was necessary for a
telephone report but were not sufficient -- complaints which
did not contain these criteria gel have still received a
mobile police response.
When a complaint was assigned a mobile police response,
a patrol officer had the responsibility of performing the
initial investigation. The activities performed and the
information collected by the patrol officer during the
initial investigation were recorded on a series of report
forms: the investigative report form (Appendix A) and
supplements for the narrative (Appendix B), the modus
Operandi descriptor form (Appendix C), the personal
descriptor form (Appendix D), the vehicle descriptor form
(Appendix E), and the property form (Appendix F). Depending
on the particular crime and the amount of information
available, any combination of these reports may have been
completed (however, the investigative report was always

completed). The forms were completed by the responding

officer and were most often in handwritten form. After the

56

reports were completed, they were held in the uniform
division office until the beginning of the next morning
shift when each of the detective sergeants received the
reports pertinent to their squad.

The sergeant from each squad then had the task of
deciding which cases to assign to the detectives for
follow-up investigations. At the Landau P.D. this decision
was guided by solvability factors. Solvability factors are
key pieces of crime related information which, when present,
increase the probability of the perpetrator being
apprehended and hence, the crime being ”solved” (Hastings,
1980). The solvability factors used by the L.P.D. were
listed on the back of the investigative report form and
consisted of the following:

1) Were suspects arrested?

2) Was there a witness to the crime?

3) Can the suspect be identified by a witness?

4) Can a suspect be named?

5) Is a suspect described?

6) Is the suspect known and/or can he/she be located?

7) Was there a significant modus Operandi present?

8) Was there significant physical evidence present?

9) Is the stolen property identifiable?

10) Is there a significant suspect vehicle description?

11) Are there undeveloped leads?

12) Gravity of offense...

Value over $1,000?
Damage over $1,000?
Serious injuries / hospitalization required?
Weapons involved?
These factors were used on an infeppel basis to sensitize
the sergeant to factors which should be considered when

determining which cases to assign for a follow-up

investigation. The factors were net used to calculate a

57

solvability score as is common with the use of screening
devices (Eck, 1983). In fact, often the presence of these
factors was not noted on the investigative report form.
Therefore, the case screening decision at the L.P.D. most
closely resembled an unaided-analytic strategy as described
in Chapter Two.

Cases where the suspect was arrested (the first
"solvability factor" listed above) involved situations where
a patrol officer made an arrest prior to the selection
decision. These cases were formally referred to as
"in-custody cases" -- where the culprit was, at the time of
the initial review of the report, being held in the police
department detention facility. All "in-custody" cases were
assigned to a detective for a follow-up investigation.

Detectives were responsible for conducting follow-up
investigations on those cases they were assigned. In
conducting follow-up investigations, detectives recorded on
the ”case log” or ”turn back sheet" (Appendix G) the
activities they performed in the investigation (e.g.,
reviewed report, interviewed victim, interviewed witness,
talked to prosecutor, etc.) as well as the total amount of
time spent on the investigation (from when the case was
received until it was closed in some manner -- see below).
In nearly all of the cases, a narrative of the activities
performed, and the information produced, was provided on
supplemental sheets and was attached to the case log. These
reports were most often handwritten although more detailed

58

and complicated investigations tended to have typewritten

narratives.

Cases were assigned one of several statuses upon

conclusion of the follow-up investigation. Each case status

represented a means by which the case could be "closed.”

They were:

1)

2)

3)

WWW - This
status is entered when the accused has been charged
with another crime and will not be charged under this
incident. (Must meet first four criteria [a-d] for
exceptional clearance below).

e ar e - This status is used
when all of the following questions (a through d) can
be answered "yes": ‘

a) Has the investigation definitely established the
identity of the offender?

b) Is there enough information to support an arrest,
and an acceptance by the prosecutor?

o) Is the exact location of the offender known so
that he could be taken immediately into custody?

d) Is there some reason outside police control that
prevents an arrest, charge, and acceptance by the
prosecutor?

In addition, for an offense to be exceptionally
cleared it must fall into one of the following
categories:
e) Is there no other more specific Landau Police
Department status definition?
f) Did the offender commit suicide?
g) Was the offender a victim of a homicide?
h) Did the offender die after making a
confession?
i) Was the offender killed by the police?
j) Was extradition denied?
k) Was the incident referred to a non-police
agency such as Probation, Parole, Postal
Service?

h ve - This status is used
when all leads have been exhausted and without
additional information the investigation can go no
further.

4) WWW - This status is

used when a completed investigation is reviewed by

59

the prosecutor’s office and they refuse to issue a
warrant to the victim or detective.

5) Re re r a e t — This status is used
when a juvenile accused is identified and petitioned
to probate court.

6) o d N - This status is assigned when it has
been determined that a reported crime was not
committed.

7) V c ‘ Ca v t V - This status is

assigned upon verification with the victim that
he/she no longer desires to pursue the complaint, or
when the victim fails to take effective action in the
investigation within a reasonable amount of time.

8) V t use 8 u e V - This status is ,
assigned upon verification with a victim that he/she
does not wish to prosecute, or when the victim fails
to take effective action toward prosecution within a
reasonable period of time.

9) Wa a t t WPA - This status is
assigned when an offender has been identified and a
valid warrant is received by the prosecutor’s office.

When a case was closed with a "WPA" status, the

outstanding warrant ("want") was entered into the LEIN (Law
Enforcement Information Network) computer system. The
patrol division then executed the warrant and an arrest
would be made. As a result, the detectives made extremely
few physical arrests.

In addition, to the above statuses, two others were used

when the complaint was currently being investigated or was

not assigned for a follow-up investigation:

10) Open Aeaigned (92A) - This status is used when the

squad supervisor initially assigns a case to a
detective.

11) Qpep_flpeeeigne§_(9£fll - This status is used when the

squad supervisor determines that the case will not
be assigned to a detective. This status also
represents a case closure.

60

There were no formal policies or procedures on how
investigations should be conducted or on how long cases may
remain open ("OPA") before they needed to be closed in some
manner. There were no formal guidelines regarding the
completion of the turn-back sheet or the calculation of
amount of time spent on the investigation (an assessment of
the accuracy of the reports is offered in Chapter Six).

After a case was closed, some case information
(complaint number, sergeant who made screening decision,
detective assigned case, and amount of time spent on the
investigation) was entered into the department’s ”Data-Flex"
computer program. After the entry of this information, the
cases were taken to the records bureau where each case
(initial and/or follow-up reports) was placed in an envelope
and filed in chronological (complaint number) order within
the established filing system.

The Landau Police Department operated a Law Enforcement
Management System (LEMS) computer system which contained
case data (complaint number, type of crime, victim name,
address, age, race, sex, value of property loss, sergeant
who selected the case for an investigation, detective
assigned to the case, and status of the investigation) on
all of the complaints taken by the police department. Also
contained within the LEMS system was a name file which
listed all individuals who had a criminal contact (as either
a victim or an accused) with the Landau Police Department
during the past ten years. The file provided the queried

61

individual’s identifying data (date of birth, sex, race,
age, etc.) and criminal history.

Another capability of the LEMS system was the tracking
of pawn shop property. All pawn shops were required by
state law to furnish to their local police agency a listing
of all property which they purchased. The store was
required to complete a form which contained the seller’s
name, address, sex, age, race, thumbprint, and the serial
number (if applicable) of the merchandise being sold. This
form had to be submitted to the police department within
forty-eight hours of the transaction. Once received by the
police department, the information was entered into the pawn

shop property file within the LEMS system.

Summary
This chapter contained a description of the study site.
The City of Landau and its police department were described
and the mechanics of the investigative process within the

Landau Police Department were discussed.

62

CHAPTER FOUR

METHODOLOGY

Three data collection methods were used in this study.
First, case characteristics were coded from investigative
reports and the resulting data were used to determine the
relationship between the case characteristics and decision
outcomes. Second, an information board was used to collect
verbal protocol data from detectives. These data provided
insight into the cognitive processes associated with
decision making. Third, observations of, and interviews
with, detectives allowed for additional insight into the
factors which influence decision making and the cognitive
processes associated with decision making. The observations
and interviews also provided a means by which the meanings
ascribed to case characteristics could be explored. Each of
these methods and their associated procedures are discussed
in this chapter. In addition, the variables of interest in
the study are defined and the research questions and

hypotheses are stated.

63

Decisions as Outcomes

The data for this component of the study came from
investigative reports completed by patrol officers and
detectives. Reports which identified: (1) a burglary or
robbery; (2) the initial investigation report as being
referred to the investigations division for a possible
follow-up investigation; (3) an arrest of a suspect as not
having occurred prior to case assignment; and (4) the crime
as having occurred between July 1, 1989 and June 30, 1990 (a
one year time frame) were included in the sample. In
addition, cases which were investigated only after
information was obtained which would enable the cases to be
immediately closed were excluded from the sample. The
rationale for each of these criteria is discussed below.

Burglary and robbery investigations were the focus of
this study for two reasons. First, burglary and robbery are
relatively common offenses which consume a large proportion
of a police department’s investigative resources.
Therefore, to understand the investigation of these crimes
is to understand much of police investigations more
generally. Second, the strength of the evidence associated
with the two crimes is often quite different (e.g., there is
often an eyewitness in a robbery but not in a burglary).
Thus, it is possible to examine the relative impact of
evidence strength on decision making not only within each

crime type but also across crime types.

64

it.
8e

He‘

Only criminal incidents where a mobile police response
unit was dispatched to conduct an initial investigation were
included in the study. Because reports taken over the
telephone ("NRF reports”) very rarely came to the attention
of the investigations division, it would be inappropriate to
include these types of cases in the sample.

Crimes in which an arrest was made prior to case
assignment (e.g., an arrest was made during the initial
investigation; ”in-custody” cases) were not included in this
study because all of these cases received investigative
attention. Thus, variance in the dependent variables of
interest would not be provided.

Criminal incidents which occurred between July 1, 1989
and June 30, 1990 were included in this study. The twelve
month time frame balanced the need for adequate frequencies
with the limited project resources.

Finally, cases assigned to a detective only after
information became known which would enable the case to be
immediately "closed” were excluded from the population. As
a result, two types of cases were eliminated from the
population. First, cases where a person confessed to
committing a crime (which was not initially selected for an
investigation) while being questioned about another crime
were excluded. Second, cases where the detective(s)
inferred that an identified individual was responsible for
several other crimes (which were not initially selected)

were eliminated.

65

Case Selection Procedure

The case selection process consisted of several steps.
First, a computer print-out of all of the burglaries and
robberies which were reported between July 1, 1989 and June
30, 1990 was obtained from the department’s LEMS computer
system. The print-out listed the cases in chronological
order by complaint number and also identified the status of
each case. The list identified 339 robberies and 1,674
burglaries.

All of the reports which identified a ”No Report
Forthcoming" (NRF) status were then excluded from the list.
No robberies were excluded while 679 burglaries (of 1,674;
41%) were excluded due to a NRF status.

All of the remaining cases were then reviewed and, as
discussed above, several more categories of cases were
eliminated. First, all of the reports which stated that an
arrest was made prior to case assignment were excluded from
the population of cases. This resulted in the exclusion of
32 robberies and 73 burglaries.

Second, all of the cases which were assigned to a
detective only after information became available which
would allow the case to be immediately closed were excluded
from the population. On the basis of this criterion, no
robberies were excluded while 32 burglaries were excluded.
Third, all of the cases which were missing from the files
were excluded from the population. Accordingly, two

robberies and 33 burglaries were excluded.

66

As a result of the entire case selection procedure, 305
robbery cases and 857 burglary cases were included in the
sample for a total N of 1,162 cases. For each of the 1,162
cases, the initial and/or follow-up investigation reports
were content analyzed and a case data form (Appendix H) was

completed.

Variables

The independent variables included in this portion of
the study consisted of victim and offense characteristics.
Data on most of the independent variables were obtained from
the initial investigation reports and were transcribed as
recorded by the patrol officer who completed the report.
Specifically, the victim characteristic variables consisted
of: victim type (business/individual), age, sex, race
(white/non-white), employment status (employed/not
employed), victim-offender relationship (yes/no), desire for
investigative effort (yes/no), and income.

When a crime was directed toward a business (as in a
burglary) or a representative of a business (as in a
robbery) the other demographic characteristics of the victim
(i.e., age, sex, race, income, and employment status) were
coded as missing. Based on discussions with detectives, an
assumption was made that the business characteristic would
override the influence of any other victim characteristics.

For example, it was assumed that a robbery perpetrated

67

against a bank would carry a significance not mitigated by
whether or not the teller was a male or female.

Data on the presence of a relationship between the
victim and offender were obtained from the initial
investigation reports and recorded as disclosed by the
victim. As defined here, a relationship between a victim
and an offender could have taken the form of acquaintances,
friends, lovers, spouses, children, etc. Also included were
terminated relationships (e.g., ex-lovers, ex-spouses,
etc.). These data were based on who the victim and/or
witness believed to be the culprit. This belief however,
was not always based on eyewitness knowledge. For example,
burglary victims often offered a guess as to who they
thought might have perpetrated the act. If it was not
”known” who committed the crime at the time of the initial
investigation, a relationship was considered as being
absent.

Data on desire for investigative effort were obtained
from the follow-up investigation reports. If, upon
detective contact with the victim, it was learned that the
victim no longer wished to pursue or prosecute the
complaint, or if the victim failed to cooperate in the
investigation (e.g., return phone calls), desire for effort
was coded as ”no.” In these situations, the case would be
closed by the detective as either ”VCI” (victim cancels

investigation; if the culprit was not positively identified)

68

or
pass
eff
inv
des
the
in'

811

CE

US

de

or "VRP" (victim refuses to prosecute; if the culprit was
positively identified). If the victim did not state that
effort was not desired and the victim cooperated in the
investigation, or if the victim was not contacted, then
desire for effort was coded as ”yes." Because the wishes of
the victim first became explicit during the follow-up
investigation, this variable was not included in the
analysis of the case selection decision.

Victim income was measured by the median income of the
census tract in which the victim resided at the time of the
incident. To obtain these data, the victim’s address was
first recorded from the initial investigation report and
then, using a map of the city, the address was placed in its
respective census tract. Using 1980 census data, the median
income of the victim’s census tract was then obtained. This
procedure was the same as that used by Bynum et al. (1982)
and similar to that used by Smith (1987). The estimation of
an individual’s income (or any other characteristic) based
on aggregate data poses certain analytical difficulties.
However, if detectives wish to consider "income” in making
choices, they are likely to infer this information from the
address listed on the initial investigation report because
no other more direct information is provided in the report.

Offense characteristic variables consisted of: weapon
used in offense (yes/no), dollar value of stolen property,

degree of personal injury (none / minor (bruises, black

69

36
be
in

Va:

eyes, cuts, swelling] / serious [broken bones, broken teeth,
loss of consciousness, stitches]), and several evidence type
variables: physical evidence available (yes/no), suspect
vehicle described (yes/no), suspect vehicle license plate
known (yes/no), stolen property identifiable (yes/no), and
the cornerstone of all evidence, strength of suspect
information. Weapon use and degree of personal injury were
included only in the analysis performed on robbery
investigations.

The dollar value of the stolen property was defined as
the value of the property taken by the culprit as a result
of the crime. The amount of loss was recorded as estimated
by the victim at the time of the initial investigation.

For significant physical evidence to be available,
fingerprints had to be lifted from the crime scene, shoe or
foot prints had to be photographed or tracked, or what was
believed to be the culprit’s personal belongings (clothing,
tools, notes, etc.) had to be present at the crime scene.
Tool marks and unphotographed shoe/foot prints were
considered insignificant (but common) types of physical
evidence.

In order for stolen property to be identifiable, the
serial number or some other engraved number (or name) had to
be known and reported by the victim at the time the initial
investigation was conducted. For stolen property to be of

value in an investigation, it had to be traceable (e.g.,

70

through pawn shop records). In order to be traceable, the
property had to have identifying information such as serial
numbers, etc.

The suspect information variable incorporated what was
known about the culprit (description, identification, name)
end how this information was produced (on the basis of a
victim [or witness] guess, an eyewitness to the incident, or
some other witness account). A description refers to
information about the physical characteristics of the
culprit while an identification means that the witness could
recognize the culprit if seen again (either in person or by
photograph). This measurement scheme appears to capture
more of the construct’s complexity than a simple tally of
how many suspect related solvability factors were present at
the conclusion of the initial investigation (e.g., suspect
named? suspect described?, etc.; as in Bynum et al., 1982).

As collected, the strength of suspect information
variable had ten values with (1) representing the weakest of
information and (10) being the strongest: (1) no significant
suspect information available, (2) a suspect was described
and could be identified on the basis of a (victim) guess
only, (3) the culprit could be described by an eyewitness to
the crime, (4) a suspect could be named by a guess, (5) the
culprit could be described and identified by an eyewitness
to the crime, (6) the culprit could be named by being seen

in the area at about the time the crime occurred, (7) the

71

culprit could be named on the basis of a guess which was
supported by other witness information, (8) the culprit
could be named by an individual who saw the culprit with
stolen property or heard the culprit confess to committing
the crime, (9) the culprit could be named in some other way
(i.e., through patrol activities), (10) the culprit could be
named by an eyewitness to the crime. Based on observations
and detailed discussions with detectives, and given the
expectations outlined in Chapter Two, this scale was divided
into three categories for the analyses: weak suspect
information (1,2,3), moderate suspect information
(4,5,6,7,8,9), and strong suspect information (10).

Data on the dependent variables, case selected for
follow-up investigation (yes/no) and amount of time spent on
the follow-up investigation (in hours), were obtained from
detectives’ follow-up investigation reports which were
contained in the case files. If a case had a follow-up
report contained in the file, then the case was considered
as having been selected for a follow-up investigation. If a
follow-up report was not in the file, and the computer
print-out verified that the case was not selected for a
follow-up investigation, then the case was considered as not
having been selected for a follow-up investigation.

Data on the amount of time spent on follow-up
investigations were obtained from the "turn back sheets”

which required the detectives to record the total amount of

72

time spent investigating the case from the point when the
case was received until the case was closed. Completion of
the "turn back sheets" was an existing procedure of the
department, one not introduced by the researcher.

Along with the total amount of time spent on
investigations, descriptive data were also collected on neg
time was spent -- on the activities performed during the
follow-up investigations. Through a review of the follow-up
investigation narratives, it was possible to identify (at
least most of) the activities which were performed in each
of the investigations. An initial sample of 43 burglary and
robbery follow-up reports were reviewed and, on the basis of
this review, 13 activities were identified as consuming the
vast majority of all investigative time. These activities,
which were similar to those specified by Eck (1983), were
then listed on the case data form. The activities consisted
of: victim interviewed, witness(es) interviewed, witness
canvass conducted, others interviewed (e.g., individuals
other than those involved in the crime -- mother of suspect,
non-Landau police personnel, probation officer, pawn shop
personnel, etc.), crime scene searched, physical evidence
submitted for analysis, computer files searched, photo
line-up conducted, mug-shot books shown, physical line-up
conducted, suspect interviewed, informants (”street
sources”) interviewed, and prosecutor consulted. Two

activities were constant for all investigations: reviewed

73

initial investigation report and wrote the follow-up
investigation report. Because of their non-variability,
these activities were not included on the data form or in

the univariate analysis.

Analysis

The data in this component of the study were analyzed
through the use of univariate, bivariate, and most
importantly, multivariate statistical procedures. The first
set of analyses had ”selected for a follow-up investigation”
as the dependent variable. Because the overwhelming
majority of all robberies were assigned for a follow-up
investigation (96%), the selection decision did not often
exist in the investigation of robberies. Therefore, the
only selection decision that was analyzed was in reference
to burglaries. Because linear regression and analysis of
variance approaches are often considered inappropriate when
employing a dichotomous dependent variable (King, 1986), a
loglinear regression (probit) model was used to determine
the relative impact of the independent variables on the
dependent variable.1 The second set of analyses were
conducted on those burglary and robbery cases which were
selected for follow-up investigations. In these analyses,
the amount of time spent on the follow-up investigation (in
hours) was used as the dependent variable in a linear

regression procedure. The case characteristic variables

74

were then used to predict this variable. Separate analyses

were conducted on cases within each crime type.

Hypotheses

Congruent with the propositions developed in Chapter

Two, the hypotheses examined in this component of the study

were 3

Hypothesis

Hypothesis

Hypothesis

Hypothesis

Hypothesis

1:

Victim income influences detective decision
making. Crime victims who live in higher
income census tracts are more likely than
victims who live in lower income census tracts
to have their cases selected for an
investigation and have more time spent on the
investigation.

° Victim gender influences detective decision

making. Male crime victims are more likely
than female crime victims to have their cases
selected for an investigation and have more
time spent on the investigation.

° Victim employment status influences detective

decision making. Crime victims who are
employed are more likely than victims who are
not employed to have their cases selected for
an investigation and have more time spent on
the investigation.

Victim-suspect relationship influences
detective decision making. Victims who do not
have a relationship with the suspect are more
likely than victims who do have a relationship
with the suspect to have their cases selected
for an investigation and have more time spent
on the investigation.

° Victim race influences detective decision

making. White crime victims are more likely
than non-white victims to have their cases
selected for an investigation and have more
time spent on the investigation.

75

Hypothesis 4:

Hypothesis 5:

Hypothesis 6:

Hypothesis 7:

Hypothesis 8:

Hypothesis 9:

The type of victim influences detective
decision making. Businesses are more likely
than non-businesses to have their cases
selected for an investigation and have more
time spent on the investigation.

Victim desire for effort in an investigation
influences detective decision making. Victims
who desire effort are more likely than victims
who do not desire effort to have more time
spent on their investigation. (Time as
dependent only).

Degree of injury influences detective decision
making. The more injury which results from
the crime, the more likely the case will be
selected for an investigation and have more
time spent on the investigation. (Robberies
only).

Value of property loss influences detective
decision making. Cases with a higher value of
stolen property are more likely than cases
with lesser value to be selected for an
investigation and have more time spent on the
investigation.

Weapon use influences detective decision
making. Cases which involve the use of a
weapon are more likely than cases which do not
involve a weapon to be selected for an
investigation and have more time spent on the
investigation. (Robberies only).

Evidence influences detective decision
making. Specifically:

Hypothesis 9a: Cases with stronger suspect information

are more likely than cases with weaker
suspect information to be selected for an
investigation. (Selection as dependent).

Hypothesis 9b: Cases with suspect information of

moderate strength are more likely than
cases with weak or strong suspect
information to have more time spent on
the investigation. (Time as dependent).

Hypothesis 90: Cases with physical evidence are more

likely than cases without physical
evidence to be selected for an
investigation and have more time spent on
the investigation.

76

Hypothesis 9d: Cases with a suspect vehicle description
are more likely than cases without a
vehicle description to be selected for an
investigation and have more time spent on
the investigation.

Hypothesis 9e: Cases with a suspect vehicle plate known
are more likely than cases without a
vehicle plate known to be selected for an
investigation and have more time spent on
the investigation.

Hypothesis 9f: Cases with identifiable stolen property
are more likely than cases without
identifiable stolen property to be

selected for an investigation and have
more time spent on the investigation.

Decisions as Processes
Data on the cognitive processes associated with

detective decision making were collected through the use of
an information board as well as through observations of, and
interviews with, detectives while they performed their
decision tasks. Discussed below are first, the details of
the information board data collection effort and second, the
observational methodology. While the information board data
were collected and analyzed in reference to each particular
investigative decision, the data which were derived from the

observations focused only on the time allocation decisions.

The Information Board Method / Verbal Protocol Analysis
Subjects

For the case selection decision in burglaries, the three

77

detective sergeants who were assigned to the investigations
division during the data collection time period and within
the crimes against property squad for at least some time in
the past year participated in this segment of the study.
For the prioritization of burglary cases, all five
detectives2 who were assigned to the property squad during
the data collection time period and routinely investigated
burglary cases participated in the information board
exercise. For the prioritization of robbery cases, both of
the detectives who were assigned to the crimes against
persons squad during the data collection time period and
routinely investigated robbery crimes participated in the
exercise. Considered together then, ten of the ten eligible
investigators (100%) participated in this component of the

3
study.

Information board structure and content

The decision task for the detective sergeants was to
decide which of five cases to assign to detectives for
follow-up investigations. The decision task for the
detectives (burglary and robbery) was to identify which
case, out of the five they were assigned, would receive top
priority (defined as the case on which they would be willing
and likely to spend the most investigative time) and then
prioritize the remaining four cases. As discussed in

Chapter Two, both of these decision tasks reflect the

78

typical sort of decisions the actors must make on a daily
basis.

In each situation, an information board was used to
present case information. The information board consisted
of small index cards arranged in a matrix (alternatives
[case numbers] x dimensions [case characteristics]) and
required the decision maker to manually turn over the cards
on which information was desired. The information board
used for the case selection decision (burglaries) contained
thirteen elements of information on five cases (see Appendix
I, p. 196, for an illustration of the information board as
presented to the detective sergeants). The information
elements contained within the information board were
consistent with the independent variables used in the
outcome oriented analysis (e.g., sex of victim, physical
evidence present, etc.; in Appendix I, p. 197, the values
for each alternative-dimension pair in the information board
are specified). The information board used for the
prioritization decision in burglaries contained five cases
and fourteen information elements (see Appendix J, p. 198,
for an illustration of the information board as presented to
the burglary detectives). The information elements were
once again consistent with the independent variables
included in the burglary investigation regression analysis
(in Appendix J, p. 199, the values for each alternative-

dimension pair in the information board are specified).

79

Finally, the information board used for the prioritization
decision in robbery investigations contained information on
five cases and sixteen information elements (see Appendix K,
p. 200, for an illustration of the information board as
presented to the robbery detectives). These elements were
once again congruent with those included in the robbery
regression analysis (in Appendix K, p. 201, the values for
each alternative-dimension pair in the information board are
specified). For each information board, the alternative-
dimension values which were specified provided for some
variation yet were typical of the sort of cases the

detectives normally confronted.

Procedure

Each subject attended one session which occurred near
the end of the participant observation period. Subjects
participated in the exercise individually. A small
interview room in the investigations division of the police
department was used for the exercise. At the beginning of
each session, the subject was provided a brief introduction
as to the purpose of the exercise and instructions on how to
proceed. Each sergeant was asked to imagine that he had
five burglary cases to either "OPU" (not assign) or "OPA"
(assign). Each detective was asked to imagine that he was
assigned five cases to work, and he was to determine the

priority each case would receive. Accordingly, at the

80

beginning of the exercise, before searching any information,
the investigator knew only that the cases were either
burglaries or robberies. In order for any other information
concerning the cases (e.g., sex of victim, dollar value of
property loss, etc.) to be disclosed, the cards on which
information was desired had to be turned over by the
investigator. The subjects were told to begin their search
for information with what they considered to be the most
important and to discontinue their search when they felt
they knew enough about the case to make a judgement. The
investigators were free to search for the information in any
way they wished (i.e., within or across alternatives). Each
subject was then familiarized with the information board and
_each information dimension was defined and the possible
values were specified.

Each subject was also instructed to "think aloud" while
reviewing and deciding upon the cases. The subjects were
asked to state the information they were looking at and what
they were thinking while looking at the information. The
subjects were also asked to state any other information not
provided in the information board which would have been
helpful in making their decisions (for example, see Appendix
L for the instructions provided to the detective sergeants
in the selection of burglaries exercise). The verbal
reports provided by each of the subjects were mechanically

recorded (through the use of a tape recorder and with the

81

subject’s consent) and from this, complete transcripts of
the verbal reports were made. The verbal protocols provided
the means by which search behavior, and ultimately decision
strategies and processes, could be identified. As a safety
net for the procedure, the researcher also used pencil and
paper to record the order and content of each subject’s
search. No time constraints were placed on the decision
makers during the exercise. The sessions ranged from
sixteen to forty-four minutes depending on the subject’s
extent of search and verbal activity. The mean amount of
time for the exercise across subjects was approximately 30

minutes.

Research Questions
The following questions were addressed in this part of
the study:

Question #1: It is feasible to collect process data from
detectives through the use of an information
board?

Question #2: What proportion of case information do
detectives search in deciding whether or not to
select a case for an investigation and
prioritizing cases which are assigned for an
investigation? (Depth of search).

Question #3: What case information elements are most often
considered by detectives in deciding whether or
not to select a case for an investigation and
prioritizing cases which are assigned for an
investigation? (Content of search).

Question #4: To what extent do detectives use compensatory
(vs. non-compensatory) strategies in deciding
whether or not to select a case for an
investigation and prioritizing cases which are
assigned for an investigation? (Linearity of

search).
82

Analysis

To address the first question, an overall assessment as
to how the detectives performed in the exercise and the
quality of the verbal reports was made. This issue is
discussed in Chapter Six.

To measure the depth of search, and address the second
question, the specific information elements accessed by each
of the subjects were noted from the verbal protocols. The
number of elements accessed was summed across dimensions and
this number indicated depth of search. When this number was
divided by the total number of information elements
available, the proportion of case information searched was
determined. (Appendix M contains the formula and examples
of calculations for depth of search.)

Content of search was measured by determining which
information elements were accessed in the information board
matrix and the order in which they were accessed. As with
the studies conducted by Payne and Ragsdale (1978) and
Einhorn et al. (1979), the accessed attributes were viewed
as the most important stimuli in the decision task. In
addition, congruent with the instructions provided to each
subject, the elements accessed in the beginning of the
search were viewed as more important than those accessed
later in the search. Accordingly, to determine the
importance of the items searched, an "importance scale” was

created. The first dimension searched in each alternative

83

received a score equal to the total number of dimensions
available in that alternative (n). The second dimension
searched received a score of n-1. The third dimension
searched received a score of n-2, etc. If a dimension was
not accessed in a given alternative, it received a score of
zero. Through this procedure, a mean importance rating was
calculated for each information element within and across
subjects. (Appendix N contains the formula and examples of
computations for content of search.)

To measure linearity in decision making, the procedure
developed by Doherty (1987) and refined by Gilliland (1990)
was used. As discussed in Chapter Two, the examination of
a constant number of dimensions across alternatives suggests
the use of linear strategies and the examination of a
variable number of dimensions across alternatives implies
the use of non-linear strategies (Payne, 1976). With this
realized, the following linearity measure was used. First,
the alternative with the largest number of dimensions
accessed was identified. These dimensions were treated as
the "standard dimensions" by which information search of the
other alternatives was compared. If a tie existed among two
or more alternatives in terms of the number of dimensions
accessed, the standard was the alternative examined first.
When comparing the standard dimensions with the dimensions
accessed on other alternatives, each time a standard

dimension was not examined, a score of one (1) was assigned

84

to that alternative-dimension pair. To determine the extent
of linearity in decision making, the number of one (1),
alternative-dimension scores were tallied and then divided
by the following denominator: ((the number of dimensions
accessed in the standard * the number of alternatives used
in the comparison, including the standard) - (the number of
dimensions in the standard + the number of alternatives used
in the standard - 1)). This index produced coefficients
between zero and one, with zero (0) indicating perfect
linearity and one (1) indicating perfect non-linearity.
(Appendix 0 contains a summary of the linearity index
formula and several computational examples.) A separate
index was calculated on each subject and then a mean

linearity index was calculated across subjects.

The Observation Method

"The researcher must get close to the people he studies;
he understands that their actions are best comprehended when
observed on the spot - in the natural, ongoing environment
where they live and work" (Schatzman & Strauss, 1973, p.
5). Accordingly, for a period of thirty weeks, from
September 1990 to March 1991, investigators were observed
for approximately 370 hours. This constituted a total of
approximately fifty eight-hour shifts with observations
usually taking place during two shifts per week. On some
days however, observations were limited to a few hours on a
given shift.

85

Time was split equally between the crimes against
property squad (burglaries) and the crimes against person
squad (robberies). Robbery detectives were the focus of the
initial observations and then, after a period of about
twelve weeks, observations switched exclusively to burglary
investigations and detectives. For the final seven weeks,
observations of robbery and burglary detectives were made on
an alternating basis. With robberies, time was equally
split between both of the detectives who investigated
robberies. With burglaries, time was spent with three of
the five detectives who investigated burglaries. These
burglary detectives were identified by the sergeant early in
the observation period as the ones who ”didn’t mind having
someone along and would be good to work with.”

The observations usually began during the morning
briefing session when the sergeant assigned cases to the
detectives. During this time a determination would be made
as to who would be observed for the day. This determination
was primarily a function of detective availability. As a
general rule, detectives who were to spend much of the shift
”in court" or engaged in administrative type tasks were
avoided. Most often after the detective read the newly
assigned reports he would provide the researcher with the
reports to review. By about 9:00 a.m., one hour after the
beginning of the shift, most all of the detectives would be

out of the station and ”on the road.”

86

During the observations, the tactic of "tracing" was
used. Schatzman and Strauss (1973) define tracing as when
"the researcher attaches himself to a single person and
follows him about through the entire course of a single
task, or even an entire shift" (p. 41). All of the
activities the detective performed during the course of the
day would be observed. This included suspect
interrogations, street stops, victim interviews, witness
interviews, discussions with other detectives, etc. At no
time was the researcher prohibited from observing any event,
situation, or interaction. During the shift, discussions
(or "informal interviews") also took place with the
detective. Most conversations took place in the detective’s
car while traveling from one point to another but
discussions also took place at the prosecutor’s office
(while waiting for a case to be ”screened” for an arrest
warrant), at restaurants, and at the police station.

The conversations with detectives usually focused on
several related and overlapping issues and were most often
in reference to particular cases. First, how does the
detective view this particular case? What is significant
about this case? Is this case viewed as a particular "type"
of case? If so, what features make this case "fit the mold"
of a "type?” Second, how does the information about the
case (e.g., "case characteristics") guide the conduct of the

investigation? Do certain features of the case make the

87

performance of certain activities more or less likely? Do
certain features of the case make the case more or less
worthy of effort? Or more or less likely to receive
effort? Essentially, what factors determine how the case is
worked? Finally, what meanings are attached to significant
features of the case? What does it mean, for example, if
the case is viewed as "a drug related (type of) case"? By
addressing these questions, insight into the thought process
associated with detective decision making could be
obtained. At the same time, insight could be obtained on
the factors which influenced decision making, which was of
primary concern in the outcome oriented analyses. In
addition to these questions and observations, it was also of
interest to explore the validity of the investigative
reports as a source of data. Accordingly, questions
regarding the process of completing reports and the
detective’s perception of report accuracy were often asked.
On the basis of the observations and questioning,
detailed field notes were written. The notes were usually
written away from the research setting although sometimes
they were written while the detective completed his
paperwork. The notes consisted of several sections which
were congruent with the question categories outlined above.
In addition, a summary of events which occurred during the
shift was written. Notes were most often written in

reference to particular cases that were worked. Along with

88

the raw field notes, a few short draft essays were written
which brought together various observations from the field.
As a result of the field observations, approximately 200
notebook pages of field notes were produced.

The data which were produced from this effort were
intended to compliment the data collected through the other
methods. Hence, these observational data offered a means by
which the other data could be supported or refuted. The
observational data were also used extensively in the
interpretation of the outcome oriented (statistical)
analyses. The observational component of the study also
provided an Opportunity to build rapport with the detectives
thus creating a more favorable atmosphere for the collection

of the verbal data through the information board exercise.

89

Footnotes

However, as seen in Chapter Five, several statistical
problems with the probit analysis required the use of
linear regression to assist in the analysis of the data.

One of the burglary detectives was a patrol officer
temporarily assigned (for two years) to the investigations
division.

The seemingly small number of research subjects (N=10)

is not uncommon when using this methodology. For example,
in the study by Payne (1976), six subjects were used. In
Eihnhorn, Kleinmuntz, and Kleinmuntz (1979), one subject
was used. In Isen and Means (1983), 22 subjects were
used.

90

CHAPTER FIVE

RESULTS

In this chapter the results of the data analyses are
presented. The results are organized into two primary
sections -- decisions as outcomes and decisions as
proceSses. Within the "decisions as processes" section, the
results of the information board exercise and the

observations are presented separately.

Decisions as Outcomes

The following results were obtained from the analysis of
investigative reports completed by patrol officers and
detectives. The results of each of the decisions of

investigators are presented separately.

The Selection of Burglaries
Table 1 presents descriptive data on the independent

(victim and offense) variables and dependent (case selected

91

TABLE 1

INDEPENDENT AND DEPENDENT VARIABLES:
VALUES AND DESCRIPTIVE STATISTICS BY ALL BURGLARIES

#

 

 

Variable Value N %
Victim Type 0=Individual 637 74
1=Business 220 26
Victim Sex 0=Male 330 52
1=Female 307 48
Victim Race 0=White 460 72
1=Non~white 177 28
Victim Employment 0=Not employed 128 28
Status 1=Employed 329 72
Victim Age N8 631
(in years) 2: 40
SD: 14.5
Min/Max: 15-95
Victim Income N: 609
($lyear) X: 16,070
SD: 4,212
Min/Max: 4,212-31,672
Victim-
Offender
Relationship 0=No 604 71
Present 1=Yes 244 29
Strength of 0=Weak 678 79
Suspect Info 1=Moderate 113 13
2=Strong 66 8
Physical 0=No 694 81
Evidence 1=Yes' 163 19
Suspect Vehicle 0=No 825 96
Described 1=Yes 32 4
Suspect Vehicle 0=No 850 99
his Plate Known 1=Yes 7 1
Stolen Property 0=No 722 85
Identifiable 1=Yes 131 15
Weapon Used 0=No e O
1=Yes Q 9
Degree of 0=No injury e C
Injury 1=Minor Injury e O
2=Ser Injury 9 0
Value Stolen N: 802
Property X"= 1,207
(in dollars) SD: 2,199
Min/Max: 0-30,000
Case Selected
For Follow-up 0=No 540 63
Investigation 1=Yes 317 37

 

# missing data are excluded from table
e variable not appropriate for burglaries

92

for investigation) variable for all of the burglary cases (N
= 857). Table 1 also reflects the coding scheme used in the
bivariate and multivariate analyses.

As seen in Table 1, 637 of the 857 burglary cases (74%)
involved individuals as victims while 220 (26%) involved
businesses. A slight majority of the cases with individuals
as victims, 330 of 637 (52%), involved male victims. In
approximately three-quarters of the cases, 460 of 637 (72%),
the victims were white. The same percentage, 329 of 457
(72%), involved victims who were employed. The mean age of
burglary victims was 40 years old and their mean "income”
was $16,070. In 244 of the 848 cases (29%), the victim had
some sort of relationship with the suspected offender.

The vast majority of cases, 678 of 857 (79%), contained
weak suspect information; 113 of 857 (13%) contained suspect
information of moderate strength; and 66 of 857 (8%)
contained strong suspect information. Of the 678 cases
which contained weak information, 491 (72%) had ”no
significant suspect information.” Most cases, 694 of 857
(81%), did not have physical evidence available. It is also
seen that there is little variance in the vehicle
information evidence variables. In only 32 of 857 cases
(4%) was a vehicle described and in only 7 of the 857 cases
(1%) was a vehicle license plate known. The great majority
of cases, 722 of 853 (85%), involved property which was not

identifiable. The mean value of the stolen property was

93

$1,207 with a range from zero to $30,000. Finally, it is
seen in Table 1 that 317 out of the 857 burglary (37%) cases
brought to the attention of the crimes against property
squad were selected for a follow-up investigation.1

Table 2 contains the correlation coefficients (Pearson’s
r) among the independent variables and the dependent
variable. As for multicollearity among the independent
variables, at least as evident through simple associations,
there appears to be relatively little. The strongest
associations are .46, .32, and .25 with the remainder at or
below .17.

Concerning the variables related to whether or not the
case was selected for an investigation, strength of suspect
information and the presence of a victim-offender
relationship are the strongest (.55 and .51, respectively).
Although not tabled, bivariate analyses indicated that 156
of the 678 cases (23%) which contained weak suspect
information were selected for a follow-up; 96 of 113 cases
(85%) which contained moderate suspect information were
selected; and 66 out of 66 cases (100%) which contained
strong suspect information were selected for a follow-up
investigation. This pattern clearly illustrates that the
stronger the suspect information, the more likely the case
was to be selected for a follow-up investigation. As for
the victim-offender relationship, 127 out of the 604 cases

(21%) without a victim-offender relationship were selected

94

 

:Owuomfiuuo>sH you omuuuaum menu u «a mwzncofluoamm HoosomuoIfiwuuH> u b
muummoum smaoum no 05Ho> m n ma ufioosH fiwuofi> u o
maneeueucmcH muuoaoum n ma mm< smuoa> u m
csoem enema mausnm> u as museum ucosaoaasm asuos> u e
confluence eaofinu> u 0H comm Ewuuw> n n

cosmofi>m Houseman u a wow afiuufi> u m
:OfluosuousH poemmom u 0 came awuuw> u H
ouusmaou on possno usmfiofluumoo “sumo mswmmfle *

00.H H0. N0.I NH. ha. «0. mm. am. 50.: 00.I N0. 00. dd. 00.! ea
00.H mo. 00. H0. ma. HH.I 0a.! mm. AH. 00. H0.I 00.! m0. ma
00.H no. mo. m0.l m0.l 0a.: 00.1 H0. 50. H0.I H0.I n0.l NH

00.H we. H0.I ma. m0. m0. #0. H0. m0.l m0. m0.l Ha

00.H 00. NH. 00. 00. mo. 00. no. mo. 50.: 0a

00.H 00.1 no.1 mo. ea. NH. 0H.i No.1 «0. m

00.H an. 0a.! NH.I >0.I ma. NH. HH.I w

00.H 0a.! bH.I 00.! 0H. 0H. MH.I h

00.H ha. 00. 0H.I H0. N0. 0

00.H 00.! HH.I 50.I * m

00.H 00.: m0.l m0. e

00.H m0. * m

00.H * N

00.H H

 

0H

0H

NH HH 0H m m h m

 

Ahmmflzv mmHm¢AUmDm and mom mmgdem4> UZOS<

mBZWHUHhmHOU 20Hadflmmm00

N mdmdfi

95

for a follow—up while 183 out of 244 cases (75%) with a
victim-offender relationship were selected for a follow-up
investigation. Therefore, with a relationship between the
victim and offender, the case was more likely to get
assigned. It is of some consequence to note however, that
the relationship between the suspect information variable
and the victim-offender relationship variable is .32; when
there was a relationship between the victim and offender, a
name of the suspect was usually provided. None of the other
correlation coefficients between the independent variables
and the dependent variable are of appreciable strength.

Because it was of primary interest to isolate the
relative influence of the victim and offense variables on
the case selection decision, multivariate statistical
procedures were used. In Table 3, the coefficients,
standard errors, t-values, and standardized weights from
probit and OLS regression analyses are presented.

Although it is generally considered inappropriate to
conduct OLS regression analyses on a dichotomous dependent
variable, two anomalies in the probit results make
consideration of OLS regression analyses necessary. First,
the derivative at the meanz for strength of suspect
information is quite small (.05) yet significant at the p <
.01 level. Second, the derivative at the mean for suspect
vehicle plate known is extremely large (.72) but not

significant (p = .989). It appears that these seemingly

96

TABLE 3

PROBIT (AND MULTIPLE REGRESSION) ANALYSIS OF
BURGLARY CASE SELECTION
AS A FUNCTION OF VICTIM AND OFFENSE CHARACTERISTICS
(regression results in parentheses)

 

 

 

Independent Standard derivative
Variables Coefficient Error t (beta)
Victim Type -.033 .361 -7.622 .00
( .012) (.028) ( .406) .01)
Victim Sex .054 .125 .429 .02
(-0013) (0029) ( -0466) (-001)
Victim Race .136 .142 .955 .05
(‘0036) (.032) (-10125) (‘003)
Victim Emp Status .293 .170 1.721 .11*
( .053) (.038) ( 1.398) .04)
Victim Age -0004 0004 -0870 .00
( .000) (.001) ( .345) .01)
ViCtlm Income 0000 0000 -0497 .00
( .000) (0000) ( -0019) (-002)
Victim-Off. Rel .489 .069 7.089 .19**
( .423) (.029) (14.396) .39**)
Suspect Information 1.750 .152 11.495 .05**
( .354) (.022) (16.020) .44**)
Physical Evidence .342 .130 2.625 .13*‘
( .098) (.032) ( 3.075) .08*#)
Vehicle Described .931 .310 3.000 .36**
( .298) (.073) ( 4.081) .12*')
Vehicle Plate Known 5.170 386.788 .013 .72
(-.015) (.154) ( -.098) .00)
property Id’able -.040 .089' -.452 -.01
( .045) (.034) ( 1.304) ( .03)
$ Value of Property .000 .000 -1.459 .00
( .000) (.000) ( 3.095) ( .08**)
2
Pseudo R 8 .32
2
(Adjusted R I .46)
Significance = .00
(Significance = .00)
1
N = 857
1
(N = 857)

 

* p<.05; ** p<.01 (one-tailed test)
1 the mean of each variable was substituted for missing data

97

unreliable coefficients are at least a partial result of
multicollinearity.3 As a result of these anomalous probit
results, the attractiveness of the multiple regression
analyses increases. In short, it appears that given the
nature of these data, the regression results (beta weights)
are less biased than the probit results (derivative at the
means).

It is of comfort to note that the probit and regression
analyses are quite similar in terms of those factors which
display a significant impact on the case selection
decision.4 As seen in Table 3, both sets of results
indicate that strength of suspect information, the presence
of a victim—offender relationship, description of a
suspect’s vehicle, and availability of physical evidence
exert a significant impact on the case selection decision.
The only inconsistencies in the results are that the dollar
value of the property loss is significant in the regression
analysis but not in the probit analysis while employment
status of the victim is significant in the probit analysis
but not in the regression.

According to the regression results, the largest impact
on the case selection decision is exerted by the strength of
suspect information; the stronger the suspect information,
the more likely the case was selected for an investigation
(b = .44; p < .01). Presence of a relationship between the
victim and offender also exerts a significant influence;

cases where a relationship existed were more likely to be

98

selected (b = .39; p < .01). The variable with the third

strongest impact was a description of the suspect’s vehicle;

when this information was known, the case was more likely to
be selected (b = .12; p < .01). The final two variables
which display a statistically significant effect on case
selection are presence of physical evidence and the dollar
value of the stolen property. When physical evidence was
available or when more value loss was involved, the case was
more likely to get selected (b = .08; p (.01 for each).

On the basis of these analyses, the following hypotheses
are supported:

Hypothesis 7: Cases with a higher value of stolen property
are more likely than cases with lesser value
to be selected for an investigation.

Hypothesis 9a: Cases with stronger suspect information are
more likely than cases with weaker suspect
information to be selected for an
investigation.

Hypothesis 9c: Cases with physical evidence are more likely
than cases without physical evidence to be
selected for an investigation.

Hypothesis 9d: Cases with a suspect vehicle description are
more likely than cases without a vehicle
description to be selected for an
investigation.

In addition, the opposite of the following hypothesis is

supported:

Hypothesis 4: Victims who do not have a relationship with
the suspect are more likely than victims who
do have a relationship with the suspect to

have their cases selected for an
investigation.

99

Time Allocation in Burglary Investigations

Table 4 presents the frequency distribution for the
independent variables (victim and offense) and dependent
variable (time spent on the follow—up investigation) on the
subset of burglary cases which were selected for a follow-up
investigation (N = 317). Table 4 also reflects the coding
scheme used in the bivariate and multivariate analyses.

As illustrated in Table 4, 252 of the 317 burglary cases
(80%) which received a follow-up investigation involved
victims who were individuals, while 65 of the 317 cases
(20%) involved businesses. A slight majority of the
selected cases, 138 of 252 (55%), involved female crime
victims. Most of the cases, 169 of 252 (67%), involved
victims who were white, and employed 126 of 172 (73%). The
mean age for the burglary victims who had their cases
selected was 36 years and the mean "income” was $15,687. In
45 of 308 cases (15%), the victim did not desire
investigative effort. A relationship between the victim and
the suspected offender was present in 184 of 310 cases (59%)
selected for an investigation.

It is also evident from Table 4 that while 96 of the 317
selected cases (30%) contained moderate suspect information,
221 of 317 (70%) cases contained weak or strong suspect
information. Specifically, 155 of the 317 cases (49%)
contained weak suspect information while 66 of 317 (21%)

contained strong information. Physical evidence was

100

TABLE 4

INDEPENDENT AND DEPENDENT VARIABLES:

VALUES AND DESCRIPTIVE STATISTICS BY "SELECTED" BURGLARIES

#

 

 

Variable Value N %
Victim Type 0=Individual 252 80
1=Business 65 20
Victim Sex 0=Male 114 45
1=Female 138 55
Victim Race 0=White 169 67
1=Non—white 83 33
Victim Employment 0=Not employed 46 27
Status 1=Employed 126 73
Victim Age N: 249
(in years) 2: 36
SD: 13.8
Min/Max: 17-95
Victim Income N: 247
(S/year) X: 15,687
SD: 4,785
Min/Max= 7,260-31,672
Victim Desires 0=No 45 15
Effort 1=Yes 263 85
Victim-
Offender
Relationship 0=No 126 41
Present 1=Yes 184 59
Strength of 0=Weak/Strong 221 70
Suspect Info 1=Moderate 96 30
Physical 0=No 250 79
Evidence 1=Yes 67 21
Suspect Vehicle 0=No 292 92
Described 1=Yes 25 8
Suspect Vehicle 0=No 310 98
Lic Plate Known 1=Yes 7 2
Stolen Property 0=No 270 85
Identifiable 1=Yes 46 15
Weapon Used 0=No Q Q
1=Yes C 9
Degree of 0=No injury 9 Q
Injury 1=Minor Injury 0 C
2=Serious Injury 0 9
Value of Stolen N: 317
Property X= 19243
(in dollars) SD= 2,680
Min/Max: 0-30,000
Time spent N: 317
on Follow-up X: 3.7
Investigation SD: 4.6
(in hours) Min/Max= .1-52.5

 

# missing data are excluded from table
9 variable not appropriate for burglaries

available in 67 of the 317 cases (21%). In 25 of the 317
selected cases (8%), a suspect’s vehicle was described and
in 7 of 317 cases (2%) a license plate number was known. In
270 of 316 cases (85%) the stolen property was not
identifiable. The mean value of the stolen property was
$1,243 with a range of zero to $30,000. Finally, the mean
amount of time spent on burglary follow-up investigations
was 3.7 hours with a range of .1 hours to 52.5 hours.

To gain an understanding of how time was spent on
investigations, the activities performed in each
investigation were recorded from the follow-up reports.
Table 5 contains these findings. Briefly, it is seen that
the most common activity performed in burglary follow-up
investigations was interviewing victims. In 213 of the 317
investigations (67%), the victim was interviewed at least
once. The second most frequently performed activity was
interviewing others who were not directly involved in the
crime in question (e.g., mother of suspect, pawn shop
personnel, parole officer, etc.); in 126 of 317 cases (40%),
this activity was performed. The third most common activity
was interviewing suspects; in 116 of the 317 cases (37%) a
suspect was interviewed at least once. The remainder of the
activities, rank-ordered in frequency, are: searched
computer files (23%), consulted prosecutor (21%),
interviewed witnesses (20%), submitted physical evidence

(15%), searched crime scene (6%), conducted photo line-up

102

TABLE 5

FREQUENCY OF DETECTIVE ACTIVITIES
IN BURGLARY FOLLOW-UP INVESTIGATIONS (N=317)

 

Investigations in Which

 

 

Activity Activity was Performed

H 1
Victim Interviewed 213 67
Others Interviewed 126 40
Suspect Interviewed 116 37
Computer Files Searched 72 23
Prosecutor Consulted 67 21
Witness(es) Interviewed 62 20
Physical Evidence Submitted 47 15
Crime Scene Searched 18 6
Photo Line-up Conducted 16 5
Witness Canvass Conducted 15 5
Informants Interviewed 7 2
Physical Line-up Conducted 6 2
Mug-shot Books Shown 4 1

 

103

(5%), canvassed for witnesses (5%), interviewed informants
(2%), conducted physical line-up (2%), and showed mug shot
books (1%).

Table 6 presents the correlation coefficients (r) among
the independent variables and the dependent variable for
those burglaries selected for a follow-up investigation. It
is seen that several of the coefficients between the
independent variables are of at least moderate strength
(.51, .34, .32), however, none appear to approach
”dangerous" proportions in terms of collinearity.

As for the variables related to the time spent in the
follow—up investigation, the dollar value of the stolen
property is the strongest (.29). The higher the value of
the stolen property, the more likely more time was spent on
the investigation. Although not displayed, a cross-
tabulation procedure showed that 36 of 163 cases (22%) which
involved under $300 of stolen property received over 3.7
hours (the mean) of investigative time. Twenty-seven of 82
cases (33%) which involved $300 to $1,300 of stolen property
received over 3.7 hours. Twenty-nine of 72 cases (40%)
which had over $1,300 of property taken received over 3.7
hours of investigative time.

Three variables, suspect information, physical evidence,
and victim income, are all positively related to the amount
of time spent (r = .21). With suspect information, a

crosstab procedure demonstrated that 34 of 155 cases (22%)

104

 

 

mHnmsoHumHom muosuuuoIEHuoH> n m
coHuomHumm>sH no woman mEHB u mH uuomum mouHmea EHHOH> u 5
muwumoum suHoum mo msHo> m u eH eaoocH aHuoH> u 0
OHQCHHHDGOCH muhumoum n nH mom EHHOH> n m
axons mumHa mHoHem> n ma museum ucmsaoHasm sHuoH> u e
omnHHomuo OHOHno> u HH doom BHHOH> u m
mocmeH>m Hmonmna u oH xmm sHuoH> u m
soHuoEuoucH poemmsm u m 0959 EHHOH> u H
ouusmsoo on poscoo ucoHOHuumoo «sumo ocHumHE ¢
00.H mm. H0.: 0N. 0H. Hm. Hm. 0H.: wH. Hm. 0H. 00. 0H.: 00.: 50. mH
00.H H0.: 00. 00. mm. «0.: 5H.: HH. on. em. 00. 00.: m0.: 00. 0H
00.H 00. H0. #0.: 00. eH.: 50. «0.: N0. NH. H0.: o0.: n0.: MH
00.H Hm. m0.: H0.: m0.: 00. «H. 00. H0. 50.: 00. 00.: NH
00.H H0.: H0. 0H.: N0.: NH. NH. mo. e0. 00. 00.: HH
00.H m0. 0H.: 5H. mH. Hm. 0H. vH.I m0.: 00. 0H
00.H 0m.: mH. H0. 00. N0. 00.: 5H.: HH. 0
00.H NN.: 0H.: HN.I 0H.: 0N. oH. MH.: 0
00.H mH. 5H. mH. 0H.: wH.: 0H. 5
00.H mH. HH. NH.: N0. m0. 0
00.H 00.: m0.: MH.: * m
00.H 50.: HN.I m0. e
00.H m0. 5 m
00.H * N
00.H H
mH vH nH NH HH 0H m 5 o m e m m H

 

A5Hmflzv WNHM4AUmDm somfiumqmme mom mmflm<Hm¢> UZOZG
mBzmHUHmmmOU ZOHB<ANMMOU

0 mflmdfi

105

with weak suspect information received more than 3.7 hours
of time. Similarly, in cases which contained strong suspect
information, 14 of 66 (21%) received over 3.7 hours of

time. However, in cases which contained moderate suspect
information, 44 of 96 (46%) received more than 3.7 hours of
investigative time. In addition, when suspect information
is treated as three distinct categories and then correlated
with the amount of time spent on the investigation (in raw
form) the correlation coefficient drops to .03 from .21.
With this evidence, there is at least initial support for
the theoretical expectation that cases with moderate suspect
information receive more time than those cases with weak or
strong suspect information. More importantly, there is
justification for combining weak and strong suspect
information into one analytic category for the multivariate
analyses.

To assess the relative impact of the independent
variables on the dependent variable, an OLS multiple
regression procedure was used. Table 7 contains the results
of these analyses. Although the primary intent of these
analyses was to test a limited set of hypotheses which
predict the amount of time spent in an investigation, it is
of interest to note that 17 percent of the variation in time
spent on burglary investigations is accounted for by victim
and offense characteristics.

As seen in Table 7, the greatest impact on time spent is

106

TIME SPENT ON

AS A FUNCTION

TABLE 7

MULTIPLE REGRESSION OF
BURGLARY FOLLOW-UP INVESTIGATIONS
OF VICTIM AND OFFENSE CHARACTERISTICS

 

 

 

Independent Standardized

Variables Beta
Victim Type .03
Victim Sex -.04
Victim Race -.07
Victim Employment Status -.06
Victim Age .06
Victim Income .05
Victim Desires Effort .08
Victim-Offender Relationship .00
Suspect Information .19**
Physical Evidence .09*
Vehicle Described .00
Vehicle Plate Known .19**
Property Identifiable -.03
$ Value of Stolen Property .22**

Multiple R .46
2
Adjusted R .17
F 5.73
Significance .00
1
N 317

 

* p<.05; ** p<.01 (one-tailed test)
1 the mean of each variable was substituted for missing data

107

exerted by the dollar value of the stolen property; cases

with greater loss were more likely to have more time devoted

to them (b= .22; p < .01). Strength of suspect information
also has a significant impact -- cases with moderate
information were more likely to receive more time than cases

with weak or strong suspect information (b = .19; p < .01).

The only other variables which display a statistically

significant effect on time spent are knowledge of the

suspect’s vehicle plate number (b = .19; p < .01) and
physical evidence (b = .09; p < .05). When a plate number
was known or when physical evidence was present, more time
was likely spent on the investigation.

On the basis of these analyses, the following hypotheses
are supported:

Hypotheses 7: Cases with a higher value of stolen property
are likely to have more time spent on the
investigation than cases with lesser value.

Hypothesis 9b: Cases with suspect information of moderate
strength are likely to have more time spent
on the investigation than cases with weak or
strong suspect information.

Hypothesis 90: Cases with physical evidence are likely to
have more time spent on the investigation
than cases without physical evidence.

Hypothesis 9e: Cases with a suspect vehicle plate known are
likely to have more time spent on the

investigation than cases without a vehicle
plate known.

108

Time Allocation in Robbery Investigations

In Table 8, descriptive data are presented on the
independent variables and dependent variable for the
robberies which were selected for a follow-up investigation
(N = 292). Also reflected in Table 8 is the coding scheme
used in the bivariate and multivariate analyses.

Because bank robberies are unique from other robberies
in that they often involve very large sums of money and
necessitate the involvement of other law enforcement
agencies (i.e., F.B.I.), it was of interest to explore the
impact of bank robberies (N=9) on the overall distribution
of scores. To do so, analyses were conducted on two sets of
robbery investigations: robberies with bank robberies
included and robberies with bank robberies excluded. It is
apparent in Table 8 that the nine bank robberies affect the
distribution of scores in two important, but not surprising,
ways: value of stolen property and the amount of time spent.
in the investigation. First, in the inclusive robbery
category, the value loss ranges from zero to $9,132 with a
mean value loss of $260. With the exclusion of bank
robberies, the loss ranges from zero to $4,000 and the mean
amount of loss drops to $173.5 Second, when bank
robberies are included, the mean amount of time spent on
robbery follow-up investigations is 4.5 hours with a range
of .4 to 50.8 hours. When bank robberies are excluded, the

mean amount of time spent on an investigation decreases

109

TABLE 8

INDEPENDENT AND DEPENDENT VARIABLES: 3

VALUES AND DESCRIPTIVE STATISTICS BY "SELECTED" ROBBERIES

 

 

Variable Value Robberies Robberies
w/bank w/o bank
N % N %
Victim Type 0=Individual 189 65 189 67
1=Business 103 35 94 33
Victim Sex 0=Male 121 65 121 65
1=Female 66 35 66 35
Victim Race 0=White 132 71 132 71
1=Non-white 54 29 54 29
Victim Employment 0=Not employed 77 50 77 50
Status 1=Employed 78 50 78 50
Victim Age N= 186 186
(in years) X= 31 31
SD: 12.4 12.4
Min/Maxa 12-75 12-75
Victim Income N: 145 145
(s/year) X‘= 14,940 . 14,940
SD: 5,163 5,163
Min/Max= 7,260-37,238 7,260-37,238
Victim Desires 0=No 64 24 64 25
Effort 1=Yes 200 76 191 75
Victim-
Offender
Relationship 0=No 260 90 251 89
Present ltYes 30 10 30 11
Strength of 0=Weak/Strong 103 35 102 36
Suspect Info 1=Moderate 189 65 181 64
Physical 0=No 258 78 256 90
Evidence 1=Yes 34 12 27 10
Suspect Vehicle 0=No 249 85 242 86
Described 1=Yes 43 15 41 14
Suspect Vehicle 0=No 275 94 266 94
Lic Plate Known 1=Yes 17 6 17 6
Stolen Property 0=No 285 98 278 98
Identifiable IIYes 7 2 5 2
Weapon Used 0=No 109 37 108 38
1=Yes 182 63 174 62
Degree of 0=No injury 191 66 182 65
Injury 1=Minor Injury 58 20 58 21
2=Serious Injury 40 14 40 14
Value of Stolen N8 278 272
property 2: 260 173
(in dollars) SD= 855 387
Min/Max: 0-9,132 0-4,000
Time spent N: 279 270
on Follow-up X= 4.5 3.7
Investigation SD: 5.5 2.9
(in hours) Min/Max= .4-50.8 . -16.0

 

 

# missing data are excluded from table

6
to 3.7 hours with a range of .4 to 16.0 hours.

As for the distribution of the values for other
variables (as presented within the inclusive category), it
is seen that in the majority of cases, 121 of 187 (65%), the
victims were male. In 132 of 186 cases (71%) the victims
were white. Seventy-seven of 155 cases (50%) involved
victims who were employed. The mean age of robbery victims
was 31 years and their mean ”income" was $14,940. In 64 of
264 cases (24%) the victim did not desire investigative
effort. In 30 of 290 cases (10%) the victim had some sort
of relationship with the suspected offender.

Most of the cases, 189 of the 292 (65%), contained
moderate suspect information while 103 of the 292 cases
(35%) contained weak or strong suspect information.
Specifically, 68 of the 292 cases (23%) contained weak
information while in 35 of 292 cases (12%) the information
was strong. Physical evidence was available in 34 of the
292 cases (12%). In 43 of the 292 cases (15%) a suspect’s
vehicle description was available and in 17 of the 292 cases
(6%) a suspect’s vehicle license plate was known. In only 7
of the 292 cases (2%) was stolen property identifiable.
Weapons were used in 182 of 291 cases (63%). The majority
of cases, 191 of 289 (66%), involved no injury to the
victim; 58 of 289 (20%) involved minor injury; and 40 of 289
(14%) involved serious injury.

As with burglary investigations, it was of interest to

111

highlight the activities performed most consistently in
robbery follow-up investigations. Table 9 contains these
findings. As seen, the most common activity performed was
interviewing victims. In 245 of the 292 cases (84%), the
victim was interviewed at least once. The second most
common activity was interviewing others; in 106 of the 292
cases (36%) this activity was performed. In 82 of the 292
cases (28%) a photo line-up was conducted -- the third most
common activity. The remainder of activities, rank ordered
in frequency, are: searched computer files (26%),
interviewed witnesses (23%), consulted prosecutor (23%),
interviewed suspect (19%), showed mug shot books (19%),
canvassed for witnesses (14%), searched crime scene (11%),
interviewed informants (11%), submitted physical evidence
(10%), and conducted physical line-up (10%).

Table 10 contains the correlation coefficients (r)
between all variables for those robberies selected for a
follow-up investigation. In looking for collinearity among
the independent variables, it is seen that there are several
potential problematic associations (.60, .44, .22) but such
associations do not appear to be pervasive.

Concerning the correlations between the independent
variables and time spent on the follow-up investigations, it
is seen that the dollar value of the stolen property has the
strongest association (.62). The higher the value of the

stolen property, the more likely more time was spent on the

112

TABLE 9

FREQUENCY OF DETECTIVE ACTIVITIES
IN ROBBERY FOLLOW-UP INVESTIGATIONS (N=292)

 

Investigations in Which

 

 

Activity Activity was Performed

N 1
Victim Interviewed 245 84
Others Interviewed 106 36
Photo Line-up Conducted 82 28
Computer Files Searched 76 26
Witness(es) Interviewed 68 23
Prosecutor Consulted 66 23
Suspect Interviewed 55 19
Mug-Shot Books Shown 54 19
Witness Canvass Conducted 40 14
Crime Scene Searched 31 ' 11
Informants Interviewed 31 11
Physical Evidence Submitted 30 10
Physical Line-up Conducted 30 10

 

113

 

 

Amazon :Hv SSOCM uuon OHOHnu> u NH uaoocH EHDOH> u 0
:OHvomHumu>:H so ucwmm OEHB u 5H COnHuommn OHOHnm> n HH mom EHHOH> n m
auummoum smHoum no msHo> 0 n 0H ousuoH>m HouHmmam n 0H .uoum .mem EHDOH> u e
muonsH no common u 0H COHuosuOmsH uuommsm u 0 0000 EHuOH> n m
Coma somouz u «H QHnmsOanHum MOCSOMHOISHDOH> n 0 x00 sHuOH> u N
anoHuHusooH muummowm u mH vacuum How emHmoo EHDOH> u 5 OQ59 EHDOH> u H
omuomsoo on #0: CHSOO usmHoHuuooo “sumo msHmmHE 5
00.H N0. 5H.: 0H.: HN. 00. 00. 5e. 0N. 0H.: 0N. 00.: 00. 00. 00.: 00. 0N. 5H
00.H 00.: 00.: «N. 00.: H0.: 0N. no. 00.: 0H. N0. NH. N0. 0H. H0.: 00. 0H
00.H NN. «0. 00.: «H.: HH.: 0N.: 00. 00.: 00. «0.: 00. 00. 0H.: mn.: 0H
00.H 00. 50. NH.: 0H.: 5H.: HN. 0H.: NH. <0. 0H.: 00.: 0H. NN.: «H
00.H 00.: 00. NN. N0. N0. 00. NN. H0.: 00.: 00.: N0. 50.: MH
00.H 00. e0.: 00. e0.: «0.: H0. mH.: 00. 0H. 0H. H0. NH
00.H no. NH. H0.: H0. 00. NH.: N0.: 0H. 00. H0. HH
00.H HH. NH.: 0H. H0. H0. 50. NH.: 00. 0N. 0H
00.H «0.: N0.: 0H.: 00.: 0H. 00.: 00.: 0N. m
00.H 0H.: H0. 00.: 5H.: HH. 0H. 0N.: 0
00.H N0. 0H. 0H. 00.: 00. MN. 5
00.H 50.: HH. .00.: 0H. 5 0
00.H e0. H0.: HH. 5 m
00.H HN.: 0H. 5 e
00.H 50.: e m
00.H 5 N
00.H H
5H 0H 0H 0H mH NH HH 0H m 0 5 0 m e m N H

 

Ammmflzv mmHmmmmom somBUMAflw: mom mmqm<Hm¢> Uzozd
wBZWHUHhmmOU ZOHB<AWMMOU

0H mqmdfi

114

investigation. Results from a cross-tabulation (not
illustrated) showed that 54 of 179 cases (30%) which
involved stolen property of under $100 received more than
4.5 hours of investigative time (the mean amount of time
spent). Twenty-eight of 80 cases (35%) which involved
losses ranging from $100 to $650 received more than 4.5
hours of time. And 21 of 33 cases (64%) which involved over
$651 of stolen property received over 4.5 hours of time
spent on the follow-up investigation. Five other variables
had at least a slight relationship with the amount of time
spent: availability of physical evidence (.47), desire for
effort (.28), victim type (.25), identifiability of stolen
property (.21), and suspect information (.20).

In regard to suspect information, a cross-tab procedure
showed that 11 of 68 robbery cases (16%) with weak
information received more than 4.5 hours of investigative
time. Similarly, of the 35 cases which contained strong
suspect information, six (17%) received over 4.5 hours of
time. However, in cases which contained moderate suspect
information, 87 of 189 (46%) received more than 4.5 hours of
investigative time. In addition, when suspect information
is treated as three distinct categories and then correlated
with the amount of time spent on the investigation (raw
form), the correlation coefficient drops to .03 from .20.
As with burglaries, this evidence provides at least initial

support for the hypothesis that cases with moderate suspect

115

information receive more time than cases with weak or strong
suspect information. More importantly, this evidence is
justification for combining weak and strong suspect
information into one analytic category for the multiple
regression procedure.

To determine the relative impact of the victim and
offense variables on the amount of time spent on the
investigation, and hence test the hypotheses stated in
Chapter Four, OLS multiple regression was used. These
results are presented in Table 11.

As seen in Table 11, analyses on robberies with bank
robberies included are very similar to the analyses with
bank robberies excluded with one important exception -- the
influence of victim type. When bank robberies are included,
victim type exerts a significant impact on the amount of
time spent on an investigation; "business" robberies were
more likely to receive more investigative time (b = .14; p <
.01). However, when bank robberies are excluded from the
analysis, the impact of victim type disappears (b = .02; p >
.05).

In robbery investigations, the dollar value of the
stolen property once again exerts by far the most influence
on the amount of time spent on an investigation; greater
property loss led to the expenditure of more time even with
bank robberies (which account for much of the variance in

property loss) excluded (b = .50; p < .01 with bank

116

TABLE 11

MULTIPLE REGRESSION OF
TIME SPENT ON ROBBERY FOLLOW-UP INVESTIGATIONS
AS A FUNCTION OF VICTIM AND OFFENSE CHARACTERISTICS

 

Standardized Beta

 

 

 

Independent
Variables Robberies Robberies
w/bank w/o bank
Victim Type .14** .02
Victim Sex .00 .01
Victim Race -.03 -.02
Victim Employment Status -.02 -.03
Vietim Age -.02 -002
Victim Income -.02 -.01
Victim Desires Effort .16** .19**
Victim-Offender Relationship .07 .06
Suspect Information .14** .15**
Physical Evidence .22** .25**
Vehicle Described .05 .04
Vehicle Plate Known .00 .00
Property Identifiable .03 .02
Weapon Used .04 .05
Degree of Injury -.01 .03
$ Value of Property Loss .49** .50**
Multiple R .72 .71
2
Adjusted R .49 .48
F 18.78 17.82
Significance .00 .00
1
N 292 292

 

4* p<.01 (one-tailed test)
1 the mean of each variable was substituted for missing data

117

robberies excluded and b = .49; p < .01 with bank robberies

included). The availability of physical evidence also

displays a significant impact on time spent; when physical
evidence was available, more time was spent on the

investigation (b = .22 inclusive; b = .25 exclusive; p < .01

for both). Another significant contribution is made by the

victim’s desire for an investigation. If the victim did not
wish an investigation, less time was spent on the

investigation (b = .16 inclusive; b = .19 exclusive; p < .01

for both). The strength of suspect information also

influences the amount of time spent on the case. Cases with
moderate suspect information were likely to receive more
time than cases with weak or strong suspect information (b =

.14 inclusive; b = .15 exclusive; p < .01 for both).

Finally, in terms of the variance explained by the victim

and offense variables, the model did quite well accounting

for 49 percent in the inclusive category and 50 percent in
the exclusive category.

On the basis of these analyses, the following hypotheses
are supported:

Hypothesis 4: Businesses are likely to have more time spent
on their investigation than non-
businesses (only when bank robberies are
included).

Hypothesis 5: Victims who desire effort are likely to have
more time spent on their investigationthan
victims who do not desire effort.

Hypothesis 7: Cases with a higher value of stolen property

are likely to have more time spent on the
investigationthan cases with lesser value.

118

Hypothesis 9b: Cases with suspect information of moderate
strength are likely to have more time spent
on the investigation than cases with weak or
strong suspect information.

Hypothesis 9c: Cases with physical evidence are likely to

have more time spent on the investigation
than cases without physical evidence.

Summary

By considering the findings across the three analyses,
one can identify several patterns. First, none of the
victim demographic variables (i.e., age, race, sex, income,
employment status) displayed significance (except, of
course, victim type in robberies with bank robberies
included). Second, the dollar value of the stolen property
exerted an impact on all three decisions, and was the most
influential in both of the time allocations decisions.
Third, strength of suspect information displayed a
significant influence across all three analyses. Fourth,
the presence of physical evidence had a significant impact
on all of the decisions. Fifth, identifiability of the
stolen property did not have an effect on any of the
decisions. Finally, victim desire for an investigation, the
presence of a victim-offender relationship, knowledge of the
suspect’s vehicle license number, and vehicle description

displayed inconsistent effects across decisions.

119

Decisions as Processes

Information Board / Verbal Protocol Analysis

The following data were derived from the analysis of
information board search patterns and the verbal protocols.
The results for each investigative decision are presented

separately.

The Selection of Burglaries

To identify the amount of information searched (depth of
search) in deciding whether or not to assign burglary cases
to detectives, the proportion of thirteen case information
elements accessed was calculated for each alternative and
then across alternatives and sergeants. In Table 12 it is
seen that the mean amount of information searched across
alternatives and detectives was 46 percent. There exists
substantial variance among subjects with Detective Sergeant
A searching, on average, 26 percent of the available
information and Sergeant B searching 57 percent. An
examination of the variability of search within subjects and
across alternatives shows that the depth of search is
generally quite similar with relatively little variation
from the mean.

Table 13 presents data on the importance7 of case

information in the decision of whether or not to select a

burglary case for a follow-up investigation. With the

120

TABLE 12

PROPORTION OF AVAILABLE INFORMATION SEARCHED
IN THE BURGLARY SELECTION DECISION

 

Detective Sergeant
Case Alternatives

 

 

A B C Mean
1 .31 .62 .46 .46
2 .23 .62 .38 .41
3 .23 .46 .62 .46
4 .23 .62 .69 .51
5 .31 .54 .62 .49
Mean .26 .57 .55 .46

 

121

TABLE 13

THE IMPORTANCE OF CASE INFORMATION
IN THE BURGLARY SELECTION DECISION

 

Detective Sergeant

Case Information

 

 

Victim Type

Victim Sex

Victim Race

Victim Employment Status
Victim Age

Victim Address
Victim-Offender Relationship

Suspect Information
Physical Evidence

Vehicle Described

Vehicle Plate Known
Property Identifiable

$ Value of Stolen Property

HH
ONONHN OOOOOOU

OOOOOO OOOOOOO

HHH
O’CDCDONN NHOOOOO}

“Gab-505th QNOOOOW

HH
H0.:-000 OlCDF-‘OOOCD

AGNIOOO QO’OOOOG

l-t

HH
(”whom-AN NH

 

122

largest scores representing the most important information,
it is seen that suspect information (12.5), physical
evidence (11.2), and victim type (9.8) are, on average, the
most important pieces of case information. These elements
were generally accessed the soonest within each
alternative. None of the subjects searched victim sex,
race, or employment status within any of the alternatives
(0.0).

When analyzing the content of search in the verbal
protocols, the importance of other considerations in
deciding whether or not to assign cases for investigations
were highlighted as well. For example:

Ok, let me see here. Case number one. First, one of

the considerations I would be making is how many people

I’ve got available to work... (Detective Sergeant C).

... Depending on case load this case might get
assigned... (Detective Sergeant B).

... Sometimes a residential ah, canvass can be done.
Only if I had detectives standing around with hands in
their pockets [would this case get assigned]...
(Detective Sergeant B).

... Chances are that I would assign that case unless I
was very short of people... (Detective Sergeant C).

... but it would also go along with whether or not we’ve
got a problem in that area, got suspects that are
working that area, and perhaps the type of property that
is taken and if its something that is unique and we’ve
got people that are hitting that type of stuff...
(Detective Sergeant C).

... there is some relationship. Again, now when I see
this I want to know what the relationship is, if its an
ex-boyfriend, or girlfriend, whatever the situation is
here that we are talking about. And now the problem, we
have an individual occupied dwelling, and there is some
relationship and it could be an ex-husband, or ex-wife.

123

Did in fact they now have legal standing in that
residence?... (Detective Sergeant C).

... we’ve got a lead here, a vehicle. And here again,
if that is something that would click with myself or any
of the investigators, then it would give us something to
go on... (Detective Sergeant C).

... There again, I would want to know if that’s an area

where we’ve been hit hard or not... (Detective Sergeant
C).

In Table 14 the linearity scores, as calculated through
the procedure outlined in Chapter Four, are presented for
each of the subjects in the decision to select a burglary
case for a follow-up investigation. The mean linearity
index score is .28 which reflects a high degree of linear
(or compensatory) decision making (0 reflects perfect
linearity and 1 represents perfect non-linearity). Although
the linearity scores vary from .14 to .38 across subjects,
all of the scores fall on the linear side of the decision
strategy "continuum."

In an analysis of the sergeants’ verbal protocols, the
linear style of decision making is also apparent.
Compensatory decision strategies are most often reflected
when the decision maker considers and weighs a range of
information elements before rendering a decision. When a
sergeant uses a compensatory strategy in deciding whether or
not to assign a case for an investigation, a combination of

information elements is considered and the additive

124

TABLE 14

LINEARITY OF SEARCH
IN THE BURGLARY SELECTION DECISION

 

 

Detective Sergeant Linearity Score
A .33
B .14
C .38
Mean .28

 

125

weight of these elements determine whether or not the case
will get assigned. For example, the following two excerpts
from the protocols reflect compensatory decision making
styles:

Case number five. Ah, look for physical evidence. Yes
there is physical evidence. Do we have a suspect?
Described and could be identified on the basis of an
eyewitness to the crime but no name. Ok, we’ve got a
good description. I’ve got good physical evidence.
Stolen property is identifiable. That is enough right
there probably, we’d have an investigator assigned.
Because the value of the property is up to $1,000, that
confirms it even more (Detective Sergeant B).

With case number three we really don’t have a whole lot
to go on except for some physical evidence. Right now
I’m kinda wondering, on this particular case, what
neighborhood this might be in. So I guess somewhere 8
along the line here I’d be kinda reading that. Ok, --
Street. But it would also go along with whether or not
we’ve got a problem in that area, got suspects that are
working that area, and perhaps the type of property
that’s taken and if its something that’s unique and we’ve
got people that are hitting that type of stuff. That
would make the assignment of this case more likely
(Detective Sergeant C).

Although the linear index scores indicate a high degree
of linear decision making, this is not to imply that
selection decisions are made exclusively through the use of
this type of strategy. In the analysis of the verbal
protocols, the use of non-compensatory strategies are
evident as well. For example, the partial protocols from
Detective Sergeant A:

Ok, on case number one, I’m going to look at the victim

type. I’m looking to see if it’s a business or an

individual. It is a business and ah, looking for a

suspect. Name provided, accused was seen committing the

crime. Ok, ah, physical evidence? No physical

evidence. Was anything taken? No serial number. Ok, I
would probably assign this case simply because I have a

126

witness who saw an individual commit the crime and

provided his name. If he didn’t provide a name, I

wouldn’t assign the case.
In this protocol it is seen that the lack of other
information such as physical evidence could not distract
from the weight attached to the knowledge of a suspect’s
name. However, given the apparent importance of suspect
information in the selection decision for this subject, it
is difficult to understand the reason for the search of any
other information elements. It appears that the same
decision outcome would have been rendered if "suspect
information” was the only dimension searched. Perhaps if
other information was found to be present, the qualifier of

"probably" would not have been necessary.

And case number five. Individual unoccupied dwelling.

Let’s find out if we have a suspect. A§_xen_enn_eeee_1[

Described and could be identified on the basis of an
eyewitness to the crime but no name. Ok, let’s see if
we have any physical evidence here so we can come up
with an identification. 0k, yes. See if there was a
vehicle involved. No vehicle. Probably not assign this
case even though you have an eyewitness. Without a name
its just a shot in the dark and I probably would not
assign that case (emphasis added).

In this excerpt, it is seen that other factors, such as the
presence of physical evidence, could not compensate for the
lack of a named accused and could not move the case over the

"assignable" threshold.

Prioritization of Burglaries
Table 15 contains data on the proportion of the fourteen

127

TABLE 15

PROPORTION OF AVAILABLE INFORMATION SEARCHED

IN THE PRIORITIZATION OF BURGLARY CASES

 

 

 

Detective
Case Alternatives

A B C D E Mean

1 .64 .43 .50 .43 .64 .53

2 .57 .79 .64 .43 1.00 .67

3 .50 .71 .64 .36 .79 .60

4 1.00 .86 .64 .50 1.00 .80

5 .86 .71 .57 .57 .93 .73

Mean .71 .70 .60 .49 .87 .67

 

128

information elements searched in each alternative for each
of the detectives when prioritizing burglary cases. It is
seen that the mean amount of information accessed across
alternatives and detectives was 67 percent. The amount of
information searched ranges from, on average, 49 percent
(Detective D) to 87 percent (Detective E). An examination
of the variability of search within subjects and across
alternatives shows that the depth of search is generally
quite variable but some subjects display more search
variability (Detectives A & E) than others (C & D).

Table 16 contains data on the importance of case
information in the prioritization of burglary cases. It is
seen that the presence of a victim offender relationship
(12.2), followed by suspect information (12.0), and presence
of physical evidence (11.9) are, on average, the most
important pieces of case information. The factors of least
importance are victim race (.2), employment status (.6), and
address (2.4).

With these general patterns realized, it is worthwhile
to highlight the variation in importance scores across
individual detectives. For example, victim desire for
effort received a score of 13 with Detective C and a score
of 7.4 with Detective A. Knowledge of a suspect’s vehicle
plate received a score of 11 with Detective B but 0.0 with
Detective D. With other examples available, it is clear

that there are individual differences among detectives on

129

THE IMPORTANCE OF CASE INFORMATION

TABLE 16

IN THE PRIORITIZATION OF BURGLARY CASES

 

 

 

Detective
Case Information

A B C D E Mean
Victim Type 14.0 5.6 14.0 14.0 8.0 11.1
Victim Sex .4 1.6 0.0 0.0 5.6 1.5
Victim Race .2 0.0 0.0 0.0 .8 .2
Victim Emp. Status 1.4 0.0 0.0 0.0 1.8 .6
Victim Age 3.2 9.0 0.0 1.6 4.6 3.7
Victim Address 1.8 2.8 0.0 1.8 5.6 2.4
Victim-Off. Rel’ship 11.2 14.0 11.0 13.0 12.0 12.2
Victim Desires Effort 7.4 10.0 13.0 9.4 9.0 9.8
Suspect Information 13.0 13.0 9.0 12.0 13.2 12.0
Physical Evidence 10.6 12.0 12.0 11.0 13.8 11.9
Vehicle Described 8.6 5.8 4.6 0.0 4.6 4.7
Vehicle Plate Known 5.0 11.0 6.0 0.0 1.0 4.6
Property Identifiable 8.8 5.0 10.0 7.4 11.0 8.4
6 Value of Property 6.6 4.2 6.6 1.6 10.0 5.8

 

130

the degree of importance attached to case information.
The verbal protocol content of search analyses

illustrates the importance of other factors in the

prioritization of burglaries as well. For example:

The categories that you have here, A through M, all
represent very important information. You could hardly
work a case without knowing this information. I think
the only other thing is, ah, say under H, nature and
source of suspect information. I want to know more
about that person because I want to take that person to
the computer, to LEMS, to LEIN. I’m going to research
him before I do anything. I want to know who I’m
talking to. Does he have accessibility to that area?
Has he committed a number of crimes? Crimes like this
before? I want to know as much about him before I talk
to him. An investigation is not as clear out as a lot
of people think. In terms of knowing about the victim,
I would like to know how often he reported crimes in the
past. I could find that out through LEMS. I’d want to
know who lives at the location of the crime. Does the
suspect live there? Did he used to live there?
(Detective D).

... If you can’t put the case together and information
doesn’t seem to come, or you’re not getting any closer,
you dump it because you don’t have pine (Detective B;
emphasis added).

... Another thing to do on case one is find out how many
crimes have occurred at this business [which did not
desire effort]. I’d see if there was a certain trend or
a certain picture here. If they had one last month and
one the month before, then you find out who the
insurance company is and go from there (Detective E).

If its a high value loss and the victim doesn’t want to
prosecute, I want to see why. I want to see if they are
insured. I want to see if they are employed or not
employed. I want to see where they live. And I want to
see if they were an accused in a crime somewhere
themselves. So I’ll usually look up their [criminal]
history as soon as I see that there is a large property
loss and they don’t want to prosecute (Detective A).

The linearity score for each of the burglary detectives

is presented in Table 17. As seen in the table, the mean

131

TABLE 17

LINEARITY OF SEARCH
IN THE PRIORITIZATION OF BURGLARY CASES

 

Detective Linearity Score

 

.38
.25
.09
.32

MUOW>

.15

Mean .24

 

132

linearity score for all of the detectives is .24 which again
represents a high degree of linear, or compensatory,
decision making. Although the index scores range from .09
to .38, all of the scores reflect extensive use of linear
decision making strategies.

The compensatory style is also reflected in the partial
protocols provided below. In reading the protocols, it is
apparent that an understanding of each case (on which
decisions of priority are based) is achieved only after
particular pieces of information are considered and weighed
together. Hence, the meaning of each case develops only
after all the information elements are "added together" in a
compensatory style. In addition, it is seen that in several
of the following excerpts (as well as some of those
previously presented) certain "questionable circumstances"
lead to an increased depth of search. For example:

Case number four... ... What’s the value of it? $795.
That’s a lot of property. Let’s see if there was any
evidence available? No evidence available. Well, what
does the victim want to do? Yes they want to prosecute.
With that much property taken I’m kind of curious as to
where they live. What side of town? What area they
live in? That’s getting to be pretty high in value. --
Avenue. My first thought is that that is a lot of
property taken from that area up there. There are some
nice houses over there but there are also a bunch of
dirt-baggy houses too. I’d be interested in what type
of property was taken. If it was cash I would really
question the situation. I would go down and see where
this guy works. His employment status. Not employed.
Now I’ve got some real questions about it. The first
thing that comes to my mind is that it is an insurance
rip. This is a more common situation for a male so I’m
going to look at victim sex. It’s a male. I’ll look at
the race. Non-white. This kind of case, the more I see
the more I went tcgknow before I even go talk to
somebody. I want to know as much as I can because there

133

are some real unusual circumstances. I might put this
case on the back burner until I see what comes up for
awhile (Detective A; emphasis added).

From the same detective:

Case five... ... Ok, how much was it worth? $1,000.
I’m assuming that the person is going to want to
prosecute with that much loss. Yes, they want to
prosecute. I’ll see where they work and live with that
much property loss. -- Avenue. Kind of a working class
neighborhood. Ok, they’re employed. That’s not too
unreasonable then. How old is this person if they’ve
got that kind of property to lose? 47. That is about
what I would have expected... This case appears to be
probably the case which would be the most time
consuming.

From Detective E:

Ok, normally on my regular case investigations, all of
the information, if it’s available, you go ahead and
correlate everything together...

Now another thing that I look at here is that in case
four we got a male, 32 years old, non-white, not
employed, and lives on -- Avenue here in Landau. --
Avenue is an area where, it was a good area years ago
but now we got a lot_of problems with narcotics, dope,
cocaine, and so forth and so on. So associating --
Avenue with a non-white male and $795 ripped off, first
thing I’m going to be looking at is a dope rip-off of
some sort. (Priority was fourth of five cases.)

Another example of the need for an increased depth of
search is provided below:

In terms of property loss... if it is a real high value
single piece of property, the owner is going to have
something to prove that they own it. Like a $1,000
T.V. There are not a lot of $1,000 T.V.s and they don’t
come in neighborhoods where, you know, there are
dirt-bag houses. If someone lives in a house that is a
twenty, twenty-two, twenty-three thousand dollar house
and they’ve got a $1,500 T.V. and they don’t have a
receipt for it? Now I got a real problem with that,
right off the bat. Now I want to see where the guy
works. See what other pieces of property he has in the
house that is going to show me that he’s going to spend
$1,500 on a T.V. (Burglary Detective A).

134

In the analysis of the burglary detectives’ protocols,
the existence of non-compensatory decision making is evident
as well but only in reference to when the victim did not
desire investigative effort. For example:

... and the last [dimension] I’m going to look at as far
as all cases is victim desires effort. The reason that
this is done is that before you get too involved in the
case you want to know if your victim gives a shit. I’m
going over the categories that I’ve turned over to try
and determine if I want to turn over some more of these
before I make a determination or if I’m ready to make or
get rid of some of these cases. Ok, at this point in
time, case number one, I would not follow-up. We have a
business that doesn’t care. Under victim desires effort
there is a no so I’d dump it at that point. This one
would not be worked on (Detective B).

... What’s big to me is this right here, victim desires
effort, because I could be working a case and spending
five, six, seven days on it. So this is big to me. I
don’t want to waste seven days on even a legitimate
crime if you’re not going to get cooperation from the
victim (Detective C).

... then I would have to look at victim desires effort.

No. So that closes this case. That’s the determining

factor for closing this case out (Detective D).

... in most cases if the victim desires no effort,

basically I don’t, I go on to the next case because

number one, the prosecutor will never authorize [an

arrest warrant] (Detective E).

The excerpt from Detective A’s protocol (below)
concerning the victim’s desire for effort is unique from
those above in that although it initially reflects
non-compensatory decision making, it allows for the
possibility of other factors (i.e., strong suspect
information) to compensate:

Case one... ... Victim desires effort? No. Well, that

puts this one on the bottom burner real quick. Although

I might still be interested in it because it might be a

case that I’m working in conjunction with something

135

else. However, the most valuable piece of information
here is that we’ve got a name provided, the accused was
seen committing the crime. That leaves me still real
interested in the case because I still may be able to
convince somebody that I still want the case. So this
case, I would not can it. It’s a good case, not a piece
of junk. So I may go over his head or at least pressure
him a little bit.
With four of the five burglary detectives then, nothing
could apparently compensate for the victim not desiring
effort -- even the fact that the accused was named and seen
committing the crime. Accordingly, with four of the five
detectives, this case was assigned the lowest priority of

all the cases reviewed.

Prioritization of Robberies

Table 18 contains data on the proportion of the sixteen
information elements searched within each alternative by
robbery detectives when prioritizing robbery cases. As seen
in Table 18, the mean amount of information accessed across
alternatives and detectives is 59 percent. The mean amount
of information searched ranges from 42 percent (Detective A)
to 75 percent (Detective B). An examination of the
variability of search within subjects and across
alternatives shows that there is more variance with the
search of Detective B than with Detective A.

In Table 19, data on the importance of case information
in the prioritization of robbery cases is presented. As

seen, victim desire for effort (12.2), suspect information

136

TABLE 18

PROPORTION OF AVAILABLE INFORMATION SEARCHED
IN THE PRIORITIZATION OF ROBBERY CASES

 

 

 

Detective
Case Alternatives
A B Mean
1 .13 .25 .19
2 .13 .81 .47
3 .56 .88 .72
4 .63 .94 .79
5 .63 .88 .76
Mean .42 .75 .59

 

137

TABLE 19

THE IMPORTANCE OF CASE INFORMATION
IN THE PRIORITIZATION OF ROBBERY CASES

 

 

 

Detective
Case Information

A B Mean
Victim Type 0.0 16.0 8.0
Victim Sex 0.0 7.0 3.5
Victim Race 0.0 0.0 0.0
Victim Employment Status 0.0 .8 .4
Victim Age 0.0 5.0 2.5
Victim Address 0.0 6.4 3.2
Victim-Offender Relationship 16.0 7.2 11.6
Victim Desires Effort 10.8 13.6 12.2
Suspect Information 9.0 14.8 11.9
Physical Evidence 8.4 7.8 8.1
Vehicle Described 7.8 4.2 6.0
Vehicle Plate Known 4.8 5.4 5.1
Property Identifiable 6.8 6.4 6.6
Degree of Injury 5.2 8.4 6.8
Weapon Used 10.8 9.2 10.0
$ Value of Stolen Property 6.2 5.6 5.9

 

138

(11.9), and the presence of a victim-offender relationship
(11.6) are, on average, the most important elements of case
information. These elements were generally accessed the
earliest in each of the alternatives. The information
elements of least importance are victim race (0.0),
employment status (.4), and victim age (2.5). In this
exercise at least, these information elements were rarely
considered in determining case priority.

With these general patterns realized, it is important to
call attention to the apparent individual differences which
exist in the data. For example, victim type was generally
one of the first information elements accessed by Detective
B, while Detective A never accessed this information in any
of the alternatives. Detective A never searched any of the
victim demographic characteristics, while Detective B did.

The content of search analyses calls attention to other
factors which are considered in prioritizing robbery cases.
For example:

I want to know where this thing happened but that

information is not provided. I want to know if there

are any parallels in these cases. Maybe the same guy

did several of these. That is something that I would be

looking for. Then you would bunch these cases

together. My biggest concern with everything that has

happened over the weekend, do we have a guy who wants to

go out and rob everybody or not (Detective B).

Table 20 contains the linearity scores for detectives in
the prioritization of robbery cases. The mean linearity

score is .35 which, congruent with the other decisions,

represents a high degree of linear decision making. With

139

TABLE 20

LINEARITY OF SEARCH
IN THE PRIORITIZATION 0F ROBBERY CASES

 

 

Detective Linearity Score
A .47
B .23

Mean .35

 

140

the two linearity scores being .47 and .23, individual

differences among detectives are once again highlighted.

An example of compensatory decision making is provided

in the following partial protocol from Detective B. It is

seen that all of the information elements considered and

weighed together (especially, degree of injury, suspect

information, age of victim, and sex of victim) elevate the

priority given to the case.

Let’s go to case number three. Victim type? See

suspect information. Property identifiable? See how
much it was worth? $180. Does she want to prosecute?
Yes. Victim Sex? Female. I knew it. I don’t really
care if its a male or female victim. I don’t really
care about the race of the victim. That’s the least of
my concerns. Weapon used? No. Was she hurt? Yes,
serious, broken bone. Ok, so now I need to know a whole
lot of information. She deserves some immediate
contact. She doesn’t appear to know who the accused was
in this thing and she doesn’t have a relationship or she
isn’t giving up that she has one with the suspect. She
wants to prosecute. _She is out some money. I would
want to find out how old she is. Above 50? 69. I knew
it. Seeing as to that she is older, I have a soft spot
in my heart. I got some real concerns about her.

Before I even read case number four I might give her a
call and just find out how she is doing... Case three
will receive top priority. I base that on the fact that
she did receive an injury, there wasn’t a weapon, but I
want to know more about the case. She might be able to
identify. If she can’t identify, then she drops in
priority to maybe third.

Non-compensatory decision making is also evident in the

protocols of the robbery detectives. As with burglaries,

non-compensatory strategies were used when the victim did

not desire effort. For example:

Case number one would be the last case that I would
work. I would call the victim, ask why he didn’t want
to follow through on prosecution, and then close the
case with no further action (Detective B).

141

... so the next thing that is important in these two
cases would be ah, desire for effort. I guess that
would be the next thing. Ok, case number one, that case
would be gone right away. Unless you see that there is
an on-going problem, you might push that victim a little
harder. So this one could be a real quicky (Detective
A).

Summary

Consideration of the findings across the three analyses
shows that the greatest amount of information was searched
in the prioritization of burglaries (67%), followed by the
prioritization of robberies (59%), and then the selection of
burglaries (46%). In all of the decisions, the offense
characteristics appeared to be of more importance than
victim characteristics. Suspect information was the most
consistently important factor across all decisions. Victim
race and employment status appeared to be the least
important across all the the decisions. Regarding the
extent to which detectives use linear strategies in decision
making, it was found that all of the detectives (and
detective sergeants), used a primarily compensatory style.
The highest degree of linearity was displayed in the
prioritization of burglaries, (.24), then the selection of
burglaries (.28), and then the prioritization of robberies

(.35).

142

Observations

The following discussion is based primarily on
observations of, and interviews with, burglary and robbery
detectives. Detective sergeants were not systematically
observed and therefore, these data do not reflect the
complexities of the case selection decision. While the data
derived from the observations were broad and diverse, the
discussion presented here focuses specifically on describing
how cases are processed by detectives. Accordingly,
additional light is cast on the decisions made by
detectives, how detectives make these decisions, and how
these decisions determine the amount of time spent on a
case.

The discussion centers around Figure 2 which is an
illustration of the process by which cases are interpreted
and disposed of by detectives. The organizationally
recognized case dispositions (means by which cases can be
"cleared") are noted within the parentheses (see Chapter
Three pp. 59-60 for definitions of the statuses). This
illustration also represents the framework for some of the
discussion presented in Chapter Six.

As portrayed in Figure 2, a follow-up investigation
consists of a series of decision stages. The decisions made
at each of these stages reflect the meanings attached to the
case and guide the conduct of the investigation. The model

is a simplification, and perhaps an over-rationalization, of

143

FIGURE 2

A DECISION

MAKING MODEL

OF BURGLARY AND ROBBERY INVESTIGATIONS

 

does this sound li e a legitimate crime?

yes

victim no e—"verify facts"

longer
wishes to
pursue/
close
case
(VCI/VRP)

 

1?
does this case

no

initial investigation report assigned/reviewed

"verify facts”—+victim no

 

<r
does this case

longer
wishes to
pursue/
close
case
(VCI/VRP)

have legitimacy? lack legitimacy?

yes no yes no
clclse clise
case case
(UNF/NFI) (UNF/NFI)

 

 

are there :good leads?"
yes no

exhaust good leads
perpetratdI identified?

yes no

I

v~

 

*§
close case is this case "serious" or "unusual?" QR

(WPA/RPC/COP) is there enough time to keep working
this case? \&
yes no
cultivate close case
information (NFI)
"leads developed?"
yes 0

exhaust leads
perpetrator identified?

yes no

I

. ‘ .
continue to cultivate information/
close case
(NFI)

 

close case
(WPA/RPC/COP)

 

144-

a very complex process but it does seem to capture much of
how burglary eng robbery detectives think, and how they go
about "working" cases.

The process begins when the patrol officer’s report is
assigned to, and reviewed by, a detective. After reviewing
the report, the first determination is whether or not the
reported crime sounds legitimate. That is, is the victim’s
"story" truthful? and/or has there really been an
"injustice” done here?

As evident in the protocols provided earlier, ”stories"
told by certain types of victims, and incidents which
involve certain circumstances, are viewed by detectives as
questionable in their truthfulness. Most fundamentally, did
a crime really occur? Is this a phony report to rip-off an
insurance company? Or, say in the context of a robbery of a
taxi cab driver or a gas station attendant, did the "victim"
pocket the money and then claim to have been robbed?

Reports are also questioned on their legitimacy if it
appears that the crime resulted from illegal activities in
the first place. For example, given the facts and
circumstances of the incident, does it appear that the
"robbery” was really "a drug deal gone bad?" Was it a
”payback" where ”the victim got what he deserved?" Was the
”victim,” for example, carrying money to buy drugs? Or was
it money received from selling drugs? Did a ”hooker rip him

off?" Generally, crimes with young black male victims, or

145

crimes which occurred in a "had” (high drug, high crime,
high non-white population) part of town, especially after
dark, are questioned the most in terms of their overall
legitimacy. Crimes with older white females are questioned
the least.

By "reading between the lines” of the patrol officer’s
report and considering the details of the incident, the
detective formulates an idea of what reelly happened -- he
gets "a feel" for the case. At this point in the
investigative process, the detective can make a tentative
judgement that yes, this does sound like a legitimate crime,
or no, this does not sound like a legitimate crime. With
either interpretation, there is a need for the detective to
"verify the facts" of the case and find out what actually
happened —- to ”see if the victim’s story check’s out." In
doing so, the detective’s preliminary assessment regarding
the legitimacy of the case can be confirmed or
disconfirmed. This is most often accomplished by
recontacting the victim and asking him to once again to tell
what happened. Specifically, this recontact can have
several purposes. First, it is a means by which the
patrol officer’s interpretation of the incident can be
verified. Second, additional information which would help
in the investigation could possibly be produced. Third, by
recontacting and re-interviewing the victim, the detective

can ”test" the victim to see if he can provide the same

146

story twice. If there are inconsistencies between the
patrol report and the account provided to the detective, it
may signal an untruthful victim and a report lacking in
legitimacy.

Fourth, a personal contact with the victim allows the
detective to see the victim and judge his credibility in
person. As one burglary detective stated in a protocol, "1
think case four... that’s a go-down-and-see person. I
wouldn’t do any of that over the telephone. I’d want to see
the person in person to judge their credibility." When
"judging credibility," the victim’s ”body language" can be
observed when asked ”the tough questions." With such
observations, further evidence could be offered as to the
(il-) legitimacy of the complaint. The following message,
which was on a note attached to one of the follow-up
investigation reports reviewed for the statistical analyses,
provides an example of how detectives use the victim’s body
language to give meaning to a case: "He did fair in the
interview. When I asked him if he knew who robbed him, he
dropped his eyes and said no. Maybe a little payback?"

Finally, as a result of the contact, it may be learned
that the victim, for whatever reason, does not wish to
pursue the complaint. While detectives may question the
motivations of the victim for such a desire (i.e., is this a
false burglary to rip off an insurance company?), the case

is most often closed without much hesitation (Victim Cancels

147

Investigation; "VCI," if culprit is not known; Victim
Refuses to Prosecute; ”VRP," if culprit is known). In some
cases, when questionable circumstances exist, the first
question the detective asks the victim during the interview
is whether or not he would "be willing to go to court and
testify against the the person who committed the crime.” If
the victim says no, it means that the victim does not desire
effort and the case is closed (as either "VCI" or "VRP").
In addition, if attempts are made to contact the victim but
the victim never responds, it is assumed by the detective
that the victim does not wish to pursue the complaint which
also allows the investigation to be quickly closed ("VCI,”
”VRP").

The information produced by ”verifying the facts" can
contribute additional weight to the initial judgement that
the report is either legitimate or lacking in legitimacy.
As stated by a burglary detective in a protocol, "I think
that case number four would be easy enough to either figure
out and write off as a bullshit case or maybe it is a
legitimate case." If there is a strong conviction that the
case is "bullshit" or "not legitimate" after the facts are
verified, then the case would most likely be closed (as
Unfounded, "UNF;" or No Further Investigation, "NFI”).

For the large proportion of cases which remain, the
question becomes: ”are there ’gocd leads’ for an

investigation?" A ”good lead" is a piece of information

148

which is likely to "lead" directly to the culprit and
accordingly, is associated with a specific investigative
activity. For example, a suspect’s name provided by an
eyeWitness to the crime is a "good lead" and an "interview"
of the suspect is the specific activity. If a robbery
victim feels that she can identify the perpetrator, and the
description she provided matches that of a known robber
(leads), then a photo line-up would be the specific
activity. If there are are fingerprints available and a
named accused is known or suspected, then the prints could
be submitted for a comparison, etc.

If there are "good leads" and, as a result of performing
the activities associated with these leads, the perpetrator
was positively linked to the crime ("tied down"), the case
could be closed (Warrant Pending an Arrest, "WPA" if an
adult; Refer to Probate Court, "RPC" if a juvenile; or
Accused in Other Prosecution, ”COP" when identified but not
charged with the present crime). If good leads were
exhausted and a perpetrator was not positively identified,
_p if there were no good leads to begin with, then, before
more time was spent on the investigation, the following
questions had to be answered: (1) is this case "serious" or
”unusual" in any way? and (2) is there enough time to keep
working this case?

A case can be defined as "serious" or "unusual" in

myriad ways. For example, the victim could have been

149

severely injured, a large amount of property may have been
taken, the crime may be part of a pattern, the victim may be

H

of high status and have "connections, the perpetrator(s)
may be seen as particularly dangerous (e.g., an "Asian
gang"), or the suspected perpetrator may have personally
challenged the detective at some time in this or other
investigation (e.g., "come back when you can arrest me but
now get the fuck out of my face"). These factors are always
considered in relation to the legitimacy of the crime and
the circumstances surrounding the incident.

Consideration of time pressure is also important in
determining how much time will be spent on a given
investigation (and if the case is assigned at all as seen in
the earlier protocols). If the detective is assigned
numerous cases on a daily basis (vs. only a few cases per
week), there is less time to work on any one case and, in
order to maintain control of the workload (”avoid being
swamped"), there is more pressure to close cases more
quickly. If there is not enough time to keep working the
case, or if the case is not unusual or serious in any way,
the detective might tell the victim to call if s/he finds
out who committed the crime. The case would most likely
then be closed ("NFI").

If a great deal of time pressure is not perceived, or
the case is serious or unusual, then the task of the

(self-motivated) detective is to cultivate information, the

150

most time consuming aspect of a follow-up investigation.
The importance and usefulness of cultivating information is
also increased when it is believed that just "a little bit
more" information is needed to identify the culprit. When
cultivating information, the activities performed do not
emanate from specific pieces of information. As a result,
the activities do not have as much direction as those
associated with "following leads;" information which is
cultivated is often of poor quality sometimes leading the
detective "in the wrong direction." Detectives may be able
to cultivate information by showing mug shot books,
conducting neighborhood canvases, talking with parole or
probation officers to see if they know of anyone matching a
given description, searching the crime scene, interviewing
informants, talking to detectives from other departments to
see if "anything rings a hell" with them, or "patrolling"
the area of the crime and (1) making "street stops" to ask
individuals if they know anything about the crime (e.g.,
"Who rides the white mountain bike around here?" "Has
anybody been talking about the old lady who was ripped
off?") or (2) trying to locate someone who matches a given
description. As a result of these activities, "good leads"
may be developed. If, in exhausting these leads, a
perpetrator is identified, the case would be closed ("WPA;"
”RPC;" "COP"). If a perpetrator is not identified as a

result of exhausting the newly created leads, or if no leads

151

were present in the first place, activities oriented around
cultivating information could continue indefinitely or cease
immediately ("NFI"). The continuation of effort and the
allocation of time to such a case would once again be a
function of time pressure, and the seriousness or uniqueness
of the crime.

In conclusion, as portrayed here, the follow-up
investigation resembles a filter where cases with certain
combinations of characteristics get the most amount of time
devoted to them and other cases with other combinations of
characteristics get the least amount of time devoted to
them. Given the "big picture" sketched here, decision
making of detectives sometimes reflects the non-compensatory
style and at other times it reflects the compensatory
style. For instance, little if anything can compensate for
a victim who does not wish to pursue the complaint in
determining the amount of time to devote to the case.
Similarly, if a case is interpreted as ”bullshit," nothing
can compensate to allow the case to receive more time or
effort. Alternatively, the uniqueness or seriousness of the
crime or the availability of time (lack of time pressure)
can compensate for the lack of evidence in a case. Cases
are not always closed simply because there is a lack of
evidence. Evidence is the fuel of the investigation but for
some cases, as described here, efforts are made to create

fuel. In the next chapter, this perspective on

152

investigations, its implications, and some of the causal

linkages on which it rests are discussed further.

153

Footnotes

If one included those burglaries which received a ”No
Report Forthcoming" status, the percent of burglaries
assigned to detectives would drop to 21% (317 out of
1,536).

The derivative at the mean is analogous to the
standardized beta weight in linear regression.

This is particularly true in explaining the

inconsistency in the derivative for ”suspect vehicle plate
known.” Suspect vehicle plate known and suspect vehicle
described have a correlation of .46. This appears to be
reflected in the standard error of the "plate known"
variable (SE[B] = 386.788). The inflated standard error
ultimately distorts the derivative at the mean statistic
(.724) in the probit analysis. The multicollinearity does
not appear to affect the regression results, however.

However, although common in the literature (e.g.,

Sommers & Baskin, 1990; Smith & Klein, 1984), it is not
appropriate to rank the impact of variables based on their
degree of significance (except if the variables are
measured on the same scale). With consideration of the
significance levels, one may reject the null hypothesis
but that is all (Johnson, Johnson, a Buse, 1987).

The value losses involved in the bank robberies were:
$460, $926, $2,000, $2,145, $5,100, $7,576, $9,132 (two
missing).

The following were the total amounts of time (hours)
spent in the bank robbery follow—up investigations: 6.0,
16.7, 18.0, 20.0, 27.8, 31.8, 37.0, 37.5, 50.8.

It is necessary to distinguish between ”importance" and
”influence." To determine influence, the value of the
dimension searched and its impact on the decision outcome
must be assessed. The determination of importance, as
outlined in Chapter Four, rests with simply identifying
when the dimension was accessed in each alternative.
Because of the conceptual and empirical differences
between "importance" and "influence," any causal
inferences which are made on the basis of "importance"

must be interpreted with caution.

For confidentiality reasons, street names are not
included in the protocols.

154

CHAPTER SIX

DISCUSSION

Chapter Six contains a discussion based on the results
of the study. The chapter is divided into five sections.
In the first section, the decision outcomes of detectives
are placed in the context of previous research. The second
section discusses the impact of case characteristics on the
decision outcomes of detectives. The third section
discusses the "black box” of the outcome oriented results --
the cognitive processes associated with detective decision
making. The chapter concludes with a discussion of the

study’s limitations and directions for future research.

Decision Outcomes in Context
Before discussing the stimuli which provoke the decision
responses of detectives and the cognitive processes
associated with the decisions of detectives, it is necessary
to briefly compare the decision outcomes obtained in this
study with those identified in previous research. This task

is complicated by the fact that only one study in the

155

literature provides statistics on a comparable population of
cases. For example, Bynum et al. (1982) state that 82
percent of "all property and personal crimes” ”received
little or no investigative effort" (p. 315). According to
Ericson (1981), 30 percent "of all cases” that came to the
attention of the detective bureau received "one or more
hours" of investigative time. Greenwood et al. (1977)
report that approximately 63 percent of robberies, 36
percent of non-residential burglaries, and 30 percent of
other burglaries received "at least a half hour of a
detective’s time" (p. 130). Eck (1983) however, does
provide comparable figures. In Table 21, the percent of
burglary and robbery cases assigned to detectives along with
the mean amount of time spent on the cases in the three
jurisdictions studied by Eck are presented along with those
same figures obtained in this study.1

In examining the burglary statistics, it is seen that in
Landau the lowest percentage of cases were assigned.
Perhaps for this reason, those burglary cases that were
assigned, on average, received a relatively large amount of
investigative time. With robberies, there was little
variance across jurisdictions in the percent of cases
assigned but Landau was second to DeKalb in the amount of
time devoted to them. From this comparison it appears as
though the screening of burglaries in Landau followed rather

stringent criteria, at least in relation to those

156

TABLE 21

 

 

DECISION OUTCOMES FOR BURGLARY AND ROBBERY CASES
IN ECK (1983) AND PRESENT STUDY
Jurisdiction % Burg’s Hours % Robb’s Hours
Assigned Spent Assigned Spent
DeKalb 45.4 1.9 100.0 5.1
St. Petersburg 35.3 1.1 100.0 .9
Wichita 76.1 .8 100.0 2.4
Landau 20.6 3.7 96.0 4.5

 

157

 

jurisdictions examined by Eck. Accordingly, although devoid
of case load information (number of cases assigned / number
of detectives available), the time spent on cases in Landau

does appear reasonable given the context of Eck’s study.

Case Characteristics and Decision Outcomes

According to Black’s (1976) theory of the behavior of
law, the seriousness of a criminal incident, and thus the
"amount of law" mobilized as a result of the incident, is a
function of the social structural characteristics of victims
(and/or offenders). To test Black’s theory, the influence
of victim characteristics on the case selection and time
allocation decisions of detectives was examined. Overall,
the findings of this study do not lend support to Black’s
theory. In the statistical analysis, neither victim sex,
race, employment status, income, nor age2 had an
appreciable effect on the "amount of law" invoked by
detectives in any of the three decision situations
analyzed. Only the ”organization” of the victim ("victim
type") appeared to affect decision responses; robberies of
"businesses" were more likely than robberies of
”non-businesses" to have more time spent on the follow-up
investigation but only when banks were included in the
analysis. When bank robberies were excluded from the
analysis, the impact of "organization” disappeared.

In addition, opposite of the expectations outlined by

158

 

Black, it was found that the presence of a relationship
between the victim and the offender inopeased the likelihood
that a burglary report would be selected for a follow-up
investigation. If Black’s hypothesis on this issue would
have been supported, the presence of a relationship between
the victim and the offender would have deepeneed the
likelihood of a report being selected. In the two time
allocation decisions, this variable did not have a
significant effect in either direction. In short, of the 18
hypotheses (6 variables x 3 decision situations) derived
from Black’s theory which asserted a causal relationship
between victim characteristics and decision outcomes, 16
were not supported, one was partially supported, and one was
contradicted.

Results from the information board exercise,
specifically the content of search analysis, are also
generally non-supportive of Black’s predictions. Although
it is difficult to make causal inferences on the basis of
these data, it does appear that the social structural
characteristics of victims are less important than offense
characteristics in the case selection and time allocation
("prioritization") decisions of detectives. In the case
selection decision, of the six victim characteristics which
predict decision outcomes according to Black, only one
(victim type) received an ”importance score” which ranked in

the top half of all scores. In the prioritization of

159

burglaries and robberies, the "importance scores" for only
two of the six victim characteristics (victim offender
relationship and victim type) ranked in the top half of all
scores.

While the regression and content of search analyses
indicate that the social structural characteristics of
victims do not, for the most part, affect decision outcomes
of detectives, the observational data, along with several of
the verbal protocols presented in Chapter Five, suggest that
victim characteristics do have an impact on detective
decision making. However, these influences do not take the
form of independent causal relationships as assumed by
Black. Rather the influences result from the combination of
certain characteristics. For example, in the statistical
analyses it was found that victim sex and age did not have
independent effects on the time allocated to an (burglary or
robbery) investigation. However, observations suggest that
crimes which involved older female victims3 were more
likely to be viewed as legitimate, and, in robberies
particularly, much more likely to be viewed as "unusual" or
”serious" than crimes with younger male, younger female, or
even older male victims. On the other hand, robbery and
burglary cases with younger, non-white, male victims4 were
more likely viewed as lacking in legitimacy or as

"bullshit," than cases with older, white, female victims.

However, due to the relative rarity of such victim

160

characteristic combinations in the data, it is not possible
to test these observations through statistical analyses, at
least with the data available here.

In addition to the victim characteristic hypotheses
derived from Black, this study also examined the expectation
that the preferences, or wishes, of the victim affect the
decision responses of detectives. In the statistical
analyses, victim desire for investigative effort was found
to be influential in robbery and, to a lesser (statistically
non-significant) extent, burglary time allocation decisions;
victims who desired effort were more likely than victims who
did not desire effort to have more time spent on their
investigations ("desire for effort” was not included in the
case selection decision). I

The content of search analyses were generally congruent
with the statistical results although once again, causal
inferences are problematic with these analyses. Victim
desire for effort was ranked as the most important case
information element (of sixteen) in the prioritization of
robberies and fifth most important (of fourteen) in the
prioritization of burglaries.

Based on observational data, "desire for effort" was the
only victim characteristic which had a direct and
independent effect on time allocation decisions. If the
victim did not wish to pursue the complaint, then typically

any additional investigative effort was viewed as "a waste

161

of time" by the detective because the prosecutor’s office
would not issue a warrant for the arrest of the perpetrator
even if s/he was identified. Closing a case due to victim
preferences was a "quick and easy" way of disposing of a
case (cf. Ericson, 1981). The predictive power of this
variable was perhaps mitigated in the statistical analyses
by the fact that even if a victim desired effort, it did not
necessarily mean that time was spent on the investigation.
The case could have been quickly closed for a variety of
other reasons (e.g., lack of legitimacy, lack of good leads,
etc.).

In placing the present "victim characteristic" findings
into the context of previous detective decision making
research, it is seen that the findings of this study are
similar to those presented by Bynum et al. (1982). Recall,
these authors found that victim sex, race, age, and
employment status did not (at least directly) influence
detective decision making in terms of the effort expended in
follow-up investigations, although victim income in burglary
cases did. Support for the expectation that victim income
affects detective decision making also received support from
Ericson (1981) and Waegel (1981). In addition, Waegel
(1981) found that the race of the victim had an impact on
detective decision making. However, none of the available
studies provide a discussion of the influence of other

victim characteristics on detective decision making.

162

Therefore, this study supplements those studies previously
reported by providing insight into the effect (or
non-effect) of victim type, victim-offender relationship,
and victim desire for effort, as well as providing
additional support for the previous findings that victim
age, race, sex, and employment status do not directly affect
detective decision making. Finally, the observational
component of this study, as well as the verbal protocol
analysis, further contributes to our understanding of the
influence of victim characteristics by highlighting the
complex calculus by which these causal influences are
manifested in the decisions of detectives.

When placing the ”victim characteristic” findings of
this study (as well as of the other detective decision
making studies) into the larger literature on police (patrol
officer) and criminal justice decision making, it appears
that victim characteristics are more influential on the
decisions of police (patrol officers) and other criminal
justice actors than they are on the decisions of
detectives. For example, Smith and Klein (1984), Smith
(1987), and Black (1970) all found victim income to have a
direct and independent impact on the decisions of patrol
officers (decisions to arrest or file a complaint). Victim
sex was also found to affect the decisions of patrol
officers (Smith & Klein, 1984; Smith, 1987) and prosecutors

(Williams, 1978). Smith (1987) found that the race of the

163

victim influenced the decision of patrol officers (but see
Smith & Klein, 1984). Similarly, numerous studies have
found that the presence of a victim-offender relationship
(or a particular type of relationship) influenced the
decisions of patrol officers (Black, 1971; Smith & Visher,
1981; LaFave, 1965; Friedrich, 1977; Worden & Pollitz, 1984)
and prosecutors (Schmidt & Steury, 1989) in the manner
predicted by Black.

The inconsistencies which appear when comparing
detective decision making research with this larger body of
police and criminal justice decision making research allow
for the possibility that the follow-up investigation
represents a unique decision stage in the justice system,
distinguishable from ”police decision making" and "criminal
justice decision making” more broadly, in that victim
characteristics generally do not have (at least independent
and direct) effects on the decisions of investigators.

Alternatively, one could argue that the number of null
findings produced here (especially in reference to those
victim characteristics identified as important by Black)
raise questions about the appropriate conception of crime

" if "seriousness" can predict the decision

"seriousness,
outcomes of detectives. It may be the case, for example,
that the behavior of law is not so much a function of victim

characteristics as it is characteristics of the offense,

specifically those indicative of harm -- value of stolen

164

property, use of a weapon, and degree of victim injury.

Of the offense characteristics which reflect the actual
(or potential) harm done, only the value of the stolen
property was significant in the statistical analyses. In
fact, of all the case characteristics included in this
study, the amount of property loss was the most influential
in both of the time allocation analyses. This variable also
displayed a significant impact in the case selection
analyses although to a lesser degree than in the other
decision situations.

Despite the major impact property loss displayed in the
statistical analyses, the content of search analyses shows
that none of the importance scores for "value of stolen
property" ranked even within the top third of all scores.
This inconsistency may be explained by the frequency in
which detectives use compensatory strategies in making
decisions. With compensatory strategies, a high value on
one dimension (e.g., $15,000 of stolen property) can
compensate for a low value on another (e.g., weak suspect
information) regardless of the order in which the
information elements are accessed. Therefore, given
extensive use of compensatory strategies and the previously
discussed inability of "importance scores" to reflect the
values of the accessed dimensions, it is seen once again
that "importance scores” based on content of search are not

necessarily valid indicators of "influence.”

165

The influence of the value of stolen property on the
amount of time allocated to follow-up investigations was
evident in the observations as well. However, as seen in
several of the protocols presented in Chapter Five, when the
circumstances which surrounded the incident seemed "fishy,"
a large amount of property loss often raised questions about
the legitimacy of the crime (recall the protocol with the
theft of the $1,000 TV in the "dirt—baggy” area of the
city). However, most of the the cases which involved an
"extreme” amount of stolen property involved victims who
were known in the community, and known to have such property
to lose. Perhaps for this reason alone were these types of
cases viewed as deserving of investigative attention.

Case in point: During the observation period, a
burglary was reported by a well known real estate developer
in the city of Landau. This individual was described by the
detectives as "owning half of the city." The claim of this
victim was that $10,000 in cash was taken from his residence
and the likely culprit was the exterminator (the "bug man")
who had access to the house the morning the cash was
discovered missing. None of the detectives doubted the
ability of this victim to have $10,000 in cash taken. After
numerous hours of time were allocated to the investigation
(including several polygraph examinations of the suspect and
a discussion of the possibility of "staking-out" the

accused) the investigation was terminated without an arrest.

166

If the victim would have been ”Joe Nobody," the legitimacy
of the incident would have most likely been questioned
perhaps to the point of being "written off as bullshit."
But the important point for consideration here is that "Joe
Nobody" does not report a loss of "$10,000." Maybe $100,
$500, or even $1,000, but not $10,000.

One of the detectives elegantly summarized the way in
which property loss affects the time allocated to cases:

The police department is a political animal. Somebody

that has got $25,000 has got some political touchings.

So therefore, I’ve got to pay attention to that person

over the person who gets robbed of his [newspaper]

money. Plus the higher the money, the more likely of
that going into the press which means that my captain
and my chief are going to want to know what is going on
here. The media is more interested in the big money
losses than the small ones.

While the dollar value of the property loss was
influential in the statistical analysis, neither the degree
of injury nor the use of a weapon were found to affect the
amount of time allocated to robbery follow-up investigations
(these indicators of harm were not, of course, appropriate
for burglaries). In the content of search results, degree
of injury and weapon use both obtained "importance scores”
which ranked among the top half of all scores (seventh and
fourth, of 16 respectively).

The results of the statistical analyses were generally
confirmed through the observations. The only time weapon

use, by itself, elevated the priority of a case or led to an

increased amount of time being spent on the investigation

167

was when the culprit used some "unusual" weapon like a
sawed-off shotgun or a machine gun (i.e., "an Uzi") to
perpetrate the act; again, a statistically infrequent
occurrence and one not captured in the operationalization of
the ”weapon use" concept. An individual who used such a
weapon was viewed as particularly dangerous, likely to cause
great harm to someone in the future and hence, one who
needed to be "dealt with.”

When serious injury was inflicted on a victim as a
result of a robbery, it often (but not always) gave the case
enhanced credibility. However, this "enhanced credibility"
did not always translate into additional time being spent on
the case. For instance, the detectives told numerous
stories about "victims" who caused self-inflicted injuries
and then claimed that the injuries were a result of a
robbery. In one memorable incident, for example, an
employee of a tire store stole cash from the store and then,
once the cash was secured, hit himself over the head with a
tire-iron, causing unconsciousness. It was not until other
employees discovered the "victim" on the floor lying
unconscious in a pool of blood that a "robbery” was
reported. Because of such incidents, detectives know that
injuries are not necessarily "the truth." In addition,
consideration of the person injured (e.g., old female vs.
young male) and the circumstances of the event (e.g.,

occurred on the street at 3:00 a.m. or in the victim’s house

168

at 4:00 p.m.) add additional meaning to the case. Finally,
degree of injury is not "serious" in the same way that
property loss is "serious." According to the detective’s
statement cited earlier, an extensive amount of property
loss is serious not because of the amount of property taken
per se, but rather it is serious because of who the victim
may be -- not anyone can have "$25,000" taken, but anyone
can be injured.

In placing these offense characteristic findings into
the context of previous detective decision making research,
the first realization is that very little of a context
exists. Bynum et al. (1982) is the only study which
directly examined the impact of property loss and victim
injury on detective decision making; no studies provide
insight into the impact of weapon use on detective decision
making. Bynum et al. did not find property loss or injury
to affect the extent of effort devoted to follow-up
investigations. Therefore, this study supplements Bynum et
al. in finding weapon use to not affect decision making,
provides additional evidence that injury has no additive
effect, and also contributes strong contrary evidence as to
the effect of property loss. The study also provides
insight into the meanings attached to these case
characteristics and thus, how these case characteristics
come to affect (or not affect) the decisions of detectives.

When considering these findings in relation to those of

169

the larger police decision making literature, several
parallels emerge. For example, both studies which examined
the impact of weapon use on patrol officer decision making
(Smith 8 Klein, 1984; Smith, 1987) found that weapon use did
not affect decisions. Similarly, four of five studies which
examined the effect of victim injury on police decision
making (Berke 8 Loseke, 1981; Smith 8 Klein, 1984; Worden 8
Pollitz, 1984; Smith, 1987) found no (direct) effect (but
see Waaland 8 Keeley, 1985). No studies of police decision
making have assessed the impact of property loss on decision
making. The consistencies between this study and studies of
police decision making more broadly in terms of these
offense characteristics suggest that detective decision
making is similar to patrol officer decision making in that
direct effects of weapon use and injury do not exist.

In comparing these findings to those of previous
criminal justice decision making research however, several
inconsistencies do emerge. For example, Adams and Cutshall
(1987) found property loss to affect prosecutorial decision
making. However, Schmidt and Steury (1989) found the extent
of victim injury and weapon use did impact on prosecutorial
decision making. Nagel (1983) did not find weapon use to
affect judicial (pre-trial release) decision making.

The final set of offense characteristics examined in
this study consisted of evidence, or solvability factors,

associated with the crime. Of the evidence variables

170

examined (strength of suspect information, presence of
physical evidence, description of suspect vehicle, knowledge
of vehicle plate number, and identifiability of stolen
property) only suspect information and physical evidence
were found to influence decision making in all three
decision situations. If physical evidence was available,
the report was more likely to be selected and have more time
spent on the investigation. Further, the stronger the
suspect information, the more likely the report was to be
selected and when there was moderate suspect information
more time was spent on the investigations than when there
was weak or strong suspect information. Identifiability of
the stolen property did not display an effect on any of the
decisions examined. The other evidence variables displayed
inconsistent effects. When a suspect vehicle was described,
the report was more likely to be selected. When a suspect’s
vehicle plate was known, the burglary case was more likely
to receive more time.

In the content of search analysis, the "importance
scores" for suspect information and physical evidence
consistently ranked in the top one-third of all scores. In
fact, suspect information ranked either first or second in
all of the decision situations. The remaining evidence
variables achieved importance scores of varying strengths.

The observations support the other analyses which

illustrate the impact of suspect information and physical

171

evidence on decisions of investigators. However, the impact
of the other evidence factors could not be isolated through
the observations. In the selection decision, the strength
of suspect information most often had a direct impact on
decision outcomes. In fact, in reviewing the reports for
the statistical analyses, it was not difficult to
successfully predict those cases which were not assigned --
if a name of a suspect was not provided (or was provided
only on the basis of the victim’s guess), the case was very
rarely selected for an investigation. In order for physical
evidence to be predictive of case selection however,
generally other evidence (albeit minimal) had to be present
as well. For example, a case where a name of a suspect was
provided only on the basis of the victim’s guess (weak
suspect information), and fingerprints were available, would
have had an increased chance of selection. Given this
combination of evidence it would be relatively easy for the
detective to either ”make” or eliminate the named suspect.
Therefore, it was seen that all of the evidence weighed
together (again, in a compensatory style) was more important
than each evidence element considered individually.

In the time allocation decisions, peep "good leads” were
derived from suspect information. If no leads were present
(e.g., there was weak suspect information), then the
detective typically did not realistically expect much chance

of an arrest clearance regardless of the time spent in the

172

investigation. As a result, the investigation was most
often quickly closed through some other means (i.e.,
”NFI"). The only exception to this, as described in Chapter
Five, was when the case was viewed as "serious" or when time
was available; when either of these conditions were present,
the investigation may have still had considerable time
devoted to it (e.g., time was spent cultivating evidence).
At the other extreme, when suspect information was quite
strong (i.e., the name of the accused was provided by
someone who saw the accused committing the crime), little
time was needed to identify the culprit -- perhaps an
interview of the suspect and a visit to the prosecutor’s
office. When suspect information was of moderate strength
however, it was believed that there was at least a
possibility that the case "could be broken" with just a bit
of additional information. These cases were therefore
considered more promising and consequently, time was often
"created" for them. Once again however, strength of suspect
information could have been enhanced by the presence of
other evidence which could affect the overall "strength of
evidence" in the case and ultimately, the amount of time
devoted to the case.

In looking at the detective, police, and criminal
justice decision making literature, it becomes apparent that
evidence (however defined) has a consistent effect on

decision making across all decisions in the justice system;

173

more evidence leads to ”more law" (e.g., Adams 8 Cutshall,
1987; Albonetti, 1986; Schmidt 8 Steury, 1989; Black, 1971;
Eck, 1983; Greenwood et al., 1977; Bynum et al., 1982).
Accordingly, considering the previous discussions in this
chapter, it seems reasonable to conclude that evidence
strength is the one factor on which decisions of all
criminal justice actors turn. At the same time however,
this study empirically demonstrates the seemingly unique
curvilinear relationship between strength of suspect
information and the time allocated to follow-up

investigations.

The Processes of Decision Making

This is the first known study which has attempted to
explicitly trace the cognitive processes associated with the
decision making of detectives, or any criminal justice
actor. Accordingly, this study should be considered as only
a first step toward an understanding of how detectives make
decisions. In the previous section of this chapter, the
statistical results and observations were discussed in
relation to the factors which influence the decisions of
detectives. In this section, the results of the information
board exercise and observations are discussed in order to
cast light on the processes by which detectives transform

information inputs into decision outcomes. Specifically,

174

search behaviors of detectives are discussed in terms of
their linearity, depth, and content.

When detectives were presented with case information and
asked either to decide which cases to select for a follow-up
investigation or determine which case would receive top
priority and rank the rest, it was found that all of the
detectives used, to a large extent, linear decision making
strategies. This result is rather surprising given the
expectations outlined in Chapter Two. A substantial amount
of the process tracing literature has illustrated that
decision tasks characterized by a high degree of complexity
(i.e., the amount of information available -- number of
alternatives, number of dimensions, or both) are associated
with the use of non-compensatory decision strategies (Payne,
1976; Ford et al., 1989). Although task complexity was not
manipulated in this study, the amount of information
available in the information boards provided what would be
considered a relatively complex task in the context of
previous research.

In the bigger picture, the observations also lead one to
believe that decision making of detectives is accomplished
through primarily (but not exclusively) compensatory
strategies. Based on the description of the investigative
process provided in Chapter Five, it was seen that there are
basically five determinations made in an investigation: (1)

Is the case legitimate? (2) Does the victim wish to pursue

175

the complaint? (3) Are there good leads? (4) Is the case
serious or unusual? and (5) Is there enough time to keep
working the case? Each of these determinations is either
directly or indirectly influenced by case information. In
considering these factors, and hence determining the amount
of time to spend on a case, compensatory and
non-compensatory strategies are used. For example,
non-compensatory strategies are used when the victim does
not wish to pursue the complaint or when the case is viewed
as lacking in legitimacy; if the case does not have
legitimacy or the victim does not wish to pursue, little
time is spent on the case. Conversely, compensatory
strategies are used when the case is serious; seriousness
can compensate for lack of evidence ("good leads").
Availability of time (lack of time pressure) can also
compensate for the lack Of evidence (or the lack of strong
evidence); if time constraints are not perceived, the case,
regardless of the amount of evidence or degree of
seriousness, will have time allocated to it (of. Sanders,
1977). Hence, on the basis of the observations of
detectives, one is left to conclude that detective decision
making involves a mix of compensatory and non-compensatory
strategies, with most decisions made through the use of
compensatory strategies.

Regarding the amount of information searched in the
information board exercise, it was seen that the detective
sergeants searched, on average, less information in the

176

selection decision than the detectives in the prioritization
decision. Although determination of statistical
significance is problematic give the small N, a possible
explanation for this finding is that the detective sergeants
have a lesser burden for action. As evident in several of
the protocols, the primary task of the detective sergeants
in reviewing initial investigation reports was to look for
leads which a detective could pursue. If leads were not
available, then the seriousness of the case would have to be
established (vis—a-vis time pressure) before the case would
be assigned. While the sergeants may formulate initial
impressions about the legitimacy of the complaint, it was
not their task to verify its legitimacy. Therefore this
information was not necessary in order to make a
determination as to whether or not to assign the case. The
detective assigned the case, on the other hand, not only had
to determine if there were leads and the seriousness of the
incident, but he also had to be concerned with information
which could assist in making an overall determination of
legitimacy.

Perhaps the most interesting results of the search
pattern and verbal protocol analysis related to the
detective’s content of search. As discussed in the previous
section, victim characteristics were generally considered
"less important" than offense characteristics in the search

for information. More importantly perhaps is the evidence

177

provided in the protocols which supports a "gestalt theory"
of decision making -- with detective decision making, the
sum of the information which relates to a given case is
greater than each of its individual parts. As a
consequence, a case takes on meaning only when the
information relating to the case is considered and weighed
together. This "gestalt theory" also, of course, has direct
implications for the depth of search in decision making. In
Chapter Five this was referred to as ”the more I see, the

more I want to know" phenomenon.

Limitations of the Study

There are several limitations to this study. First, the
data used for the statistical analyses were obtained from
official reports completed by patrol officers and
detectives. Official records have been questioned as an
accurate reflection of objective reality (Manning, 1980;
Meehan, 1986). Particularly troubling for some is the
validity of the self-report measure of time spent on an
investigation. On the basis of the observations, it is
believed that very little error in this measure was
intentionally introduced by the detectives. While
detectives were required to state the amount of time spent
on the investigation, the information was not monitored or
organizationally reported in any fashion. There was no

feedback to the detectives or supervisors on the basis of

178

 

 

ti
on
no
the

end

these statistics. Detectives were not evaluated on this
information. In addition, supervisors rarely questioned the
amount of time stated on the report. Therefore, it appears
that there was little reason to intentionally inflate or
deflate the stated amount of time spent on a case. However,
the measure did contain at least two types of errors --
memory errors and calculation errors.

Concerning memory errors, through the course of a shift,
a detective may have worked on five or more cases,
performing one activity on one case and then another
activity on another case, etc. However, sometimes the case
log would not be completed until the case was closed which
forced the detective to recall the activities performed and
the time spent on each activity. At the other extreme,
where the least amount of memory errors would be reflected,
the detective would record the activity and the time spent
on the activity as the activity was performed. However,
most of the detectives recorded the activities performed and
the time spent on the activities on a daily basis, at the
end of the shift.

Calculation errors resulted from the difficulties
associated with estimating the time spent on various
activities. For example, because of the difficulties in
allocating driving time to particular investigations, it was
most often excluded by the detectives. An exception to this

was when a detective traveled to another jurisdiction in

179

reference to a specific case. Time spent on other
activities which were not attributable to individual cases
(e.g., discussions with other police personnel in the
department, "patrolling" high crime ["slum"] areas) were
generally not included in the summary of activities or time
spent on a particular case. Finally, certain activities
which consumed a relatively insignificant amount of time
(e.g., departmental record computer checks) were often not
recorded on the case log and therefore, were not often
counted in the total amount of time spent on an
investigation.

Second, the observations and protocol analysis made it
clear that several theoretically important variables were
omitted from the analysis, most important an indicator of
time pressure. Although time pressure is a slippery
construct to measure, for the analysis here, a dummy coded
variable could have been sufficient. Given that the
greatest frequency of criminal incidents occur during the
summer months and that this is also when the fewest number
of detectives are available (i.e., vacations), a "crime
occurred" variable could have been constructed with 1 =
crime reported during summer and 0 = crime reported at all
other times.

A third limitation of the study is that the relatively
small case sample sizes did not allow for the quantitative

examination of the interaction and dependency effects

180

discussed qualitatively in this chapter. However, since the
sample of burglary and robbery cases used in this study
represented the population of cases which were reported in a
one year time period, correcting this limitation may have
required performing such a study in a jurisdiction with more
(or at least more varied) burglary and robbery incidents or
perhaps expanding the time frame of the study from one year
to three or four years.

Fourth, several problems emerged from the information
board and verbal protocol analysis methodology. First, as
evident in several of the protocols, decisions of
investigators are based on more information than what was
presented in the information board. Therefore, if this
information was available in the information board, thus
making the decision tasks more complex, subjects may have
used non-compensatory strategies more frequently than what
was observed here. Second, the information board may be
criticized for removing the study of decision making from
the natural work environment and thus, its artificiality.
However, the decision was made to use an information board
rather than hypothetical initial investigation reports in
order to reduce the likelihood of order effects on the
search for information. Based on preliminary observations
of detectives reviewing reports, it was seen that they most
often began by looking at the information placed on the top

of the first page and concluded with the information on the

181

bottom of the last page. Logical perhaps, but with such a
pattern, a meaningful description of content, depth, and
linearity of search would have been very difficult. Third
and relatedly, while the cooperation received from all of
the investigators was outstanding, several of the
investigators seemed, at least initially, uncomfortable with
the format of the board. They were used to a narrative "to
find out what happened.” As a result of this unfamiliar
format, natural decision processes may have been inhibited.
Fourth, as discussed previously, the content of search
importance scores were sometimes misleading especially when
compared to statistics which indicated "influence." As
discussed in Chapter Five, this problem was further
compounded by the frequency in which detectives used
compensatory strategies.

A final limitation of this study relates to its degree
of external validity. Are these results generalizable to
other populations? This remains unknown, an issue perhaps
which should be of concern for future research. Regardless,
given the 100% case selection rate for burglaries and
robberies and the 100% cooperation rate for detectives and
detective sergeants in the information board exercise, the
results of this study would appear representative of the

populations studied.

182

Directions for Future Research

As a result of this study, several directions for future
research have been identified. The first area which could
benefit from further research is the influence of time
pressure on detective decision making. In the observations
as well as in several of the protocols, attention was drawn
to the importance of time pressure in determining whether or
not a case was selected for an investigation (or the
likelihood of a case being selected) and how much time was
to be spent on the investigation. Yet this research, and
other criminal justice decision making research, has not
provided an adequate understanding of this important
constraint on decision making. Previous psychological
research has examined time pressure on decision making and
has generally found that time pressure affects the processes
associated with decision making. However, in such studies,
time pressure is easily defined (e.g., "Subject A had 15
minutes to perform the decision task, Subject B had 5
minutes"). In the natural work context however, time
constraints are more difficult to define and identify.
Hence, future research could examine where time pressure, or
perceptions of it, emanates from in the natural work
environment, and the effects of the perceived time pressures
on the outcomes and processes of investigative decision
making.

Second, during the observations and in the analysis of

183

1

information board search patterns, numerous differences
between detectives were seen. Specifically, in the
observations some detectives were seen as more motivated,
more ambitious, and more skillful than others. In the
search patterns it was seen that some considered certain
information very important while others considered it
irrelevant. While this study provides preliminary evidence
that individual differences exist, future research could
explore in detail the nature and sources of these individual
differences and their effects on the outcomes and processes
of investigative decision making as well as on other
outcomes associated with the investigative process (e.g.,
how cases are closed).

Third, additional research is needed on the proper
measurement of evidence strength. This study made a
significant step by dissecting "strength of suspect
information" and demonstrating the differential impact of
various elements of evidence across decision situations.
With this study as a foundation, future research could.
identify the effect of various combinations of evidence on
investigative decision making thus constructing a more
complete index of "evidence strength."

Finally, since detectives base many of their decisions
upon the information provided in the patrol officer’s
initial investigation reports, and efforts of patrol

officers are related to the amount of information collected

184

during the initial investigation (Eck, 1979), it would be
useful for future research to examine the determinants of
effort and the decision processes of patrol officers during
initial investigations. Accordingly, a study similar to the
one conducted here, except focused on patrol officers and
initial investigations, would add another dimension to our

understanding of the investigative process.

185

Footnotes

Because important details concerning the processes by
which cases were referred to the various investigation
bureaus and the validity of the self-report measure of
time allocation are not provided by Eck (1983), one should
be cautious in drawing inferences from this table.

Victim age is not included as a social structural
characteristic in Black’s theory and therefore, an absence
of a relationship between age and the decision responses
of detectives should not be considered as additional
evidence not supportive of the theory. Because age is a
demographic characteristic (as is sex, race, etc.), it is
included in the discussion here.

In the data analyzed in this study, there were three

(of 292; 1%) robbery cases which involved female victims
over 50 years of age, and fourteen (of 317; 4%) burglary
cases which involved female victims over 50 years of age.

Eighteen (of 292; 6%) robbery cases involved non-white,
male victims under the age of 30. Ten (of 317; 3%)
burglary cases involved non-white, male victims under 30
years of age.

186

APPENDICES

APPENDIX A

Initial Investigation Report Form

 

POLICE DEPARTMENT

INVESTIGATIVE OFFENSE FORM

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g OFFENSE INCIDENT CODE ATTEMPT VIOLATION LEVEL NATURE OF OFFENSE COMPLAINT NUMBER
g DY” DLocaI UFooem
I
‘5 ‘ Duo 05mm DOM
Vest Month Day Tm
w OFFENSE o“ . OFFENSE Dmsun DI2IMon Damn mew
1 h ....... b ........ q ................ q
77 com m b DISITM DIGIFn Urns“ DIGIUmnown
I“
3 -----1h---------_ ........ huh------J mcURRED
REPORTED TO omcsn D: I )DayI-gn: DQIDIIIIness UISNMDM
5 W8 DIRECTION NAME - TYPE BLDG. APT NO FLOOR
5
8 IF INTERSECTION. NAME . TYPE I NAME . 7va
.J
SUBJECT ROLE. VT-chrn CP-Comommn: INT-witness DC-Dmcoveveo Cnme F's-Person Seem-no Promo LP-Last Person In Possess-on
NAME (Last Fm. Macon. Suflu: SEX RACE DOB/Age IDENTIFY INTERVIEW PROSECUTE LIVES ALONE
C] VT
[3 C, Dm Duo Um UNo Elves UNo Um
D m ADDRESS (Nuclear. Duecuoa Name-Tyne. Bloc. Act No. Floor: TELEPHONE (II-ca»:
, 0°C 59 .......................................................................... is: ..................... I...
.a 0 PS Enema (Name Adam: Won ms
2
o 0 LP
Bu_s .
v, "S'é'o .
2 ”0 VICTIM INJURED DI: :No Inuy DIZIPoum. Bu: moo-m DIJINOfl-IOCIDOCIIOIM Ultvncaoacnmq Dana:
E U w NAME (Last First Macon, Sum: SEX RACE Dos/Age IDENTIFY INTERVIEW PROSECUTE LIVES ALONE
S DY“ DNO DY” DNo DY“ DNO DY"
> D C?
UM ADDRESS (Number. Dnecnon Mmefiyse Blog As: No. Floor: TELEPHONE IX'CI.“
0 DC 339 .......................................................................... -102 ..................... I- _I_
0 rs EmoIcyer (Name. Aaovcss: Work m:
U L, Bus
EEO
NO VICTIM INJURED 0.: mo Incury DI2IPowDI¢ Bu: UMnOwn DIJINOmIneaDocuaImq Dunncaoacnamg DEIFam
ROLE NAMEILu: Furs: Mmole Sum-I ADDRESS mumoer Ouectron Name - Tyne Blog, Act No.) TELEPHONE Inn»
SE1
D (:9
0 WT RACE Res Res
0 0C ------------------------------------------------------------------
U PS SEO :Deumv Imam/new 003. A9.
a D L“ .0 UV” DNO 0Y0! 0N0 Emmove' . Bus '
5'1
5 [3 c9 ‘
w [3 wt “(3'
'5 C : I
g D m .331 ......................................... R. 92 ....................
0 U es 550 IDENTIFY INTERVIEW DOB/Ag.
:1 0 LP NO DY” DNO DY" UNO Erosion: Bus '
SEX
: [:1 cr
" D V" RACE
” an Res
5 D Dc ------------------------------------------------------------------
3 D 95 sec :oemurv :msnvnew 003.».
D t' "0 DY“ DNo Um 0N0 Em an; I I
SEX
D or
D wt RACE
Res Res ____t__-
D DC .......................................... . ................
D es 550 05:47:91 mrssvxsw 008. Age ‘
D LP "0‘ BY” DNO Uves DNO Enclave: BIL I
O SCEIE TYPE (cadet—OTHER VICTIM/O‘FENOEB RELATIONSHIP IcooeI___OTHER
9 ASSISTING JURISDICTION W NRIS NCIGNT W MAL OEST VALUE
1 cs: PROCESSED Dm Um ‘
REPORTNG OFFICER BADGE NO DATE COMPLETED TIME
.. PAGE _OF _
a
3 Asslsmc orncsas supsnvusoa supesvnsonv JUDGEMENT .coos.

 

 

 

 

187

INVESTIGATIVE OFFENSE FORM

NOTE 1: ll addmonal welrms. complainants. or subsets are present. use addmonal lnveshgarrve Ollense Forms,

NOTEZ: Mandatory Informatron: All reports rhusl conlarn all Res Gestae Mlnesses. statements. address. and what eacl'l wrtness can «mm; :o.

accused DOB. address. tights (arm and where evldence can be located.

' NOTE 3: Use the Iolrcwlng abbreviations where applicable ”NA" (or Not Applrcable. "REF" (or Relused and "UNK" for Unknown

 

LOCATION; Number Drrecnon Name - Type 8100 Apr No Floor
(examples) 125 w Claremore Dr. 2 H 3
100 81k E Michigan Ave.

If the oflense occurred without a specific address. stale in the 'Number' box the block number as :n the second example. (Blkxalock)

'. If you are usrng a block number as in the second example. use the even numbered block numbers ( I 00. 200 300. etc. l to represent the even numbered Slde ol
the street and the odd numbered block numbers (101. 201, 301. etc ) to represent the odd numbered srde o: Ihe sneer

REPORTED TO OFFICER Relers to the time the ollrcer rs Investrgahng (he complalnt
ATTEMPT. A cnme can only be classrlied 5. an Attempt when al the lollowrng meme-ms exist:

1- A Dhywcal act rs committed (e.g. burglary where ladder :s placed on window or CSC where v:ctlm Is grabbed)
2. The perpretralor's Intentions are stated by the acoused or Ind-cared by physical evadence.
3. The aims Is not completed.

ADDRESS: Include thecrty name i: the address is other than Lansrng.

TELEPHONE It a telephone does no: exlsl. pnnl "NONE" in the space crowded. ll a-Ielephone number is oulslde the I 51 7 l area Code. Include the area code.

SEQ. NO.:_ A unlque number shOuld be entered in the SEO. NO. boxes (or every v:c1:m.sub)ec:. master: or suspect who :5 Wed on the lorms. (1 .2. 3. 4. etc.)

LIVES ALONE Place an "X" in the "yes" bar when the otlense occured In a reSIdence. the vrclrm (was In Ihal reSIdence and the victim 15 the only adult Irvrng In
the TBSIOOOCE Olnerwrse. enter "NA".

SCENE TYPE
ReSIcenIlal

01 Apanmenl

02 Duplmr

03 Single Farmly

06 ReSIdenlIaI Garage
05 Slorage Shed

06 Yard/Lawn/ Driveway

Busmess

07 Amusement/Arcade
08 Appliance

09 Auto Dealer/RV Center
10 Bar/Reslauranl

11 Commerce: Relall
12 Conven-ence Store
13 Drug Store

14 _Flnancral lnslIIulrcns
15 Gas SIaIIOOI Garage
16 Hotel/MOIeI

17 Indoor Recreahonal
18 Jewelry Store

19 Laundromat

20 Cleaners

21 Medical Facrliry

22 Dulce/Business

PATROL SUPERVISORY JUDGEMENT

1 Departmenlal Policy
2 Geographrcal Cucumstances

Industrial

23 Commerclal Storage
24 Constructlon Srle

25 Manufacturing/ Factory
26 Undeveloped Area

27 Warehouse

Publrc Premise

28 Cemetery

29 Church

30 Park/"Playground
31 Publrc Bwldmg
32 Scrlool

Slreel/Parklgg

33 Alley

34 Street/Highway

35 Parkung Lo:

36 Parkrng Ramp

37 Dumpster (arson only)
38 Vehicle (arson only)

39 Other Mobrle (arson only)

OTHER

3 Inablllty to Locate ernesses. VlCIIITIS. Suspects

4 Evadence Results No: Avarlable
5 Absence From Work

6 Omar (EnIer code and descnplron ol other)

‘ 40 Other (use space provided after Scene Type‘ (code)

188

SEX M - Male
F - Female
8 - Buslness
U - UnknOwn

RACE. W - Whlle
8 - Black
,I - Inclan (Amencan)
H - HISDanIC
A - Asmn
U - Unknown

VICTIM . OF=ENDER RELATIONSHIP
A Suspect anc Vlcum are Marned
B Exoscouses

C Suscec: and Vic-Inn are Romanlrcally Involved

0 Parent: Chlld Relationship

E Brother: 5 I r Srstetl s) RelallonShlp

F Other Farmh- Relallonshrp

G Long Term Personal Acoualnlance
H Short Term Personal Acquaintance
.l EmDIOyeermployee

K SelI-mrnczed

L POIICE Olhcer IS VICIIm

M Stranger

N Other (use space provided alter Other)
99 Unknown

SOLVABILITY FACTORS

Was there suspec1( SI arreSIed"

Was Iner'.‘ a wrlness to the crime?

Can the suspect be :denrmed by wrlness"
Can a suspect be named?

Is a suspect descrrbed"

I

Is the susoecr known ahd'cr can he'shc be localed"

Was the": Slgnmcanl MO presenl?
Was mere slgn-Izcan! :nvsmal evrdence present?
I: Ire stolen property :denlrlnable?
(5 men,- RSOHII'CBO’ suscec: vehrcle descnplron" 1
'-.'l': more undeveloped leads? 1
Gm. '. nl Ollcnse 1
VRI'J'.‘ oer $1.000
OP’Y‘TI}! over $1000

Y

1
2

DUUDUUDUUUDU E

3"“ng ."IU'IQS’ ncsollalrzalron recurred

Woormnc :nmlvbd

APPENDIX B

Supplemental Report Form

 

POLICE DEPARTMENT

SUPPLEMENT FORM

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

189

I"
>
p.
<
C
I
<
Z
0 orreuse mcroenr cope VICTIM'S um: names or orrause COMPLAINT wmaen
m
1
sums neeonrmc omega BADGE no. in" munmo TIME
.—
I
8 assented omcens n
coo: __ PAGE or __
8

j

 

APPENDIX C

Modus Operandi Descriptor Form

 

. POLICE DEPARTMENT

MOOUS OPERANDI DESCRIPTOR Fonu‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

POINT OF ENTRY ENTERED THROUGH AAEANS DP ENTRY — PROPERTY POINT OF EXIT EXITED THROUGH SAFE JOBS
tFrent AAIec IPhyaeatForee tFronI AAmc tam
2Rear IVent/AIrContatIon 25mm 2Rear BVent/AIrConoItIoh 2CamadAuay
:ISide JCIabnglnaomlent :Is-oe CDoor JWorIIaoaIScene
A Above DPeuolS-angGlaae A PrvagToela A Above DPetloISlidIngGlaea
5 Smooth E Fee Eaaoe 5 Exptoevea s Beneath E Fire Escape
”um FAttecheoGarage see-anode ”um FAttaehedGaraoe
G We! 7 Dypaaa W G Wal
5 H tntenor We! 99 Unto-own H lntenor We!
III J Whoo- e Other. J window
,1 K Floor K Floor
c L Root L Root
5 M Concadttere 99 Untied-In
‘” 99 Unknown M Other.
N Other.
CODE_ CODE— CODE____ CODE__ CODE__ CODE—
SUSPECT CHARACTERISTICS INSTNCTIONS TO VICTIM MEANS OF ATTACK - VICTIM VICTIM ABUSE
ARanaackaIMuDeat. MDIsableeTelaohoneIAtam lLiaDo-n APhyecalForce IToruaed
DSOIacbveInLoa N Ohms/WM 2EmorCODIOI/Vaul Saddle/mouse 2BlIndIoIded
m CNeatBuruar PanngannContam JWIIegatActs CDelICordIRooe llaouneraggad
D SmaaheGrab R DoeeNotTamuoeone A DoanSale/CaahRegIater 0 Vehicle A Shots Fired
5 E Mods/Removes Pmta S victims Vehicle Used 5 Face We! E firelEmlouve s Stabbed
'5 F Eats/Drinks on Pram-ea T Cut PM wire 6 Enter Vehicle E Firearm S Victim Searched
< 0 Anon/Attempt u UWIIW 7 Abnormal meme G CSGaa 7 Hit with Weapon
g HmtateaICendeUud VUeeaNotee ”W H ”Um
5 J Smom w Jumps OverCountar/Sar I Other J Threat/Verbal e Other.
5 K Mutton Pan-com x Lott Nora Ben-no K entanglement
. LHoetagea Y ThreatenstoKill LCheIncallPeiaon
5 Z Other. I: meted-belt
3! 99 um
In N Other.
3
In
CODE_______ CODE____.._ CODE—_.__.___ CODE______
FIREARM TYPE CARRIED FIREARM IN FIREARM HANDLING CALIBER: GAUGE HANOGUN DESCRIPTION
I Random I Newspaper I Showntovicurn A 22 I Nickel/ChromaISIaInlesa
’ 2 muted/Hand In Pocket 2 Poem/Coat 2 Cocks or Racks Firearm e .25 2 Blue Steel
3 thIeI Shotgun 3 Sell/Waistband J Pointed at Victim C .32 3 Unusual One!
A Served on Shotgm A Holster A Lays Weapon on Caunterraar D 38 A Rusty
5 Other. 5 Boat 5 PutsWeapon to Vicnms Boay E .357 5 Detechve
2 6 Hand 0 Other F A5 99 Unknown
3 7 Other. G 9 mm e Other
g H 410 Gauge
“- .l 20 Gauge
It 16 Gauge
L 12 Gauge
99 Unknown
u Other
CODE.— CODE__ CODE_______. CODE_.. CODE—....—
AREA MOTIVE WMSER OF FIRES SET FUEL SUPPLY HEAT SOURCE
t Halway t Abnormellehavlor 1 OheFlre I FlaruhableLiou-da 1 Matches
2 Mechanical 2 Boredom 2 Male 2 Catriona-eta Seeds 2 Lighter
JSIoraoe JCoverOtherCnma JWWImTrailera 3mm“ JCandIe
A UIIIIty Room A Domestic A Chantal: A Tmng Device
5 Other 5 Fraud 99 um s Elecmcrty
z 6 HiredAraonIet 5 Other. 6 SmolongMatanel
8 7 Protest 7 Cooking Stove
c S Revenge 8 Heating UnIt
‘ 99 Unknown 9 Molotov Cocktail
9 Other: 99 Unkmn
0 Other
CODE __ CODE __ CODE __ CODE CODE _—
neeommc omcsn one To ‘ coumrnr NUMBER
PAGE—...OE—

 

 

 

 

190

APPENDIX D

Personal Descriptor Form

' POLICE DEPARTMENT — PERSONAL DESCRIPTOR FORM

 

NATURE OF OFFENSE

OFFENSE DATE

 

OFFENSE TIME

 

 

LOCATION OF OFFENSE

 

LPD COMPLAINT NUMBER

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SUBJECT ROLE Ds-urwweo suspect DA-ARRESTEE/ACCUSED CI enumwn CI wwssmo reason
NAME (Lea. First. Hm Suffix) DOB. ADDRESS (Number. Name-Type. Bldg. ApL Nd.) TELEPHONE
Res Res
CI JquNILE
AKA Employer/School Bus
HAT SHIRT/BLOUSE SKIRT/DRESS
COAT PANTS/SUCKS SHOES
U FATHER CI GUARDIAN ADDRESS paw TELEPHONE
MOTHER ADDRESS DAY TELEPHONE
DISPOSITION/RELEASED TO:
RACE SEX AGE HEIGHT BUILD SUBJECT WORE
w White M Male 1 o- 9 t A7'oUnner t Thin A CostunsIUnitonn H Malteuumalas only)
B Black F Fennle 2 10-13 2 Very Short (At-5'21 2 Medium B Bag/Cloth wan Eyeholes J Nude/Partialy Nude
l Indian U Unknoiwi 3 lA-I7 3 StionIST-S‘b't 3 Muscular C Ski'Mask K CIoIhesoIOpooanaSes
H Hispersc A 1845 A Medium (Sr-59') A Heavy/Stocky D Stocking Over Heed L Gloves
A Assn 9 26-35 5 TallS'Itr-G’I') Obese E Halloween Mask M Nothing Unueud
U Unknoiwl 6 3645 6 Very TaIIIS'2’~OverI 99 Unknown F False Bead/Mustache 99 Unknovm
7 its-55 99 UnImom G HandkerchiellScut N Other.
9 560
99 Unknown
CODE __ CODE __ CODE _ CODE CODE CODE __ __ _—
FACIAL HAIR HAIR TYPE HAIR LENGTH HAIR COLOR HAIR FIBER TEETH
1 NoFacraI Has 1 Dyed 1 Bald/Thin 1 Black I Wavy 1 Missing
2 Unshaven/Stubble 2 Processed 2 Crew Cut 2 Blond/Strawberry 2 Kinky 2 Protruding
3 Mutton Chops 3 wig/Towed 3 Above Ear 3 Brown 3 Bushy 3 Stained/Decayed
A Mustache A Streakelerosted A Below Ear A Gray/White A Curly A Geld/Silver
5 Coarse s Alro 5 Calls: Length 5 ReoIAubum 5 Str ht 5 Chipped
6 Fu Manchu 6 Pony Tail 6 Shower Length 6 Sandy 99 UhknOwn/NA 9 Capped
7 Eu! Beard/No Mustache 7 Convows/ Braids 7 Longer than ShauIder 7 Brown/Partly Grey 6 Other 7 Nothing Unusual
a Board A Mustache B Naming Unusual 99 Unknown 8 Black/Partly Gray 99 Unknown
99 Unknown 99 Unknown 99 Unknown 9 Other.
9 Other: 9 Other. 9 Other.
CODE _ __ CODE _ CODE ...._.. CODE _ CODE ._ CODE __ _—
EYES EYE COLOR EYE BROW EARS COMPLEXION A COMPLEXION B SPEECH
t False I Brown I Thin t Caulillowu I Llwlt I Pockmarks/Acne I Impedlnierlt/Smners
2 Crossed 2 Blue 2 Bushy 2 Protniding 2 Meaum 2 Moles 2 Accent (American)
3 Sunglasses 3 Hazel 3 Connected 3 Earring 3 Dark 3 Freckles 3 Accent (FWI
A Glassesmlain) A Green A Nothing Unusual A Missing A Albino A Nothing Unusual A Foreign Language
5 Bulging 5 Gray 99 Unknown 5 Nothing Unusual s Reodish 99 Unknown 5 NothingUnusual
6 Scum/Bulk e Pink 5 Other. 99 Unknown 99 Unknown 5 Other 99 Unknown
7 Irregular 7 Unmanned 6 01hr. 6 Other. 6 Other:
B Nothing Unusual 99 Unknown ‘ “
99 Unknown e Other:
9 Other
CODE _......_. CODE CODE __ CODE __ CODE __ CODE __ CODE
SCARS TATOO/BIRTHMARKS (Describe) DEFORMITIES AMPUTEE (Describe) DEXTERITY SUBJECT INJURED
tYes tlnsignia IArm tYes RRI'ght tNoInyury
2 No 2 Pictues/Deslgna 2 Hand 2 No L Let! 2 Possible. but Unknown
It Unknown. Leave Bled: 3 Names/Words 3 Fingers 99 um 99 Unknown 3 Non-Incapacrtanng
A Initials A Torso A Incapacttating
_F.“ 5 Numbers 5 Log 9 Fatal
_Neclt 6 Brahman: 6 Foot
7 None 7 Lime
_Arm 8 Other 3 Nothing Unusual
__Hand/Wnat It Unknown. Leave Blank 99 Unknown
Toy-go __FJC9 _Ftngor 9 0mm:
(.9 ___Aml _Torsd
._Hand _Leg
CODE __ _ CODE __ CODE ._ CODE _
REPORT WRITTEN BY: BADGE NO OTHER OFFICER IN VEH. BADGE NO. DATE-TIME REPOF DESCRIPTION GIVEN BY'
PAG_§__ OF—

 

 

 

 

 

 

191

CR Use Onry (PIN)

APPENDIX E

Vehicle Descriptor Form

 

' l POLICE DEPARTMENT '

VEHICLE DESCRIPTOR FORM

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

VEHICLE ROLE: STOLEN VEHICLE
lg UtSISusdect DIAIArrestee DIUISIdIsn DISPOSITION DISIStdlan a Net Recovered Dram A Recovered in Lansing
g; v v v : DIFIRecdvered FOR Dias-do Juneacsdn
5 REGISTERED OWNER (LaeL Fest. Madras. Sills) TELEPHONE
2 3
32
of. ADOFESSINIM.Name-TydeI.ADLNd.) srAte ZIPOODE
g >
INTERIOR EXTERIOR GENERAL CONDITION BODY DAMAGE WHEELS
IBuctteISeala 7Uredueltatna tPutaadlnecndbdn IPoor tLeh IMags
ZBenchSests_ Basted/Tape 2StIcsarlDecaI 2Fair 2Right 2ChrdrneRlIna
3 CW 9 Floor Shirt 3 Run 3 Good 3 Front 3 Over Bias
ATM scum AVIIwIToa AEaceIent ARear Aweeninis
SEeI-dnierithed dOther. SDeoorPa-tt 99m and sued-wow
to 9 m Missing ' a Related Para 9 No Damage 99 Um
I: 7 Huang Uni“ 99 Ursula-n 9 Other.
0 - 99 Utltldlan
: 9 Otter
E
0
III
0 MODIFIED WINDOWS LIGHTS OUT LDAA CIRCUMSTAICE
:1 I Front I Damaged sidewiiioows s DecauPlaqiis I Len Front Know/Venice UIItYes Drama
2 Rear 2 Damaged W 7 Curtains 2 vatFrdrI
9 auction-up JDthegedRearWlnddw Beta-tom .3LellReer WW Drum 0‘2”“
z A mum A Tinted Unltndwn A Rigrileear Peynlentaw 0(1)Yes thmo
'0 99 Unknown 5 Covered . 9 Other 5 Brake Lights
’ .5 Other. 6 License Plate
99 Unknown
CODE __ COOE__ COOE_____
VEHICLE YEAR VEHICLE MAKE VEHICLE MODEL VEHICLE STYLE (code) COLORS - Too BOIIorh
.:
e
g LICENSE nuueen LICENSE sure ucexse MONTH m rem LICENSE come - Plate Numpers
L“, .
Ff) AUTO Vim MOTORCYCLE FRAME nuueen VEHICLE INSURED ev (UDAA only) nuance COMPANY Iqu only)
In
>

 

NAME (Last First. Midas Suttiitl

 

ADDRESS (Number. Direcson, Name - Type. Apt Nd)

 

 

 

 

 

 

 

 

 

 

 

 

0)
HI
)-
2 YEAR MAKE srne kuotoncvae some nuueen
ADDRESS uou necovenv DISTRICT nuueen pare necoveaeo me RECOVERED
some type (codat__ ' “
treason». BELONGINGS m VEHICLE JURISDICTION men: STOLEN owsroe JURIS COMPLAINT no

 

VEHICLE CONDITION (any A)

DuIStndpad Umlgnitlansm chIMILDesL

BIA) Tdrcned 0(5) No Damage 0(6) Other

 

TOW RECOVERY INFORMATION‘REG. OWNERI

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

192

TOWED RELEASED to ownen rowep BY: WHERE STORED REASON rowep
DYes Duo UYes ONO

oereuomts NAME DOB DEFENDANT‘S mus doe OFFENSE INCIDENT cooe

:
COMPLAINT NUMBER

2 A

neponrmc OFFICER made no. pare oomereo TIME
.-
a:
8 ASSISTING OFFICERS In

”Los—
E

lauuoleAWWI

 

APPENDIX F

Property Form

______‘_. H .... .

 

REPORTING NFICER

 

POLICE DEPARTMENT

PROPERTY FORM

 

COMPLAINT NUMBER

 

PROPERTY CODES. SoSIDIen

R-Recovevaa

HOW RECOVERED CODES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

193

F—Fwna L-Locano In Second Heno Suite 660003 PucrIasec
Li?! 0: Sum J-Reooverea Io: OIheI Arm-on N-NCIC/LEIN C-ConIIsceIeo P-CDmouIDT O-OIneI (Ewan
FROP. CODE BRAND/MAKE NAME SPECIFIC PROOUCI 1‘ TYPE IAOOEL NAME] NUMBER STYLE
MODEL YEAR PRODUCT ID/SERIAL NO OPERATION ID NUMBER LPD NUMBER SIZE COLOR
- LICENSE PLATE NO. YR mo OF PLATE [MM VALUE LOCATION OF PROPERTY wHEN STOLEN YR PURCHASED
UMens
PURCHASED AT VALUE ESTIMATED aY VALL£ (cow INSURANCE COMPANY INSURANCE AGENT
:
g FINANCE WPANY LOCATION OF PROPERTY WHEN RECOVERED WHERE STORED ION REWD CODE
0
E
m. ELISE Fall IEN O-ECK FOR STOLEN GUN REG FEOUEST DESCRIPTION
DYae ONO UNI DNOC DECO DUnreg
m m BRAND/MAKE NAME SPECIFIC PRODUCT/TYPE MODEL NAMEI NUMBER STYLE
MODEL YEAR PRODUCT ID/ SERIAL NO OPERATION ID NUMBER LPO NUMBER SIZE COLOR
LICENSE PLATE NO YR IMO OF PLATE DwM VALUE LOCATION OF PROPERTY WHEN STOLEN YR PURCHASED
OMens
PURCI-IASED AT VALUE ESTIMATED BY VALUE Imam INSURANCE COMPANY INSURANCE AGENT
N
t
g FINANCE COMPANY LOCATION OF PROPERTY WHEN RECOVERED WHERE STORED I-OW ECO/EEO CODE
0
E
m. m Fm LEIN OEO< FOR STQEN GL1! PEG REQUEST DESCRIPTION
DVD! 0ND DHII DNog DR” DUNE;
——. _r #— :—
PRJ’. m BRANDIMAKE NAME SPECIFIC PRODUCTtTYPE MODEL NAME/NUMBER STYLE
MODEL YEAR PRODUCT ID/SERIAL No OPERATION ID NUMBER LPD NUMBER SIZE COLOR
LICENSE PLATE NO. YRJMO OF PLATE mm VALUE LOCATION OF PROPERTY WHEN STOLEN YR. PURCHASED
UM":
PUROIASED AT VALUE ESTIMATED BY VALL£ Imam INSURANCE COMPANY INSURANCE AGENT
H
t I
5 FM COMPANY LOCATION OF PROPERTY WHEN RECOVERED WHERE STORED HOW IECOVEPED C1306
a
O
E
W mm PW LEN O-ECX FOR STOLEN GUN PEG EOUEST DESCRIPTION
Una 0N0 DH: DNeg UReg UUnIeg
F

PAGE—OF

 

APPENDIX G

Follow-up Investigation Case Log

DATE REC’D:

 

DAITICIIT”T:

 

ASSIGNED BY:

‘ POLICE DEPRRIMENT

 

CASE LOG

 

CRIME TYPE:

 

 

INVESEflflflOR:

 

 

 

 

V—_..——.._.._-__._._.
___._____..-_-_._..—.—

I
I
I
I
I

I
I
I
I
I
I
I
I

 

 

 

 

 

 

 

 

 

 

TOTAL THIS: I -_
03 Proseczzto: 09 case Review 14 SLrveillance 16 Bridence
05 azspoena Ser. 10 Case Research 15 Report Writing 21 Seared warrant
03 Court. 11 Scene Lmzes: 16 Meetmgs 25 Aucopsy
06 Mess/'Invo :. 1 Interxnw 17 Arrest
O7 Misc/Other 13 N. canvass
COMMENTS:

EINAL STATUS:

194

APPENDIX H

Case Data Coding Form

INVESTIGATIVE DECISION MAKING CASE DATA FORM

1. P.D. complaint number Computer ID number _ _ _
2. Crime type? 0 = BURGLARY 1 = ROBBERY _

3. Victim type? 0 = INDIVIDUAL 1-8 =BUSINESSES _
4. Victim sex? 0 = MALE 1 = FEMALE 8 = MIXED _

5. Victim race? 0 = WHITE 1 = NON-WHITE 8 = MIXED _
6. Victim age? 00 = MIXED 99 = MIS’G _ _
7. Victim address? / S of census area / _____

 

8. Victim employment status? 0 = NOT 1 = EMPL’D 8= OTHER _

9. Number of witnesses to the crime?

 

 

 

10. Could suspect be identified? 1 = YES 0 = NO

11. Could suspect be named? 1 = YES 0 = NO _
12. Could suspect be described? 1 = YES 0 = N0

13. Strength of suspect information? 1-2-3- -5-6-7-8-9-10 _
14. Was there physical evidence present? 1 = YES 0 = N0

15. Can vehicle be described? 1 = YES 0 = NO _
16. Is vehicle plate # known? 1 = YES 0 = NO

17. Is stolen property identifiable? 1 = YES 0 = N0 _
18. Victim-offender relationship? 1-2-3 _

19. Was a weapon used? 1 = YES 0 = N0 _
20. Degree of injury? 1-2-3

21. Amount of property loss? S 99999 _____
22. Was case selected for f-u investigation? 1 8 YES 0 = N0
23. Detective assigned case? 10-11-12...88...99 _ _
24. Sgt. who screened report? 1-2-3-4-5-9

During the follow-up investigation....
25. Number of victim interviews
26. Number of witness(es) interviewed
27. Number of witness canvasses _
28. Number of others interviewed
29. Number of times crime scene searched
30. Number of physical evidence items submitted
31. Number of times computer searched
32. Number of photo line-ups conducted
33. Number of times mug pictures shown
34. Number of suspect line-ups conducted _
35. Number of suspects (or times) interviewed _
36. Number of informants interviewed _
37. Number of times prosecutor consulted _

 

 

 

u
z
0

38. Does victim desire investigative effort? 1 = YES 0
39. Status of the investigation 1-2-3-4-5-6-7-8 _
40. Total time spent on investigation _ _._

195

APPENDIX I

Information Board: Selection of Burglaries

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, x , w - -m (...... m:
n a; u-hL— _1..u\— o-h ...-.m wflo aflm u .m m - .
57% J .u u.-.“ a
‘ ‘ - ‘ p 7 r1 .
7 . a 01+. ms... (-r «.0
u-.. u-.. w-.. o r a u
..-: _h-r— a-) 1-:
.2-) —.J-r - ,
A A , mm.
a (in
u m .340
20M JIM MIN hIM
at
- ‘ ..N
u u ....N wan a d m N < end
IlN JIN v.-.“ :0 N H N U
_a
0-. (L
- _ . “11m war.
| H 86a 3m. ...
s _ y «158 4.33,. .3. .3} a? x I
rm”, damn 331.3 in.“ fifi 84.5 2.32) EFL». > Emu..> Exp...) E#N.W
, . .u . . . . . .
03a: Nut—mm. but?“ «3.. m +2 ... 9 m 17.11. D U A m
.2

 

 

 

 

 

 

 

 

 

 

196

on .00-o: avenue. snowman. luao«>c

.ouo: mouse-chm vudoa menus-noucu 0:» :« pomsaoca do:

.m:00¢0h “vauddacovaucoo no:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“0002
$1... hm
_ . 1‘
...... ., ... .. ..i. .......... .... ... .7 ...... .... .... 1......_ ....
run-a .
.. I... .. ._ ...: .. ......T m I
_
. 95!: m
.....a-.afiaz.$
. agé. w _ hula . N
O 0 e O r 5»! .m
E..ElfiflE.EEE¢s_
.132 U
hos-L16 ‘UH‘ 111-0 40.4.”... alt-n. . a . all" . and“ 3 K?
4...“... fl... -..... ......m ......n. ..n... ...... _s...._ .. L H3...— _......._ ...... ...}
.2 HH.. 4 5 _IE . e u. u 3. <

197

APPENDIX J

Information Board: Prioritization of Burglaries

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7...... 2...... Jun yd. ...-.... u-u —z-u on... “Th ‘m-v — - u-v. -
2-». I-) J.) a-) h... H-) :-r u-r “T... w-.. a-r u-r hm-.. j+
\‘cN o-N u-N - {-N
77‘ - - (L
wa .952 a? .3 *5
{Jim £44...” £42”. £1... mm- $m+umv

JndU

,— \n
1 11

dadU

m
1:-

Jndu

r1
1-

\anU

JndU

'3

198

APPENDIX K

Information Board: Prioritization of Robberies

.090: v00:0|0hn undo: nodudlhoucq and an vovsuo:« no:
on «null: venue-v :OQOLIQG I«»o«>: ..coadou headdqucokucoo hon "ouoz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 1r? In” ha
in 80:, or. .2 .2 cm x. a I.- Z. .2 1.3 .32 ...: .... .3
0.300 if
m r an“ m, ”J15 *1
.. ...: L. .2 .. .... an... .I ... ..., ...... . ...... Ru .3
:11. $.86
3% o. m m “or .2 .7. «LI UV..." II To .7. 4%. 11? slow 11.3 ms
:3 _11. umcU
a. . n31;
MU) 0m 0 O o O r I'll “W “0} .... , N‘
fl .4 2 a‘ 7 inn“. ~ L w _id 11$ Jgno :10
2...... _
.2 2.. .7 .2 ... ... 1...... .5. s. a, .3. .s. a. .3... ....
...-... w
in . L 14.33 ufl1< .1...» 3.3.1; on< . [Jaha- 315 3w :5.
”......“ HR“ fir...» km... 1......“ Mn...” ....Sa 1.35 1..., -1... 1mm...» $1.5 I»; 1.5
4 2- b H I. u a a. a Q a. a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

199

APPENDIX L

Information Board Instructions for the
Selection of Burglaries

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I | | I o h'
\-n or. ...-.... if. JG 3...“ n-h a...» :-.u can Th w .u a .u u m m m < m 426
u- .. <.r r1
KI? Oar 5.0%- II) J.) In? bu? Ha} In? “IT “a? WI? 0.? 6.? 3‘0
o-n ma
a-m. o-m Ian (A J-n 2.” h-» a.» ...... ¢.m u-m u-m u.m 9n <-m
JndU
Q-" 01-." 1.!“ {Inn JON xi" bl" H|N 1|" “|~ ‘0" “0" at." u-“ a-” (‘N IN‘
dnUU
Q|\ 0" 2" II\ I-I\ 1" b0\ “i‘ 1|. *l— ‘I- wnl- OI- dI- ”In (|— —*
JndU
. gm atest 3 1.3 31. , . d v if
151.. is: 2...}. 1.3.. w. in .149.» 13!. In“ 13.33 at} 40¢. Y!» .0 HPW-m in um ...
J? (A! It?) 30mm.”- SJ .1!“ in!“ S o ‘33) 1V.) $41; .83....) ‘1'!) c5101) 16.)
WW. 2 I J v. q H ... a ..v a

200

.050: uo»:onoun taco: :oqudluou:« 0:» an tovauocu 90:
on Ann-d: cocky-v cououvca l«»o«>z .ucoulou huaalnacovmucoo act "0002

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9: .2 .2 .2 3%, _flm— .....1 :4 .2 1.} .3; ...: 3.3
5.8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Egggamygggig

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

92+. .7 .2 .2 _mur‘ “MVM I «3 _ :4— fl—.§2_ _anm_ 73.3.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Eéigzm-agagi

 

 

 

 

 

 

 

 

[I

 

 

 

 

 

 

 

 

 

 

 

é

 

E
[2
BE

 

 

 

5‘— 31-3 ’XI!) 8:- Sh,
u‘l 61 6“ C3" C11

 

 

 

 

 

 

 

 

 

 

 

201

 

 

 

 

 

F1;

 

 

 

 

 

 

 

. . ...: a . 3
O? 02 a I." .1 cm n n 1% fl .0 :
_ L 21+: 3 . 1......» 39.1! 4?. . “Ham 3 3m. “I.

 

3}.
{in

 

 

 

 

 

 

 

 

 

 

 

INSTRUCTIONS

This exercise provides a simulation of the decision of
whether or not to assign a burglary case to a detective for
a follow-up investigation. You should imagine that you
have five cases to either "CPU" or "OPA." Your task is to
decide which case(s) you would most likely assign to a
detective.

Presented here are thirteen different types of information
you can consider in choosing which of five cases to assign.
For example, you can look at the age of the victim, the
nature and amount of suspect information, the dollar value
of the stolen property, etc., all of which are typically
contained in the initial patrol report. Your task is to
uncover and look at as much information as you need to, and
then decide which of the five cases should be assigned.

While reviewing and deciding upon the cases, it would be
most helpful if you "think aloud." State the information
you are looking at and considering. State what you are
thinking as you look at the information. State any other
information which would help you in making the decision.
Say anything that comes to your mind. Finally, identify
the cases you would assign.

All responses will be anonymous.

Thank you very much for your participation.

202

APPENDIX M

Depth of Search: Formula and Examples

Depth of Search

The following formula was used in calculating depth of
search:

depth of search = NDA
TND

 

where:

NDA = the number of dimensions accessed for a given
alternative

TND = the total number of dimensions available to be
accessed for a given alternative

The following is an example of how the equation is
computed (note: an "x” indicates an accessed dimension):

Dimensions
l 2 i 1 Q
1 x x x x x
Alts 2 x x
3 x x x
4 x x x x
for...
Alternative 1: NDA=5; TND=5; depth of search = 1.0
Alternative 2: NDA=2; TND=5; depth of search = .4
Alternative 3: NDA=3; TND=5; depth of search = .6
Alternative 4: NDA=4; TND=5 depth of search = .8
In this example, the mean depth of search across all
alternatives = .7 or 70% of information was accessed.

203

APPENDIX N

Content of Search: Formula and Examples

Content of Search

The procedure outlines below was used to calculate
"importance scores."

In assigning scores to each dimension within each
alternative, the following conventions were used:

TND = total number of dimensions available to be
accessed for each alternative

TND‘ = the first dimension accessed in each
alternative

TND-l = the second dimension accessed in each
alternative

TND-Z = the third dimension accessed in each

alternative, etc.
A score of "zero" indicated that the dimension was
not accessed in that alternative

Accordingly, in an information board with four alternatives
and five dimensions, the following search pattern could
result:

Dimensions
l 2. Si :1. i
l 0 5 4 3 0
Alts 2 0 4 5 3 2
3 0 5 4 0 0
4 1 4 5 3 2

where, for example, in Alternative 1, Dimension 2 was
accessed first, 3 second, and 4 third. Dimensions 1 and 5
were not accessed.

To determine the overall "importance score" for each
dimension, the mean of the assigned scores was calculated.
Therefore:

 

 

Dimensions
1 2 .3. A Q
1 0 5 4 3 0
Alts 2 0 4 5 3 2
3 0 5 4 0 0
4 1 4 5 43 _Z_
X .25 4.5 4.5 2.25 1

In this example, dimensions 2 and 3 are tied as "most
important," followed by dimensions 4, 5, and then 1.

204

APPENDIX 0

Linearity of Search: Formula and Examples

Linearity of Search

The following formula was used in calculating the

degree of linearity in decision making (from Gilliland,
1990):

Degree of Linearity = NA
((DS * AU) - (D8 + AU - 1))

where:

NA = the number of times a standard dimension was not
accessed on a given alternative that had at least
one standard dimension accessed

DS = the number of dimensions accessed in the standard
alternative

AU = the number of alternatives used in the
comparison, including the standard alternative

The rationale for the components of the equation is as
follows:

1.

The numerator gives an indication of the degree of
dissimilarity between the standard and those
alternatives accessed on at least one dimension of the
standard. Alternatives are limited to those accessed
on at least one dimension of the standard because
perfect linearity can exist even when all alternatives
are not accessed.

The multiplicative component of the denominator gives
the size of the matrix examined for linearity.

The additive component of the denominator adjusts the
denominator for those elements that do not add into the
numerator. Specifically, the number of dimensions in
the standard are excluded because they never add into
the numerator. Additionally, one dimension of each
alternative will never add into the numerator because
each alternative must be accessed on at least one
dimension to be included.

205

 

The following are examples of how the equation is computed

(note: an ”x" indicates an accessed dimension; 0 = perfect

linearity, 1 = perfect non-linearity):

1. Dimensions
1 2 3 4 5
1 x x x x x
Alts 2 x x x x x
3 x x x x x
4 x x x x x

NA=O; DS=5; AU=4; Linearity Score = 0

2. Dimensions
1 2 3 4 5
1 x x x x x
Alts 2 x
3 x
4 x

NA=12; DS=5; AU=4; Linearity Score = 1

3. Dimensions
1 2 3 4 5
1 x x x x x
Alts 2 x x
3 x x
4 x x
NA=9; DS=5; AU=4; Linearity Score = .75
4. Dimensions
1 2 3 4 5
1 x x x x x
Alts 2 x x
3 x x
4 x x
NA=9; DS=5; AU=4; Linearity Score = .75

206

LIST 01“ REFERENCES

 

LIST OF REFERENCES

Abelson, R.P. & Levi, A. (1985). Decision making and

decision theory. In G. Lindzey & Aronson (Eds.),

 

flethgg. (pp. 231-309). New York: Random House.

Adams, K. & Cutshall, 0.8. (1987). Refusing to prosecute
minor offenses: The relative influence of legal and
extralegal factors. 1g§31§g_ggaztgglx, 5, 595-609.

Albonetti, C.A. (1986). Criminality, prosecutorial
screening, and uncertainty: Toward a theory of
discretionary decision making in felony case
processings. Criminology, 21, 623-43.

Baldus, D. Pulanski, C. & Woodworth, G. (1983). Comparative
review of death sentences: An empirical study of the
Georgia experience. 1ggLn§1_g£_gzin1nal_hag_§nd
92.111.321.931. IA. 661-85.

Beach, L.R. & Mitchell, T.R. (1978). A contingency model for
the selection of decision strategies. Academy of
WW. 3. 439-49.

Berk, S.F. & Loseke, D. (1981). Handling family violence:

Situational determinants of police arrest in domestic

disturbances. Law and Society Review, 15, 314-44.

207

Black, D. (1980). The Menneze and Cnetoms of the Police. New

York: Academic Press.

----- (1976). The Behavior of Law. New York: Academic
Press.

----- (1971). The social organization of arrest. §Len£eng
Lew Beview, 23, 1087-111.

----- & Reiss, A.J. (1970). Police control of juveniles.
American Socielegieel Review, 35, 63-77.

----- (1968). Police Encomntere and Social Oggenization:
An Observation Study. Ph.D. Dissertation, University of

 

Michigan.
Bloch, P.B. & Bell, J. (1976). Managing Investigatione: The
Rochester System. Washington D.C.: Police Foundation.
Block, R. (1974). Why notify the police: The victim’s
decision to notify the police of an assault.

Criminelogy, 1.1. 5558-69.
Brooks, W.N. & Doob, A.N. (1975). Justice and the jury.

Jomrnal of Secial Ieenee, 31, 171-82.
Burnstein, 1., Kelly, W., & Doyle, P. (1977). Societal

reaction to deviants: The case of criminal defendants.
WM. .42.. 743-56-

Bynum, T.S., Cordner, G.W., & Greene, J.R. (1982). Victim
and offense characteristics: Impact on police
investigative decision making. Qyiminelegy, 20, 301-18.

Clarkson, G. (1962). Eoyttelio Seleetion; A Singletien 91
Trust Inveetment. Englewood Cliffs: Prentice Hall.

208

Davis, K. (1969). Rieenetionany gestice. Baton Rouge:

Louisiana State University Press.

deGroot, A.D. (1965). Thought and Choice in Chess. The
Hague: Mouton.

Doherty, M.L. (1987). The Effeets 0; Cognitive Complexity
and lash Complemity on Decisien Making Behavior. M.A.
Thesis, Michigan State University.

Eck, J.E. (1983). Solving Crimes: The Investigatien of
erglary and Robbery. Washington, D.C.: PERF

----- (1979). Managing Qase Aesignments: The Runglary

Investigetion Decieien Regel Replieetien. Washington,
D.C.: PERF.

Einhorn, H.J., Kleinmuntz, B.M., & Kleinmuntz, B. (1979).
Linear regression and process tracing models of
judgement. Reyeholegicel Review, Rh, 465-85.

Ericson, R.V. (1981). Rating Crime: A Study of Detective

Renh. Toronto: Butterworths.

Ford, K., Schmitt, N., Schechtman, S., Hults, B.M. &
Doherty, M. (1989). Process tracing methods:
Contributions, problems, and neglected research

questions.

 

Proceeses, 13, 75-117.

Forst, B. & Brusi, K. (1977). A theoretical and empirical

analysis of the prosecutor. lsurnsl_gf_Lessl_§tudies. 6.
177-91.

209

Frazier, C.E., Bock, E.W., & Henretta, J.C. (1980).
Pretrial release and bail decisions: The effects of
legal, community, and personal variables. Criminology,
lg, 162-81.

Fredrich, R.J. (1977). The lmpact oi Opganizational,
Individual, and Situational Factore on Rolice Behavior.

Ph.D. Dissertation, University of Michigan.
Gaines, L.K., Lewis, B. & Swanagin, R. (1983). Case
screening in criminal investigations: A case study of

robberies. Police Studies, Q, 22-9.
Gilliland, S.W. (1990). Time Conetpeints, Infopmetien
Redundancy, and Deeision Rehaviop; A Ryeeeee Izeeing

Approach. M.A. Thesis, Michigan State University.
Gottfredson, M.R & Gottfredson, D.M. (1980). Reeieien_uehing
in Criminal Justice: Toward the Ratienel Eyereiee pi

Discretion. Cambridge: Ballinger.
----- & Hindelang, M.J. (1979a). A study of The Reheviey of
hen. American Seeielogieal Review, 11, 3-18.

----- (1979b). Response: Theory and research in the

sociology of law. Amepiean Seeielogicel Review, 41,
27-380

 

Greenberg, B. Elliot, C.A., Kraft, L.P., & Procter, H.S.

(1977). Eelony lnvestigetion Deeisien Rogel; An Annlyeie

of lnyestigative Rlemente of lnfopmation. Washington,
D.C.: GPO.

Greeno, J.G. (1976). Indefinite goals in well structured

problems. Peyehglegicel Review, R3, 479-91.
210

Greenwood, P.W., Chaiken, J.M. & Petersilia, J. (1977). The

QLiminal.1nxestisat12n_fizgge§§- Lexington: D-Co Heath.

Grudin, J. (1980). Processes in verbal analogy solution.

Rmman Pepception and Pepformence, R, 67-74.

Hastings, T.F. (1980). Criminal investigation. In Garmire,
B.L. (Ed.), Loeel Qevernment Relice Renegement.
Washington, D.C.: International City Management
Association.

Heinz, A., Heinz, J., Senderowitz, S. & Vance, M. (1976).
Sentencing by parole board: An evaluation. lenpnel_et
Criminal Law and gpiminelogy, fil, 1-31.

Hembroff, L.A. (1987). The seriousness of acts and social
contexts: A test of Black’s theory of the behavior of
law. Amepicen Jonrnal oi §oeiolegy, 23, 321-47.

Hindelang, M.J. & Gottfredson, M. (1976). The victim’s
decision not to invoke the criminal justice process, in
WoF. McDonald (Ed.), Qriminsl.iustise.end_th2_!lgtimo
Beverly Hills: Sage.

Hogarth, B.M. (1974). Process tracing in clinical
judgement. Reheyiepnl_§eienee, lg, 298-313.

Holmes, M.D., Duadistel, H.C., & Farrell, R.A. (1987).
Determinants of charge reductions and final dispositions
in cases of burglary and robbery. lenpnel_efi_Reeenpeh_in
Crime and Relinguency, g1, 233-54.

Isen, A.M. & Means, B. (1983). The influence of positive
affect on decision-making strategy. §eeiel_§pgnitien, _,
18-31.

211

Johnson, A.C., Johnson, M.B., & Buse, R.C. (1987).
Econometries:lReeic ang Applied. New York: Macmillan.

Jones, C. & Aronson, E. (1974). Attribution of fault to a
rape victim as a function of the respectability of the
victim. Jour a f Perso slit and Soci P cho ,
2R, 415-19.

Just, M.A. & Carpenter, P.A. (1976). Eye fixations and

cognitive processes. Cegnitive Reyehelogy, R, 441-80.
Karmen, A. (1984). C ime V'ct' s' c ' o

Victimology. Pacific Grove: Brooks Cole.

King, G. (1986). How not to lie with statistics: Avoiding
common mistakes in quantitative political science.
American Journal oi Political §cience, R9, 667-87.

Kuykendall, J. (1982). The criminal investigation process:
Toward a conceptual framework. lenpnel_e£_gyimlnel
Jnstiee, lg, 131-45.

LaFave, W. (1965). st: e Dec’ '0 e us e t

t u to . New York: Little, Brown and Company.

Laub, J. (1981). Ecological considerations in victim
reporting to the police. J a ' ' e,
lg, 419-30.

Lipsky, M. (1980). - v u a c ' ' s o t
lngivignal in Rublie Service. New York: Russell Sage.

Lundman, R., Sykes, R.E., & Clark, J.P. (1978). Police
control of juveniles. Jo i i e

2121113932921. 15. 74-91 .

212

Manning. P-K. (1980). Ihg_usrgls_fismei_92sanizatignal_and

1nI2raaLi2nal_LimiLa_2n_Qrus_La2_Enfgrgemento Cambridge:
MIT Press.

----- (1978). Book review of: Detective Work (by W.B.

Sanders). Sociology end Reeiel Reseepeh, g2, 654-57.
----- (1977). ic W ° 1 ' ' f

Belieing. Cambridge: MIT Press.

Mastrofski, S. & Parks, R.B. (1990). Improving observational
studies of police. Qyiminelegy, 2R, 475-496.

Meehan, A.J. (1986). Recordkeeping practices and the
policing of juveniles. Ephen_Li1e,‘l§, 70-102.

Miller, F. (1970). Rreeecution; The neeieien te ghetge e
Suspect with a gpime. Boston: Little, Brown, and
Company.

Nagel, I.H. (1983). The legal/extra-legal controversy:
Judicial decisions in pretrial release. Len_eng_§eeiety
Revieg,,l1, 481-515.

Newell, A. & Simon, H.A. (1972). a vi .
Englewood Cliffs: Prentice Hall.

Olshavsky, R.W. (1979). Task complexity and contingent
processing in decision making: A replication and
extension-.QrsanizaLignal_§2hsxigr_snd_flnman
P fo ma e, 24, 300-16.

Onken, J., Hastie, R., & Revelle, W. (1985). Individual

differences in the use of simplification strategies in a

complex decision making task. lenynel_efi_finpeyimentel
Psychology: Rumnn Pereeption end Eepieymenee, ll, 14-27.

213

Payne, J.W. (1976). Task complexity and contingent
processing in decision making: An information search and

protocol analysis. Or a i 10 e v' d
Rerfopmance, 16, 366-387.

----- & Ragsdale, E.K.E. (1978). Verbal protocols and
direct observation of supermarket shopping behavior:
Some findings and a discussion of methods. Advences in

o u e earc , 5, 571-577.
Platt-Jendrek, M. (1984). Sentence length: Interactions

.2.

with race and court. 0 of Cri 'na J t' ,
567-78.

Pogrebin, M. (1976). Some observations of the detective
role. Jonrnal of Rolice Scienee end Adminietyntion, 1,
277-84.

Reid, 8. (1989). Criminel Law. New York: Macmillan.

Sanders, W.B. (1977). Retective Work: A Study ef inminel
Investigations. New York: The Free Press.

Schatzman, L. & Strauss, A.L. (1973). Eield Reeeepch;
Strategiee tor a Retmpel Seeiology. Englewood Cliffs:

Prentice Hall.

Schmidt, J. & Hochstedler-Steury, E. (1989). Prosecutorial
discretion in filing charges in domestic violence cases.
Cyiminology, 21, 487-510.

Sherman, L.W. (1980). Causes of police behavior: The

current state of quantitative research. Journal e!

Research in Crime and Delinguency, l1, 69-113.

214

Skogan, W.G. & Antunes, G.E. (1979). Information,
apprehension and deterrence: Exploring the limits of
police productivity. Jo a o imi 1 Ju t e, 1,

217-41.

Skolnick, J.H. (1975). Justice Without Tpial: Law
Entorcement in Democratic Society. New York: John Wiley

and Sons.

Smith, D.A. (1987). Police response to interpersonal
violence: Defining the parameters of legal control.
Social Fopces, gt, 767-782.

----- & Klein, J.R. (1984). Police control of interpersonal
disputes. Seeiel Rpobleme, 31, 468-81.

----- & Visher, C.A. (1981). Street-level justice:
Situational determinants of police arrest decisions.
Seeial Ryoblems, 22, 167-177.

Spohn, C., Gruhl, J., & Welch, S. (1987). The impact of
ethnicity and gender of defendants on the decision to
reject or dismiss felony charges. gylminelegy, at,
175-191.

Sudnow, D. (1965). Normal crimes: Sociological features of
the penal code in a public defender’s office. Seeiel
Rpehleme, 12. 255-76.

Svenson, O. (1979). Process descriptions of decision

making.

 

R3, 86-112.

216

Sommers, I. & Baskin, D.R. (1990). The prescription of
psychiatric medications in prison: Psychiatric versus
labeling perspectives. Jnetice Quapterly, 1, 739-755.

Thompson, J.D. (1967). Opganizations in Action. New York:

McGraw-Hill.
Tversky, A. & Kahneman, D. (1974). Judgement under

uncertainty. Seienee, 1R , 1124-1131.

0.8. Department of Justice (1988). o t t N i
Crime and gheti e. Washington, D.C.: Bureau of Justice
Statistics.

Visher, C.A. (1983). Gender, police, arrest decisions and
notions of chivalry. Cpiminology, g1, 5-28.

Von Hirch, A. & Hanrahan, K.J. (1979). The Qmeetion ef
Repele. Cambridge: Ballinger.

Waaland, P. & Keeley, S. (1985). Police decision making in
wife abuse: The impact of legal and extralegal factors.
Law and Human Behavior, 3, 355-65.

Waegel, W.B. (1982). Patterns of police investigation of
urban crimes. Jo nal o P 'c c e a
Administretien, lg, 452-65.

----- (1981). Case routinization in investigative police
work. Social Probleme, RR, 263-75.

Welch, S. & Spohn, C. (1986). Evaluating the impact of

prior record on judges’ sentencing decisions: A seven

city comparison. Justice Quarteply, 3, 389-407.

216

 

Williams, K. (1978). e Ro e of e V' ti ' e

Rroeecutien e; Violent Otieneee. Institute for Law and

Social Research. Publication Number 12. Washington,
D.C.: GPO.

Willmer, M. (1970). Crime and lnfoymation Theoty.
Edinburgh: University of Edinburgh Press.

Wilson, J.Q. (1978). v ' a ors‘ a
Rapeoties Agente. New York: Basic Books.

----- (1968). Venieties of Peliee Behevior. Cambridge:

Harvard University Press.

Worden, R.B. (1989). Situational and attitudinal
explanations of police behavior: A theoretical
reappraisal and empirical assessment. Leg_end_§eeiety
Review, 23, 667-711.

----- & Brandl, S.G. (1990). Protocol analysis of police
decision making: Toward a theory of police behavior.
American Qenpnal ef Cpiminel Jnstice, lA, 297-318.

----- & Pollitz, A.A. (1984). Police arrests in domestic
disturbances: A further look. aw c e v' w,
lR, 105-19.

Wright, P. (1974). The harassed decision maker: Time
pressure, distractions, and the use of evidence. lenpndl

WW1. 5.2. 555-61 .

217

 

 

 

 

(ll Willi? Illll...||'f.|‘| ill]! . ‘
P1] fill-lull!) Ill.

 

"‘Wx‘fiifiiufimfifili\\\ii\’\i\\\\\\\\\«191M

 

31293007896