U.S. Department of Justice 

Office of Community Oriented Policing Services 

DOJ Seal

COPS Logo

Reducing Theft at Construction Sites: Lessons From a Problem-Oriented Project 

by 
Ronald V. Clarke and 
Herman Goldstein 


www.cops.usdoj.gov 


REDUCING THEFT AT CONSTRUCTION SITES: 
LESSONS FROM A PROBLEM-ORIENTED PROJECT 

by 
Ronald V. Clarke and Herman Goldstein 


Table of Contents 

Table of Contents 

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 

Background to the Project. . . . . . . . . . . . . . . . . . . . . . . . . . . .7 

Defining the Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 

The Nature and the Scope of the Problem . . . . . . . . . . . . . . . . . . . .13 

Risk of Break-ins and Builder Security . . . . . . . . . . . . . . . . . . . . 15 

Selecting and Implementing the Preventive Measures . . . . . . . . . . . . . . 19 

Monitoring Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 

Evaluating the Preventive Measures . . . . . . . . . . . . . . . . . . . . . . 25 

Summary of the Results of the Experiment . . . . . . . . . . . . . . . . . . . 31 

Follow-up Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 

Lessons of the Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . .35 

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 

Notes to Annex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 



Abstract 

A building boom in Charlotte, North Carolina led to sharp increases 
in the number of kitchen appliances stolen from houses under 
construction. This paper describes a problem-oriented policing 
project, extending over a period of more than two years, that was 
undertaken by the Charlotte-Mecklenburg Police Department to 
address the problem. A detailed analysis of security practices and risks 
of theft was made for 25 builders operating in one of the police 
service districts north of the city. This produced the recommendation 
that installation of appliances should be delayed until the new owners 
had taken up residence, thus effectively removing the targets of theft. 
Twelve of the larger builders agreed to experiment with this approach 
for a period of six months, though systematic checks made by police 
throughout the period found that builder compliance was variable. 
Despite this, analysis showed that delayed installation was an effective 
policy. Appliance theft declined in the district and there was no 
evidence of displacement of thefts to surrounding districts. The 
concluding discussion of the difficulties encountered by police in 
undertaking problem-oriented projects focuses on the vital role of 
crime analysis and considers ways to strengthen analytic capacity in 
police departments. 



Acknowledgments 

This project was completed under an Advancing Community Policing 
grant (#97-0C-WX-0060) made to the Charlotte-Mecklenburg Police 
Department (CMPD) by the U.S. Department of Justice Office of 
Community Oriented Policing Services. 

Most of the work on this project was carried out by Officers Daniel 
Cunius and G. Eric Rost of the CMPD. They identified the problem 
of theft from construction sites as warranting attention in the course 
of their routine patrol work. They subsequently nominated it for more 
intensive examination and, when the study was approved, compiled 
much of the needed data, acquired information from builders, 
participated in all stages of the analysis, monitored implementation of 
the agreed-upon response, prepared, and presented the findings of the 
study. Throughout all stages of the project, they were strongly 
supported by their captain, E. Charles "Chuck" Johnson, who also 
participated in all stages of the analysis, in engaging the builders, and 
in presenting the findings of the study. Technical assistance in the 
compilation and mapping of the data was initially provided by Carl 
Walter, who also acquired and integrated data from other county 
agencies. Ryan Jackson succeeded him as the crime analyst assigned to 
the Charlie One district acquired and integrated the data on 
certificates of occupancy and developed the computer program for 
recording compliance. Michael Humphrey provided the technical 
support for the final stages of the project. Steve Ward, senior 
Assistant District Attorney assigned to work full-time in the CMPD 
took a close interest in the project and attended many of the project 
meetings. The write-up of the project was greatly facilitated by the 
detailed notes taken on several of the early project meetings by 
Officer Lisa Carriker. 

Dennis Nowicki, the former Chief of Police of the CMPD, initially 
encouraged taking on this project as a case study in problem-oriented 
policing. In its final stages, the project has received strong support 
from the new and current chief of the CMPD, Darrel Stephens. The 
Director of Research and Planning, Dr. Richard Lumb, was generous 
in his arrangements for allocation and scheduling of staff time, as was 
Acting Director, John Couchell. Finally, James LeBeau of Southern 
Illinois University provided us with valuable statistical advice. 



Introduction 

Problem-oriented policing was initially advanced as a way of focusing 
attention on the effectiveness, rather than just the efficiency, of the 
police. Advocates of problem-oriented policing contend that it is not 
enough to respond, however efficiently, to incidents as they occur. 
Rather, with effectiveness as the goal, it is essential that the police 
identify patterns in the incidents they routinely handle, subject these 
patterns (labeled problems) to in-depth analysis, and explore new ways 
of intervening earlier in the causal chain so that these incidents are 
less likely to occur in the future. These new strategies are not limited 
to standard police responses that traditionally depend on law 
enforcement–i.e., on efforts to identify, arrest, and prosecute 
offenders. Rather, without abandoning the use of the criminal law 
when it is likely to be the most effective response, problem-oriented 
policing encourages a broad exploration of other potentially effective 
responses, alone or in partnership with others, with a high priority on 
prevention. Thus, by expanding the repertoire of possible responses 
and settling on a strategy that has the potential for reducing the 
problem, the ultimate and steadfast goal is to increase effectiveness 
(Goldstein, 1979, 1990). 

Problem-oriented policing is rooted in the increased awareness, 
reinforced by continually accumulating studies, that substantial 
categories of crime have been generally resistant to traditional policing 
methods. For example, car and foot patrols can do little to deter crime 
occurring in private places. Crackdowns rarely have lasting effects on 
street crimes. Stakeouts produce few arrests when crimes occur at 
extended intervals. Fast response is usually of limited value when the 
offender has departed the scene. Criminal investigation is too 
laborious and unproductive for all but a minority of serious offenses. 
And the arrest and prosecution of minor property offenders is often 
not productive, given the overburdened court systems and the 
unavailability to a judge of effective sanctions or alternative forms of 
disposition. These limitations sometimes lead to labeling many of the 
categories of crime at which the cited strategies are directed as "non-
suppressible." 

The crimes addressed in this project, thefts from residential 
construction sites in Charlotte, North Carolina, would fit the 
definition of non-suppressible crimes. As shown below, they were 
inherently difficult to deter and they did prove resistant to 
conventional police methods. However, they were reduced as the 
result of a problem-oriented project undertaken by the Charlotte-
Mecklenburg Police Department (CMPD). 



The initial purpose of the project was to illustrate a full, careful 
application of the problem-oriented policing concept. Once the 
decision was made to focus on theft from construction sites as the 
illustrative problem, the second purpose of the project was to enable 
the CMPD to deal with the problem more effectively by making use 
of the problem-oriented methodology. This paper reports on what 
was learned about the problem, describes the effort to reduce it, and 
presents data showing that the response implemented was effective. In 
the language of situational crime prevention (Clarke, 1997), this 
response would be classified as a form of target removal. Thus, this 
paper not only documents the success of the CMPD in using the 
problem-oriented methodology to reduce thefts from construction 
sites. It contributes as well, albeit modestly, to the literature on target 
removal. 

As the project evolved, it provided another benefit of potential value 
to the larger world of policing in the lessons learned about 
implementing problem-oriented policing. In name, problem-oriented 
policing has become quite popular, but the number of efforts that 
meet the original criteria of the concept is very small (Clarke, 1997a, 
1998; Goldstein, 1994a,b, 1996,a,b; Read and Tilley, 2000; Scott, 2000; 
Scott and Clarke, 2000). This is especially puzzling because the 
fundamental logic in problem-oriented policing is often quite 
appropriately described as simple or just plain common sense (Read 
and Tilley, 2000). This project afforded a unique opportunity to 
identify some of the factors that account for the small number of full 
applications of the concept–to identify the factors that, in practice, 
make a seemingly simple process complex. The opportunity was 
unique for three reasons: 

1. The ambitious nature of the project; 
2. The fact that the project had more than the usual support from 
within a police agency; and 
3. The familiarity that the authors, involved as we were in the 
project, had with the concept (Goldstein with problem-oriented 
policing and Clarke with situational crime prevention–a concept 
that in many ways parallels and complements problem-oriented 
policing). 


A major objective in preparing this paper, therefore, was to report on 
some of the complexities that were found in carrying out problem-
oriented policing, to identify the barriers that were encountered, and 
to outline some measures that need to be taken if the benefits of 
problem-oriented policing are to be more fully realized. 



Background to the Project 

Wherever construction is underway, there will be related problems of 
theft, but these may assume significance for a particular police 
jurisdiction only during a construction boom. The form taken by the 
thefts will depend on the nature of the construction which can range 
from enormous projects for new highways and airports to in-fill 
housing developments in suburbs. Construction site theft seems never 
to have been systematically studied, though occasional discussions of 
the problem can be found in trade journals such as "Construction 
Equipment," "CONTRACTOR.mag.com," and "Constructor," the 
latter being the house journal of the Associated General Contractors 
(AGC) of America (Constructor, 1999; Goldman, 1999; McGreevy, 
1999; Snyder, 2000; Stewart, 1998, 2000). These discussions tend to 
focus on the organized theft of expensive equipment such as 
bulldozers or backhoes (commonly known as JCB's in the UK). 
However, this literature draws attention to at least three other forms 
of construction site theft: theft by workers of tools and materials; 
after-hours pilfering of lumber and other materials by opportunist 
thieves, perhaps for their own use; and, in the case of homes under 
construction, thefts by habitual offenders and others of fixtures and 
appliances. 

The present project started with a focus on the general problem of 
theft from construction sites, but, in the classic pattern of problem-
oriented policing, it quickly became more tightly focused on just one 
of the specific sub-problems identified during analysis. This was the 
theft of household appliances, such as ranges (cooking stoves) and 
microwaves, from newly completed houses. 

The project was located in the Charlie One service district, the 
geographically largest of twelve such districts of CMPD, an agency 
that provides police service to the City of Charlotte and to most of 
Mecklenburg County in which the city itself is located. Charlie One 
covers most of the northern part of the county, an area with an 
estimated population in 1995 of just over 100,000. The southern part 
of the Charlie One district includes the Charlotte campus of the 
University of North Carolina and various office, mall, and light 
industrial developments. To the north, the district is largely rural with 
a scattering of separately incorporated small municipalities and 
lakeside developments. As a result of the booming regional economy, 
and because of its proximity to Charlotte, northern Mecklenburg 
county has experienced a large increase in population throughout the 
1990s, and, as a consequence, a correspondingly large increase in 
housing. About one third of the residential construction occurring in 
the entire area served by the CMPD was in Charlie One. 



This construction was mostly in the form of single-family homes built 
in separate developments or subdivisions (estates in the UK), which 
were once farms or fields. In March 1999, 66 housing developments 
were underway, involving 48 different construction companies. This 
wave of building was expected to add more than 12,000 homes to the 
housing stock by 2002 and, by 2010, it is estimated that the population 
of northern Mecklenburg County would grow to about 180,000. This 
would be an 80 percent increase in population over a fifteen-year 
period. 

During 1998, it became apparent to Captain E. Charles "Chuck" 
Johnson, who was in charge of Charlie One, that construction site 
theft in the district was a large and growing problem. For example, of 
the 485 commercial burglaries recorded in Charlie One during 1998, 
109 (22 percent) were break-ins to houses under construction with an 
appliance taken.1 Leaving aside the sheer volume of construction, it 
was not difficult to understand the reasons for this emergent crime 
wave in the district. The numerous subdivisions were scattered 
throughout a largely rural district, which made it next to impossible 
for the police to provide adequate patrol coverage at high-risk periods 
–in the evenings and weekends–when sites were usually deserted. 
Because of the high costs, few sites employed security guards or off-
duty police and, to encourage casual inspection by potential buyers 
(and because it would have been difficult to secure entrances), the sites 
were generally left open. This both afforded the opportunity and the 
excuse for thieves openly to prowl for targets. When the first residents 
moved into their homes, they afforded only minimal guardianship of 
nearby properties because, in the evening, night, and weekend hours, 
thieves could easily blend in with the employees of sub-contractors 
who sometimes worked in those hours. During the day, the large 
number of employees of sub-contractors, casual laborers on site, and 
delivery personnel made it difficult to maintain site security. Tools and 
materials were constantly at risk of theft. Finally, the large number of 
construction firms operating in the district and the large number of 
site supervisors employed could mean that thieves might be able to 
find ready purchasers for some of the appliances and home fixtures 
they stole. Altogether, in the terminology of routine activity theory 
(Felson, 1998), the construction sites ensured the convergence of 
many suitable targets for theft, an absence of capable guardians, and a 
ready supply of likely offenders. 

1 Burglary of a home under construction is 
classified, until it is occupied, as commercial.

Two of Captain Johnson's officers, Daniel Cunius and G. Eric Rost, 
had been taking an interest in the problem and, on the basis of 
discussions with them, he assigned them part-time in May 1998 and 
then full-time in March 1999 to develop a workable solution. Their 
plan had three components: 



• Contacting all existing building site supervisors to discuss their 
crime prevention practices, provide them with crime prevention 
tips, and obtain after-hours contact numbers 
• Establishing community watch schemes whereby new residents in 
subdivisions would be urged to report any suspicious vehicles or 
people 
• Undertaking intensive patrols of the construction sites during the 
evenings and working closely with other officers and investigators 
to identify and arrest suspects 


Elements of this plan–such as the move to a primarily proactive, 
preventive mode–reflected wider efforts that were being made at the 
time to establish problem-oriented policing within the CMPD. With 
grant support from the Office of Community Oriented Policing 
Services in the U.S. Department of Justice (the COPS Office), Chief 
Dennis Nowicki persuaded Goldstein to review these early efforts. As 
a result of that review, Goldstein argued that the CMPD could better 
develop its commitment to implementing problem-oriented policing if 
resources were focused on just a few projects in which a more 
intensive effort was made to address a specific substantive problem. It 
was suggested to him that, from among the many projects then 
underway, the Charlie One project on construction site theft would be 
a suitable candidate for the kind of intensive project he had proposed. 

Accordingly, about six months into their project, Goldstein met with 
Captain Johnson and the two officers involved. The efforts they had 
already made to obtain a detailed picture of the problem, the 
relationship they had cultivated with the department's crime analysts, 
and most important, the enthusiasm and openness they demonstrated 
were impressive, and it was therefore decided that the project merited 
the kind of concentrated attention that had been proposed. The offer 
to make their explorations the subject of a more intensive project was 
welcomed by Captain Johnson and the two officers. Despite the 
considerable efforts they had made to arrest offenders through 
intensive patrols, stakeouts, and working with investigators, little 
tangible progress had been achieved. Both offenders and the stolen 
property seemed to vanish into thin air. In 1998, less than two percent 
of reported construction site thefts were cleared and, while the 
clearance rate was improving (it rose to about six percent in 1999), it 
was still at a low level. The few offenders arrested (most of whom 
were drug addicts) had refused to divulge how they disposed of goods 
they had stolen. In extensive checks of area pawnshops and flea 
markets, not one appliance had been recovered out of the 159 stolen 
from houses under construction in 1998. The Charlie One team was 
anxious to consider any new ideas for dealing with the problem that 
might surface from the type of in-depth inquiry that had been 
proposed. 



Soon after this first meeting at Charlie One, Goldstein invited Clarke 
to join him in advising the project. Clarke's help was sought 
particularly with the analysis and in identifying possible preventive 
measures. This was the beginning of a collaborative effort that 
extended over two years–the time frame greatly influenced by the need 
to collect data over an extended period and competing demands on 
the time of some of the project team members (the team now 
consisting of Captain Johnson, officers Cunius and Rost, and a crime 
analyst–a position filled during the project in succession by Carl 
Walter, Ryan Jackson, and Michael Humphrey). Goldstein and Clarke 
met with the team regularly in a series of short visits. The team was 
frequently joined in these meetings by Steve Ward, an Assistant 
District Attorney with senior status in that office, who had been 
assigned to work full-time with the CMPD. Conscious of the need to 
reduce the pressures on the criminal justice system, his constant 
involvement in the project proved vital to its success. The role played 
by Clarke and Goldstein was essentially consultative–analogous to that 
of coaches–to explain the process of problem-oriented policing, to 
help talk through problems, to raise points for further inquiry or 
action, and to make suggestions about data analysis. This work 
encompassed five main areas: 

.. refining the focus of the project 
.. obtaining a better understanding of the problem 
.. calculating risks (and relating these to builder practices) 
.. selecting and implementing a workable solution 
.. assessing its effectiveness. 



Defining the Problem


In the initial meetings, much time was spent on defining the problem. 
It was decided early on to concentrate on sites being developed for 
housing and not to cover other kinds of construction sites. While 
relatively few, these other sites presented a greater variety of problems. 
Even the residential sites alone presented quite a variety of theft 
problems. While thefts of heavy equipment were rare, it appeared 
from the officers' analysis of crime reports and from talking to 
builders that there were at least three other common forms of theft: 
theft of lumber and building materials, theft of tools, and theft of 
appliances from houses under construction. 

It was decided to concentrate on the latter, theft of appliances, not 
because this was the largest problem (the procedures for reporting and 
recording of construction site theft did not permit this to be 
determined), but for several reasons. It was a costly and common 
offense. It appeared that there were some potentially effective ways in 
which to secure the houses under construction. The large size of 
appliances and the fact they carry serial numbers afforded the 
possibility of retrieving those that had been stolen. The other forms 
of theft seemed more intractable. Lumber and building supplies are 
often scattered around outside and they are almost impossible to 
identify as stolen. This is also true of tools. Even when they have 
serial numbers, the builders seemed not to keep records of them. 

Added to these practical considerations, it seemed probable that if 
thefts of appliances could be reduced, then those of other less 
valuable property might also decline. This is because the increased risk 
or difficulty of stealing appliances might dissuade offenders from 
coming to the construction sites and picking up whatever else they 
could during their visits. 

Once the decision to focus on appliances was made, the Charlie One 
team embarked on a closer study of the incidents. They quickly found 
that many of the thefts that they knew to have been reported had not 
been recorded under the correct category in the system used to collect 
crime statistics. This discovery required that they pull all the original 
incident reports of theft from construction sites and re-code those 
involving appliances–a major undertaking. Compared with the 55 
incidents of break-ins to houses under construction reported officially 
for 1998, they found 109, nearly double the number. They repeated 
the exercise for later years with similar results. 

 

The team also found that thefts were particularly concentrated on 
ranges (cookers in the UK), microwaves, and dishwashers. Of the 414 
appliances stolen during 1998 and 1999 in the entire Charlie One 
district, 34 percent were ranges, 26 percent were microwaves, and 22 
percent were dishwashers. The remaining 18 percent were distributed 
among washing machines, dryers, refrigerators, ovens, stovetops, range 
hoods, air conditioners, and garbage disposal units. Discussions with 
builders revealed that appliances that were hard-wired (i.e. that were 
directly wired into the electricity supply, were attached to plumbing, or 
were built into kitchen cabinets) were less likely to be stolen than plug-
in appliances. This suggested that the thieves were not particularly well 
organized or determined, which was consistent with the fact that at 
least those apprehended had been habitual offenders with a problem 
of addiction. 

As a result of these findings, the focus of the project was narrowed 
further to thefts of plug-in appliances from houses under 
construction. 


The Nature and Scope of the Problem 

Narrowing the focus of the project resulted in questions being raised 
about how seriously the problem, which consisted of just over 100 
reported breaks-ins in the Charlie One district in one year (1998), 
ought to be taken, and whether it warranted the time that it was 
planned to devote to it. In particular, did the number of thefts present 
an unacceptably high risk of break-ins, given the volume of 
construction in Charlie One? To answer this question, the crime 
analyst initially assigned to the project (Carl Walter) sought to find 
information about the number of houses under construction. He 
obtained data from the county building inspector's office showing that 
building permits for 3,130 houses were issued for Charlie One in 
1998. On the basis of this figure, he calculated that break-ins resulting 
in the theft of an appliance were experienced by 3.3 percent of houses 
under construction (104/3,130 x 100). This was little higher than the 
risk of reported burglary (2.8 percent) for all households in America 
in 1995–the latest available comparative data (Farrington and Langan, 
1998). 

Another way to calculate risks is per builder, of whom there were 
several dozen operating in Charlie One during 1998. A risk of 3.3 
percent translates into a risk of one break-in for every 30 houses. 
Only the contractors building as many houses as this per year (of 
which there were 25) could expect to be a victim of a break-in 
involving theft of appliances. The problem would in all likelihood be 
concentrated among the largest contractors in this group. There were 
eight who were building more than 100 houses per year (accounting 
between them for 82 percent of the building permits issued). 
Assuming a reasonably equal distribution of risk, each of these 
contractors might expect to suffer a minimum of three such break-ins 
per year. The largest contractor, who was issued 385 permits for the 
year, might expect to suffer 11 break-ins. Did these numbers represent 
an unacceptably high level risk? 

The answer to this question depends partly on the cost of break-ins, 
which, according to an analysis undertaken by Cunius and Rost of 
police reports for 1998, averaged just over $750 per incident. Of this 
amount, 66 percent represented the replacement costs for the stolen 
appliances, 13 percent was accounted for by the value of other items 
stolen at the same time, and 21 percent consisted of the cost of 
damage repairs. 

Direct costs of $750 per break-in are not high given the retail price of 
new houses in the area (averaging about $140,000) and given the one 
percent of the price that contractors are reported nationwide as 

 

generally budgeting for theft and other losses. In fact, Cunius and Rost 
had learned from talking with site supervisors while on patrol that the 
costs of break-ins were of comparatively little concern to most of 
them. Only the small builders, who saw their profits being eroded, 
were seriously concerned about the loss of the appliances and the 
costs of repairing the damage. For others, when losses climbed above 
budgeted amounts, these could be passed on to future customers in 
the form of higher prices. 

Often more damaging than the direct costs of break-ins would be the 
administrative costs involved in putting the matter right: ordering a 
replacement appliance, being available to accept delivery, scheduling 
repairs, making a police report, and reporting for insurance purposes 
or for tax write-offs. Sometimes, theft of an appliance and a delay in 
replacement might also delay a house closing, with associated financial 
penalties. These indirect costs resulting from administrative action and 
other consequences of theft can easily account for "anywhere from 
two to ten times more than direct costs" (Constructor, 1999, page 1). 

Another intangible cost of break-ins mentioned by some builders was 
that the neighborhood in which they were building might begin to 
acquire an unsavory reputation for crime, which could reduce demand 
for their houses. However, the concern did not seem to be justified 
given the large population growth and the relatively low rates of theft. 
Moreover, thefts at this stage in the development of a neighborhood, 
absent occupants, are not as productive of continuing fears as they 
would be in an established neighborhood. The reputation would 
evaporate with occupancy. Altogether, this analysis of costs gave little 
reason to think that an appeal to profits would help persuade builders 
to take preventive measures that were burdensome or expensive. 
While disappointing to the Charlie One team, this information was 
helpful in thinking about future action. 

A major factor, internal to the police department, supporting 
increased attention to the problem was the contribution that the 
number of thefts made to the overall crime rate for the district. In 
varying degrees, district commanders and their personnel are held 
accountable for the incidence of crime and especially for substantial 
increases in crime. The increase in thefts from construction sites was a 
big negative for a district that prided itself on the traditional measures 
of its performance. Moreover, the high volume of cases, along with 
the sense of frustration and futility in dealing with them, had become 
a source of annoyance to operating personnel. Thus, the desire to do 
something about theft from construction sites was probably stronger 
within Charlie One than from any source outside the CMPD. 

 

Risk of Break-ins and Builder Security 

A benefit of calculating costs was that this focused attention on 
individual builders, their losses, and their security practices. Studying 
other aspects of risk was less fruitful. Hot spot analysis showed that 
break-ins tended to be concentrated in the southern part of the 
district, but this is where most construction was occurring. 
Construction in the more northerly parts of the county was in smaller, 
more up-scale developments. Assuming that most of the thieves were 
habitual offenders taking advantage of the opportunities presented by 
the construction, they would have to travel further to reach these 
developments. Moreover, more of the appliances in these 
developments were built into the cabinetry in these more expensive 
homes. This made them both more difficult to steal and more difficult 
to sell. 

Analysis by time of day, day of week, and month revealed patterns, 
such as an apparent rise in the spring and early summer and a heavy 
concentration around weekends, which were not unexpected or which 
were difficult to interpret because of small numbers. Three makes of 
appliances (GE, Whirlpool, and Kenmore) accounted for about 75 
percent of all appliances taken but these were also the most 
commonly installed appliances. In few cases were more than one 
house broken-into and the appliances taken. Most often just one 
house would be targeted, which again suggests a low level of 
organization and planning. 

There was a tantalizing suggestion that the amount of force used by 
thieves to gain entry varied among the different builders–thereby 
suggesting that some were more diligent about locking up–but small 
numbers of cases made it difficult to be sure of this point. 

It was assumed, from the outset, that there would be substantial 
variation among builders in the risk of break-ins, and early analysis 
supported this assumption. But by this time in the project, the 
limitations on using building permits as a measure of the volume of 
construction in computing rates of theft were becoming more 
apparent. Building permits measure only planned construction, not 
that which is actually occurring. Thus a builder might apply for 
permits to construct 100 homes in a particular subdivision, but only 
start to build a portion of the houses in that year. Indeed, very few 
houses in the subdivision might reach the stage of construction when 
appliances were installed and thus be at risk of break-ins. 




Fortunately, the crime analyst involved in the project at this time (Ryan 
Jackson) learned that a Certificate of Occupancy (CO) had to be 
issued by the county before a new owner could take up residence. He 
also established that lists of COs could be provided which could be 
sorted according to builder and subdivision. Since these were only 
issued when a house was completed and ready for occupancy, they 
provided a much better, more timely basis for computing risk than the 
building permits.2 

2 The team was not aware at 
first that these data existed 
and moreover could be obtained 
in a usable form. In undertaking 
problem-oriented and situational 
prevention projects, it is often 
found (by asking the right 
questions, by probing,and 
pressing) that needed data are 
kept by other agencies and that 
these will be released even 
if pressure has to be applied.

Calculations of the risk of break-ins based on COs substantially 
changed the project team's picture of the problem. Swayed by the 
large number of incidents reported by some of the large builders, 
the team had been assuming that lack of concern about break-ins on the 
part of these builders was one source of the problem. The more 
accurate measure showed that there was still considerable variation 
among builders in risks, but not necessarily of the same pattern they 
had previously discerned. The revised analysis revealed that while 
some of the large builders did in fact have a high rate of theft, others 
among them–like the smaller builders–experienced little or no theft. 
Accurate documentation of this variation considerably strengthened 
the team's hand when later they began to engage the builders in 
discussions about changing their practices. 

While the laborious process of improving the measurement of risk 
was underway, Cunius and Rost were undertaking surveys of builder 
security practices. These surveys were undertaken at various times in 
the project, formally and informally, covering a variety of practices. 
The subjects that were covered gradually changed, as the needs of the 
project became clearer. Early on, the officers gathered information on 
the use of gatekeepers, security guards, or off-duty police officers; 
fencing and site-entrance gates; posted reward programs for 
information about offenders; and the use of temporary burglar alarms. 
Later, they focused on the use of large dumpster-style locked 
containers on site for the storage of appliances prior to installation, 
thereby reducing opportunities for theft; removal of the door from 
appliances to make them less tempting targets; and delayed installation 
of appliances until close to occupancy. The surveys were coupled with 
explorations of individual builders' attitudes toward possible new 
initiatives, such as establishing neighborhood watch schemes for new 
sites as houses began to be occupied, the use of electronic tracking 
devices concealed in appliances, and the use of video cameras to 
monitor the sites after hours. 


It proved difficult to make precise determinations of security practice 
for particular builders because of the broad discretion allowed site 
supervisors. Thus, a particular supervisor might choose not to follow 
company policies (where these existed) about receiving and storing 
appliances prior to installation. It was also difficult to link security 
practices to particular sites, because supervisors quite frequently were 
reassigned to other sites or they left to take up employment elsewhere. 
Consequently, security practices at a particular site might change 
overnight without the contractor and his top managers becoming 
aware of the change. Indeed, Cunius and Rost were frequently able to 
inform senior managers that their beliefs about site security did not 
match what was actually occurring. This became a problem later in the 
project when builders were supposed to be changing their practices to 
conform to police advice. 


Selecting and Implementing the Preventive Measures 


When the improved risk measures had been developed, thought was 
given to repeating the surveys of contractor practices in a more 
systematic manner in an effort to identify best practices. By then, 
however, the rather intensive mulling over of the newly acquired data 
led, rather naturally, to a growing consensus among team members as 
to what would constitute the most desirable workable solution. This 
was to delay installation of the appliances until the new owners 
occupied their new home. Adoption of this practice by the industry 
would effectively remove the opportunity for theft since the owners 
would provide capable guardianship of the appliances. Very few 
ordinary residential break-ins involve theft of appliances since these 
are difficult to carry away and there are many other more tempting 
targets for theft in occupied homes, such as small electronic items, 
jewelry, and cash (Clarke 1999). 

The surveys of builder practices had shown that some smaller builders 
who were delaying installation of appliances reported no thefts, but 
this might have been because they were building too few houses to 
become statistically at risk. Some site supervisors for the larger 
builders also said they did this on occasion, but there was no real 
consistency of practice. In fact, many builders were at first hostile to 
the idea. Sales personnel believed that having all of the appliances in 
place made a home more saleable, and that the new owner wanted to 
see appliances in place before completing the paperwork for purchase. 
They also believed that the absence of appliances, if attributed to 
theft, might unnecessarily alarm purchasers about the kind of area 
they were moving into. Site supervisors felt that the logistics of 
accepting delivery and installing appliances individually as houses were 
occupied were considerably more complicated than batch delivery and 
installation. Some erroneously believed that COs would only be issued 
by the building inspectors if appliances–including the plug-ins–were in 
place. Others said that this was a requirement of obtaining a 
mortgage. Finally, individual installation would mean that builders 
could not take advantage of the greater likelihood of being able to 
arrange for tightly-scheduled building inspectors to visit a site and 
issue COs wholesale. Given the bottlenecks sometimes caused by the 
unavailability of inspectors, this was a substantial concern. 

Instead of turning away from this proposed solution and investing 
more in exploring other means for securing appliances, the project 
team decided that the objections to delayed installation of appliances 

 

should be discussed with builders to see if answers could be found.3 
On deeper exploration, the problems proved not to be as intractable 
as they first seemed. Not all sales people were opposed to the idea of 
delaying installation until the new owners had occupied the premises, 
as long as the reasons for this could be properly explained. For some 
managers, the difficulties and extra costs of individual delivery and 
installation might be offset by the reduction in thefts and damage and 
in the resulting delays in closing. The COs could be issued without 
installation of the plug-in appliances, though problems would remain 
in the case of hardwired appliances. Even in the case of the latter, 
building inspectors thought these problems could be resolved. It was 
unlikely that difficulties would arise with the mortgage companies so 
long as a CO had been issued. 

3 This exemplifies the need for an iterative 
relationship between analysis and response. As 
soon as a promising response is identified, its 
costs and benefits need to be further analyzed. 
The alternative of comprehensively exploring all 
available response options runs the risk of 
losing momentum. 

A decision was therefore made at the end of 1999 to seek agreement 
from the Charlie One builders to institute a policy of delaying 
installation of plug-in appliances until the new owners occupied the 
house. It might be difficult to sell this idea to every builder, but the 
project team would focus on a group of the largest and, to date, most 
cooperative builders. Builders not involved in this natural experiment 
would serve as controls and their break-in rates could be compared 
with those who agreed to the change of policy. 

Accordingly, a presentation was developed, which the project team 
would take to each builder whose agreement was being sought. The 
presentation, which served as a vehicle for succinctly summarizing the 
results of the project to date and which was fine-tuned after critiques 
by Clarke and Goldstein, consisted of several parts: an introduction to 
the nature and severity of the problem; a description of policing 
efforts that had been made to deal with the problem; an explanation 
of why these had met with relatively little success and the change of 
direction to a problem-oriented policing project; solutions considered 
but rejected; the need for assistance from the builders and the 
proposed solution; the expected benefits for builders, police, and local 
communities; and the plan for monitoring and reporting on the results 
of the experiment. 

It was recognized early on that it was important to gain the 
endorsement of key officials before going to individual builders and to 
the local builder's association. The presentation was therefore made to 
the City Manager, County Manager, City Council Public Safety 
Committee, the District Attorney, and the Chief Building Code 


Administrator. Their endorsement was incorporated into the 
presentation. 


In seeking the builders' agreement to the experiment, they were told 
that: 

.. Hardwired appliances would be exempted 
.. The CMPD would publicly acknowledge responsibility for the 
delay in appliance installation 
.. The experiment would begin on May 1 and end on October 31, 
2000 
.. Compliance with the agreement would be monitored by police 
through checks made on houses close to completion 
.. If the experiment did not succeed in reducing the rate of break-
ins, the experiment would be discontinued 

As builders agreed to participate4, their names were added to the list 
of those supporting the proposal when the presentation was made 
subsequently to other builders. An important part of the presentation 
was a bar chart enabling builders to compare their break-in risks with 
those of their competitors. (Prepared specifically for each 
presentation, it did not include the identity of competitors.) The 
presentation and especially the bar charts invariably impressed the 
builders with the depth of the police analysis and, according to the 
project team, helped persuade many to participate in the experiment.5 

4 Some of those unwilling to participate were 
large national builders. This suggests that the 
buy-in of outside interests in a problem-oriented 
project would depend on several identifiable 
factors (1) whether the interest (or business) has 
local roots or is from outside the city–with 
corresponding degrees of commitment to the 
welfare (i.e., level of crime) in the local 
community (2) local customized operations 
versus a cookie-cutter approach developed by a 
national corporation to maximize efficiency and, 
therefore, profits, and (3) possibly the size of the 
private enterprise. Regarding the latter, a small 
contractor might not find it as easy to absorb 
the cost of losses as does the large contractor. 

5 It is awkward and somewhat intimidating for 
police to approach contractors (and other 
business) people to get them to take steps that 
will reduce crime. Police officers fear being told 
that reducing crime is their job and that they 
should go about doing it. Asking others to act 
requires exposing the limited capacity of the 
police. Moreover, police officers are supposed to 
be apolitical, and sensitive to their use of 
coercion in their work. Pressing to engage 
others, from their perspective, draws them into 
activities some see as improper. The project 
demonstrates the power of data in equipping 
the police to engage others. The police often put 
themselves in the position of complaining that 
groups or individuals are shirking their 
responsibilities. But these complaints don't get 
them anywhere and can put police officers in an 
embarrassing light. The opportunity to present 
rigorously developed data should be seen as a 
major new weapon. 


Monitoring Compliance 

Ten builders agreed at once to participate in the experiment by 
delaying installation of appliances until occupancy. Two other builders 
who had been asked to participate, but who had not formally agreed, 
were found from the beginning to be complying with the request to 
delay appliance installation. When asked again if they would agree to 
be formally included, the builders consented. This increased the total 
of participating builders from the ten agreeing from the outset to 
twelve. These twelve builders accounted for about 35 percent of 
construction in Charlie One in 2000. 

As an added measure, the builders also agreed to post decals, to be 
furnished by the CMPD, on doors announcing in English and Spanish 
that "Appliances are not in this house!" Cunius and Rost were to 
monitor the extent to which these participating builders complied with 
the agreed measures by making regular checks on houses nearing 
completion to see whether they contained appliances and displayed 
decals. These compliance checks would be made on all houses in the 
district, not merely those of the builders who agreed to participate in 
the experiment. This was decided because it was anticipated that some 
of the non-participating builders, constituting a control group of 
sorts, might decide to adopt the measures when they heard about 
them. Indeed, some of them might already have been delaying 
appliance installation. Without knowing which of the controls were 
doing so, it would be difficult to interpret the results of the 
experiment. 

Houses were selected for checks as soon as they reached the pre-
completion stage; i.e. kitchen cabinetry was installed, the houses had 
windows that could be secured, and doors that could be locked. 
Appliances were rarely installed before this stage, which was reached 
three or four weeks before completion. The officers would walk 
around each house to check for decals and would attempt to enter it. 
If they found it properly secured and they could not enter, they would 
look through the windows of the kitchen (always on the ground floor) 
to see whether appliances were present. For each house, they would 
record whether (1) it was at pre-completion stage, (2) it was secured, 
(3) decals were posted, and (4) target appliances were present (i.e., 
plug-in appliances and any appliances that were to be hardwired, but 
had not yet been installed, thereby making them easy targets for theft). 

This task was made easier by a data entry program loaded on their 
laptop computers devised by the crime analyst (Ryan Jackson). This 
program not only sped up data entry, but also enabled reports of the 
results of each round of checks to be provided to each builder. This 
served to remind participating builders of their agreement and to alert 
them to possible reneging on the agreement by site supervisors. By the 
time that the experiment was completed on October 31, Cunius and 



Rost had completed 15 rounds of compliance checks involving a total 
of 8,050 separate checks on individual houses–a truly enormous data 
gathering exercise. 

It quickly became apparent from the checks that some of the 
participating builders, who had agreed to post decals and delay 
appliance installation, were failing to comply. In fact, few of them 
made much use of the decals and one large builder was found to have 
appliances present in about 80 percent of the checks made at the pre-
completion stage. Constant reminders, in the form of the officers' 
statistical reports on their levels of compliance and personal 
approaches by the officers failed to correct this situation. As expected, 
some builders, not formally included in the experimental group, were 
also found to be delaying appliance installation (though they were not 
installing decals because these had not been supplied to them). In 
other words, the boundaries between the participating and non-participating 
builders had become blurred. 

The final tally of checks is enumerated in the Appendix. Builders are 
listed according to the percent of checks at pre-completion in which 
no target appliance was found to be present (i.e., in order of their 
compliance with the recommended practice). Throughout the six 
months of the experiment, the average compliance was 78 percent for 
the twelve participating builders; the same computation was 43 
percent for non-participating builders. Compliance did not improve 
much after the first few weeks of the project. Implementation of the 
preventive measures was therefore only partly successful. There were 
still many houses in which appliances were present at completion 
stage, even for participating builders. Of the houses completed in 
Charlie One during the six months of the experiment, about 41 
percent (an estimated 745) contained target appliances during the 
vulnerable pre-completion stage. In other words, considerable 
opportunity remained for appliance thefts. 

Given that so many houses in Charlie One were still at risk of 
appliance theft, although significantly fewer than before, one could 
expect only limited success from the experiment. Burglars would still 
be able to find appliances if they were prepared to search a little 
longer. At the same time, it could be expected that builders who 
delayed installation of appliances until occupancy (whether officially 
participating in the experiment or not) would reduce their rates of 
burglary. This preventive measure is a form of target removal, which 
has been found to be effective in numerous contexts (Clarke, 1997b). 
The best-known examples relate to cash reduction programs in 
convenience stores (Hunter and Jeffery, 1997), betting shops in 
Australia (Clarke and McGrath, 1990), and buses in the United States 
(Chaiken et al., 1974; Stanford Research Institute, 1970). In each case, 
target removal in the form of cash reduction resulted in substantial 
declines in robbery. 


Evaluating the Preventive Measures 

Accordingly, the evaluation of the experiment sought answers to three 
questions: 

1. Did builders, whether or not they agreed to participate in the 
experiment, who made it their practice to delay installation of 
appliances until occupancy, significantly reduce their risks of 
appliance burglaries? 
2. Did builders who made a formal commitment to delaying 
installation (participating builders), but did not always do so, have 
lower rates of appliance burglaries than the non-participating 
builders? 
3. Overall, did the experiment bring about a reduction of appliance 
burglaries in the Charlie One District? 


1. Relationship Between Burglary Risk and Delayed Installation of 
Appliances 

If builders never installed appliances before occupancy, they would 
obviously never lose these appliances to burglary while the house was 
under construction. Unfortunately, only a few small builders always 
followed this practice (accounting between them for a total of only 19 
completed houses during the experiment–see Annex). Given this fact, 
it became important to explore whether the risk of burglary was 
proportionate to the degree to which builders followed this practice 
(whether or not they had agreed to participate in the experiment). This 
question was examined by sorting the 59 builders covered by the 
compliance checks–irrespective of whether they had committed to the 
experiment–into three groups: 20 builders with low percentages of 
houses (<17 percent) in which targeted appliances were present at the 
pre-completion stage; 20 builders with a medium percentages of 
houses (24-66 percent) with appliances present, and 19 builders with 
high percentages of houses with appliances present (>70 percent). 
The numbers of houses completed and appliance burglaries reported 
were obtained for the duration of the experiment for these groups of 
builders. 

While numbers of burglaries are small,6 the three groups differed 
according to expectation (see Table 1). For builders with low 
percentages of targeted appliances present, the burglary rate for these 
appliances (0.9 per 100 houses completed) was about one quarter of 
the rate for the high group (3.9 per 100 houses completed). 

6 These small numbers account for the lack of 
statistical significance in the differences 
between groups. 



Table 1 
Appliance Burglary Rates for Three Groups of Builders with Varying 
Percentages of Targeted Appliances Present at the Pre-Completion Stage 
Charlie One District, May to October 2000 

Houses with targeted appliances present at pre-completion 
Number of builders 
Houses completed* 
Burglaries of targeted appliances: No.
Rate per 100 houses 
Burglaries of all appliances: No.
Rate per 100 houses 

Low percent with appliances 
20 
560 
5 
0.9 
15 
2.7 

Medium percent with appliances 
20 
891 
13 
1.5 
17 
1.9 

High percent with appliances 
19 
311 
12 
3.9 
13 
4.2 

*Excluding 58 houses not covered by compliance checks 



The high burglary rate (2.7) for all appliances for the group of 
builders who most frequently delayed installation of targeted 
appliances supports the conclusion that delayed installation does 
protect builders from burglaries, since this rate includes hardwired 
appliances–the installation of which was not delayed. 

2. Differences in Burglary Rates Between Participating and Non-
Participating Builders 

Despite the disappointing rates of compliance for the participating 
builders, targeted appliances were found to be present in many fewer 
of their homes (about 22 percent) during compliance monitoring than 
for the other builders (appliances present in about 56 percent of 
checks). One might therefore expect that the participating builders 
would experience fewer burglaries. Contrary to expectation, however, 
the participating builders experienced virtually the same rates of 
burglaries of targeted appliances as the non-participating builders, and 
somewhat higher rates of burglary of all appliances (see Table 2). 



Table 2 
Appliance Burglary Rates for Participating and Non-Participating Builders 
Charlie One District, May to October 2000 

Houses completed* 
Burglaries of targeted appliances: No.
Rate per 100 houses
Burglaries of all appliances: No.
Rate per 100 houses 

Participating builders (n=12) 
631 
10 
1.6 
22 
3.5 

Non-participating builders (N=47) 
1,131 
20 
1.8 
23 
2.0 

All builders (N=59) 
1,762  
30 
1.7 
45 
2.6 

*Excluding 58 houses not covered by compliance checks 

One possible explanation for this anomalous result is that the 
participating builders might have been more vulnerable to burglary 
because of the location or size of their sites. Even after the 
experiment was introduced, cruising burglars could find appliances 
even though many of the houses were compliant.7 Greater prior 
vulnerability of the participating builders could help explain why they 
were willing to take part in the experiment. It would also be consistent 
with the fact that the real difference in burglary rates between them 
and other builders involved hardwired appliances, which were given no 
protection by the preventive measures.8 

To check whether the participating builders were at greater risk before 
the experiment, the analysis in Table 2 was repeated for the two years 
before the experiment, 1998 and 1999. Once again, data were analyzed 
only for May to October in both years. Not all the participating 
builders completed houses in these years and the mix of other 
builders in the two years was different from that in 2000. 
Nevertheless, the analysis did provide some support for the idea that 
the participating builders might previously have been at greater risk of 
burglary. As shown in Table 3, burglary rates for the participating 
builders were higher in 1998 and 1999 than for other builders. Of the 
four comparisons possible, the burglary rates were higher in all but 
theft of target appliances in 1999 (though the difference was very 
small). 

7 The burglars' search for appliances might 
even have been narrowed by the decals telling 
them which houses had no appliances and thus 
which ones to avoid. 

8 Another force might also have been at work. 
Burglars might have been returning to a familiar, 
favorite hole (as in fishing) - one in which they 
previously found in abundance what they were 
looking for - but now were faced with the need 
to search more intensively. 



Table 3 
Appliance Burglary Rates for Participating and Other Builders 
Charlie One District, May-October 1998 and 1999 

Burglaries of targeted appliances: Rate per 100
1998
1999
Burglaries of all appliances: Rate per 100 
1998 
1999  

Participating builders 
5.8 
4.0 
6.3 
5.7 

Other builders 
4.6 
4.1 
4.9 
5.1 

3. Overall Impact of the Experiment on Burglary Rates in Charlie One 

The last and perhaps most important question concerns the overall 
impact of the experiment on rates of burglary in the Charlie One 
district. This was assessed by comparing the rates of appliance 
burglary for May to October, 2000 (the period of the experiment) and 
those for the same time period in the two previous years. This analysis 
produced the most striking finding in the study. As can be seen in 
Table 4, the rate of appliance theft in the Charlie One district dropped 
more than 50 percent in 2000 compared with the previous two years.9 

Table 4 
Appliance Burglary Rates, Charlie One District 
May to October, 1998-2000 

Houses completed 
Numbers of burglaries 
Targeted appliances 
All appliances
Rates of burglary (per 100 houses) 
Targeted appliances 
All appliances 

1998 
1,072 
54 
58 
5.0 
5.4 

1999 
1,437 
58 
76 
4.0 
5.3 

2000 
1,820 
30 
45 
1.6 
2.5 

To check that these declines in burglary in Charlie One were not 
simply part of a wider fall in these crimes, rates of appliance burglary 
were calculated for the entire area covered by the CMPD. 
Comparisons of the rates for Charlie One and the remainder of the 
CMPD show that the drop in these burglaries is much greater in 
Charlie One (see Table 5). 

9 One-way Chi square tests found significant 
differences between observed and expected 
numbers of targeted and all appliance thefts 
across the three years. 



Table 5 
Appliance Burglary for Charlie One and All Other Districts in CMPD 
May to October 1998-2000 

-Charlie One 
-Other CMPD Districts 

Houses completed: 
1998 
-1,072 
-2,212 

1999 
-1,437 
-2,892 

2000 
-1,820 
-3,343 

Appliance burglaries (target): 
1998 
-54 
-56 

1999 
-58 
-39 

2000 
-30 
-60 

Appliance burglaries (all): 
1998 
-58 
-66 

1999 
-76 
-51 

2000 
-45 
-73 

Burglary rate (target appliances): 
1998 
-5.0 
-2.5 

1999 
-4.0 
-1.3 

2000 
-1.6 
-1.8 

Burglary rate (all appliances): 
1998 
-5.4 
-3.0 

1999 
-5.3 
-1.8 

2000 
-2.5 
-2.2 

Table 5 also provides little evidence of geographic displacement of 
the problem from Charlie One to the other districts served by the 
CMPD. Rates of appliance burglary did not increase markedly in these 
districts during 2000. 

A second possible form of displacement involves the targets of theft. 
Was there evidence that reductions in appliance thefts from 
construction sites in Charlie One were accompanied by increases in 
thefts of other items? An analysis conducted by the crime analyst 
attached to the project in its evaluation phase (Michael Humphrey) 
found that, for May to October 1999, 93 other thefts were recorded 
for Charlie One construction sites, while 87 were recorded for the 
same period in 2000. These figures translate into theft rates of 6.5 per 
100 completed houses in 1999 and 4.8 per 100 in 2000. Far from 
suggesting displacement of the problem to other thefts, these figures 
suggest that, as anticipated at the outset, the reduction in appliance 
thefts might have brought wider benefits through a reduction in other 
thefts. 



Before summarizing the results of the experiment, the possibility 
should be discussed that the reduction in appliance burglaries was the 
result not of delayed installation, but of the constant presence on the 
sites of Cunius and Rost when making the compliance checks. Each 
round of compliance checks took eight to ten days to complete 
involving one visit per construction site of about 30 minutes. This can 
hardly be described as a constant presence. Indeed, the officers 
believed that before the compliance checks, when they were on regular 
patrol duties, they spent more (not less) time in the construction sites. 
In addition, during the entire period of the experiment they never 
once saw another police patrol car in the construction sites. It 
therefore seems unlikely that police presence was the cause of the 
decline in appliance thefts. 


Summary of the Results of the Experiment 

Delaying installation of appliances until occupancy removes the 
opportunity for theft and the evaluation shows that builders who tend 
to follow this practice are likely to reduce their risks of appliance 
burglary. Rates of these offenses were considerably lower for builders 
who delayed installation for a high proportion of their houses. 

The evaluation also showed that, in the period of the experiment, 
there was a substantial overall reduction in the rate of all appliance 
theft in the Charlie One District. This is a little difficult to explain 
given that the burglary rates for participating builders were not lower 
than those of the other builders in the period of the experiment. 
However, this could have been due to the greater prior vulnerability of 
the participating group, whose burglary rates before the experiment 
were higher than those of the other builders. The experiment 
succeeded in lowering their burglary rates as well, but not to the level 
of the non-participating group. 

Also difficult to explain is why theft rates declined when there were 
still so many houses (40 percent) with appliances present at the pre-
completion stage. This means that cruising burglars would still have 
been able to find appliances if they were prepared to search a bit 
longer. But they seemed unwilling to do so, which suggests that the 
preventive measures were perceived as more widespread than they 
were in reality–a phenomenon known as diffusion of benefits (Clarke 
and Weisburd, 1994). Thus, some of the burglars might have decided 
it was no longer worth the effort of driving out to Charlie One when 
they heard that the installation of appliances was being delayed, when 
they found it more difficult to locate appliances, or when they began 
to encounter houses with the warning decals. Furthermore, it seems 
that few of them displaced their attentions to building sites outside 
Charlie One. Burglary rates did not increase in these building sites 
that, in any case, had never been as tempting as the ones in Charlie 
One 

The discussion below of the lessons to be drawn from this project 
concentrates on issues of wider significance, but we should not 
overlook the value that the new response has had for Charlie One and 
the CMPD. The district now has fewer thefts to which to respond, 
saving time, money, and resources. Moreover, the district is spared the 
frustration of being held accountable for the crime but not really 
being able to do anything about it. Cunius and Rost are more content 
in their work because of their pride in having brought a more 
intelligent, more effective, and ultimately timesaving response to this 
particular problem. 


Follow-up Action 

At the conclusion of the experiment, Cunius and Rost met with each 
of the twelve participating builders to obtain views on the experiment. 
All twelve were certain that delaying installation of appliances had 
been effective in reducing thefts. Some believed that other thefts had 
also declined. All twelve said they would continue to delay installation 
as a matter of policy. Some said that it should be a countywide policy. 

Some problems of coordinating delivery of appliances with closing 
dates were reported10 and several builders said that they had 
experienced difficulties in getting site supervisors to comply with the 
policy. Only a few reported objections from purchasers who mostly 
accepted the need for the measures once the reasons were explained 
to them. Providing coupons for the delivery of free pizza to meet the 
immediate need for food pending installation of appliances (a 
precautionary measure suggested by early discussions with builders 
about the experiment) mollified the more vociferous complainants. 

Three further rounds of compliance checks made by the two officers 
after the conclusion of the experiment (during January through March 
2001) showed that the participating builders had maintained their 
levels of compliance. Indeed, the overall compliance rate for the 
group of 81.5 percent was slightly higher than during the experiment 
(78 percent). Other builders were also delaying installation in an 
increased proportion of houses (51 percent during January through 
March 2001 compared with a little under 43 percent during the 
experiment). The officers believed that this increase was the result of 
site supervisors moving from employer to employer. In some cases, 
those previously employed by participating builders converted their 
new employers to the policy of delayed installation. 

As part of a broader effort to market the practice of delaying 
appliance installation, Cunius and Rost continued their efforts to 
persuade other builders to adopt it. They also made presentations of 
the results to CMPD command staff, to local builders associations and 
to the problem-oriented policing conference held by the Police 
Executive Research Forum in December 2000. They held briefing 
meetings for officials in nearby cities and for the press, the latter 
resulting in a favorable article in the Charlotte Observer (the local 
newspaper). They were invited to draft short articles for publication by 
the Police Executive Research Forum and for the house journal of the 
state builder's association. 

10 The contractors split their deliveries, 
arranging for hard-wired appliances to be 
delivered at one time followed by immediate 
installation, and for the plug-ins to be delivered 
as homes were occupied. Apparently they 
incurred minimal costs in adopting this practice. 



In another effort to extend the practice of delayed installation, 
Captain Johnson asked Cunius and Rost to sell the idea to builders 
who were embarking on the construction of apartment complexes in 
Charlie One–at the time of writing, this was becoming a full-time 
responsibility for one of the officers. Johnson also asked the officers 
to explore the possibility of delaying installation of dishwashers, 
which had been excluded from the experiment because they are 
generally hardwired. The officers discovered that these units could be 
supplied as plug-ins, though they would still have to be connected to 
the water supply. When undertaking a further round of compliance 
checks in March 2001, they established that dishwashers were installed 
at the pre-completion stage in 64 percent of the 388 checks they 
made. However, only one dishwasher was stolen in March from the 
total of 279 houses completed, a risk of less than 0.4 per 100 houses. 
This was thought not to justify the effort that would have been 
required for both manufacturers and builders to change their practices 
to allow for delayed installation of dishwashers. 

At the time of writing, Captain Johnson had recommended two 
further actions to his superior officers building on the work to date. 
The first involved instituting a departmentwide policy of not 
investigating reported thefts of appliances by builders who had 
refused to delay installation until occupancy, and informing such 
builders that proven means were available by which, through an 
adjustment in their own practices, they could substantially reduce such 
thefts. This did not gain the approval of the CMPD's executive staff. 
The second recommendation, still under consideration, involved the 
adoption of a countywide ordinance requiring all builders to delay 
installation of appliances. 


Lessons of the Case Study 

As noted earlier, the case study served three purposes: 

1. It met its original goals by providing a comprehensive example, 
within the CMPD, of what is involved in a concentrated, careful 
application of the full dimensions of problem-oriented policing. 
2. It enabled the CMPD to address the problem of theft of 
appliances from construction sites by learning as much as possible 
about the problem, testing the effectiveness of a new response, 
and ultimately achieving a substantial reduction in such thefts. 
3. As a result of their involvement in the project, the authors gained 
new insights into the complexities of carrying out problem-
oriented policing. The case study also enabled them to see more 
clearly what is needed to enable others to realize more 
expeditiously similar results. It is these lessons gleaned from the 
project that are summarized here. 


As explained, the project began life as a conventional police operation 
–one in which, having decided to take a proactive stance, the two 
Charlie One officers, Cunius and Rost, planned to undertake an 
intensive patrol and investigative effort to identify, arrest, and 
prosecute the criminals responsible for the construction site thefts. In 
fact, this enforcement effort continued long after the project had 
become an exercise in problem-oriented policing because of the time 
taken to complete the analysis and identify alternative responses. As a 
result, the two officers spent every available hour in 1999 patrolling 
the sites and working with detectives to build cases against anyone 
suspected of, or arrested for, construction site theft. 

By one measure, this was a successful operation because cases were 
made against 20 individuals in 1999 compared with only two 
individuals in 1998. But when related to the number of such crimes 
reported, the intensified effort aimed at detection, arrest, and 
prosecution was a failure. The number of appliance thefts in Charlie 
One actually increased in 1999 to 167 from 109 in 1998, and the rate 
of theft per 100 houses increased to 5.3 from 4.7. This early analysis 
led the police officers involved to conclude that despite their efforts 
conventional police work would not solve the problem. Not only had 
the officers' extraordinary efforts to identify, arrest, and prosecute 
offenders made little impact; they had been unable to acquire 
information about where the appliances were being sold, which could 
have led to new and potentially more fruitful investigative efforts.11 

11 Cunius and Rost ended up being 
enthusiastic about the project and its potential 
and are now effective spokesmen in challenging 
the heavily ingrained commitment to always 
reverting to the use of the criminal process (i.e., 
investigation, arrest, prosecution, etc.). This was 
evident during a presentation they made on the 
project at the annual problem-oriented policing 
conference in San Diego in December 2000, 
where they strongly rebutted suggestions made 
by some in the audience that a more vigorous 
patrol and investigative effort would have 
brought results. It is also evident in a report 
made by the two officers in April 2000 to the 
COPS office. This opens with the following 
paragraph: "In May 1998, we thought we knew 
about problem solving. We worked third shift 
and we were concerned about the construction 
site thefts in our district. We approached our 
captain with a plan to reduce thefts. The plan 
was thought out over a shift and was 
comprehensive. The plan was not a band-aid 
method but was made to target the three 
elements which make up a crime (suspect, 
victim and opportunity). We were to work with 
the burglary detectives to help identify the 
suspects. Once the suspect was arrested, we 
were going to petition the courts to get them 
the maximum prison sentence. As for the victim 
builders, we wanted to get after hour contact 
numbers in case suspects were apprehended in 
their neighborhoods. We were also going to 
exchange crime prevention ideas with the 
builders in order to improve and increase the 
builder's use of crime prevention techniques. As 
for the opportunity element, we wanted to alter 
our methods of patrolling the neighborhoods 
under construction. We planned on staking-out 
neighborhoods, using marked and unmarked 
patrols as well as altering the days and times 
we patrolled. We thought this plan would 
reduce the construction site thefts. Within six 
months our plan was barely intact. The site 
managers we contacted were either reassigned 
or had left the company. This made the after 
hour contact list and distribution of crime 
prevention information worthless. We did not 
identify any suspects and our directed patrols 
did not reduce the reported crimes. We had 
thought through our plan, but left out one major 
aspect, analyzing the problem." 



If the Charlie One team had accepted defeat at this point, the project 
would have been merely one more example of the limited capacity of 
conventional policing to reduce many ordinary crimes. In fact, the 
project demonstrates that where conventional policing has failed, 
problem-oriented methods can succeed. Some earlier examples of 
similar failures followed by successes involved the efforts to control 
slug use in the London Underground ticket machines (Clarke et al., 
1994), graffiti in the New York City subway (Sloan-Howitt and 
Kelling, 1990) and pay phone fraud in the Manhattan Bus Terminal 
(Bichler and Clarke, 1996). In all three cases, the police had tried and 
failed to deal with the problem through stakeouts, intensified patrols, 
surveillance, and the prosecution of arrested offenders. Police officers 
arrested only a small proportion of the offenders, they found it 
difficult to prosecute successfully, and, even if successful, the 
sentences imposed by the courts seemed to have little deterrent effect. 
In New York, the newspapers regularly ran stories about police 
frustrations in dealing with the problems in the Manhattan Bus 
Terminal, sometimes including photographs of police standing by as 
alleged offenders continued to make fraudulent use of the phones. 

In all three cases, the problems were solved only when the managers 
of the facilities stepped in with preventive measures of their own. In 
the case of the London Underground, the ticket machines were 
modified to block slug-use and in the Manhattan Bus Terminal 
conventional pay phones were replaced with more advanced models 
programmed to make fraud impossible. In the case of the New York 
subway, the solution consisted of an intensive and sustained program 
of graffiti cleaning. 

In Charlie One, the builders had no intention of taking ownership of 
the problem and finding solutions as in the above examples. It was 
easier for them to trust to luck and swallow any losses. The Charlie 
One team, however, was unwilling to give up without having achieved 
any positive results. By the time it was clear that conventional policing 
was not going to work, the problem-oriented analysis had begun to 
point the way to the solution eventually adopted. Cunius and Rost 
were anxious to try this solution, and the district's captain, Johnson, 
who was convinced that further effort was justified, was enthusiastic 
in his support of them. Even though Johnson was working in a 
supportive management structure, this cannot have been an easy 
decision for him to make. After a disappointing period of failure in 
pursuing the first initiative, it required that he commit substantial 
resources, at a time of great competition for the limited personnel 
resources available to him, to a long and sustained effort with no 
guarantee of success. But this points to one of the strongest general 
lessons of the project. If, based on the positive results derived from 



problem-oriented policing, it is to become a standard method of 
doing police business, the police (and those to whom they are 
accountable) will have to become accustomed to measuring the 
success of their efforts over a longer period of time, and will have to 
find ways to justify the use of the resources required to produce such 
positive results. 

Furthermore, to obtain the maximum benefit from the project, police 
must not rest content with reductions in the local problem. With a 
commitment to improve the quality of policing more broadly, police, 
like the Charlie One team, must be prepared to invest additional effort 
in documenting and disseminating the results. In this way, local inquiry 
and action can have national benefits, and, reciprocally, studies 
conducted elsewhere can have local benefits. This is particularly 
important for a problem such as construction site theft, which by its 
very nature is likely to be troublesome in most places for only a 
relatively brief period of time. Once the volume of construction 
decreases, the problem will likely disappear. But somewhere else, 
where construction is just beginning, the problem will soon emerge. It 
is important to enable those in the new location affected to profit 
from the experience of dealing with the problem gained elsewhere. 

This means that these wider benefits of local action should be 
included in any assessment of the costs and benefits of such action, 
which could radically affect any conclusions about the investment of 
resources. In Charlie One, the project, by conservative estimate, 
prevented over 100 thefts (i.e., about 6 per 100 houses completed) in 
2000. The district commander characterized this, when compared to 
other allocations of resources, as a good return on the time invested 
by the two officers. However, if the practice of delaying appliance 
installation were adopted more widely–throughout the county, or in 
the entire region, the state and ultimately the nation–without the need 
for so heavy an initial investment of resources, but with similar 
benefits, the cost-effectiveness balance would be even more 
favorable.12 

12 The amount of time devoted to the project 
by Cunius and Rost should not be taken as an 
indication of what a similar effort would require 
in other circumstances and in working under a 
more restricted time frame. The officers were 
periodically involved in other police activities; 
they were involved in the investigation of crimes 
and the prosecution of offenders; they spent 
considerable time learning new techniques of 
analysis and presentation; and they had the 
luxury, in this first-time project, of engaging in 
some wide-ranging explorations that, while 
productive, would not be required in a 
replication. 

Seeking these wider policing benefits will require police administrators 
to act increasingly like the heads of local public health departments 
whose professional commitment extends beyond their communities to 
building knowledge for their profession. Thus, while the principal duty 
of police administrators will still be to safeguard the communities in 
their care, they will also increasingly have to meet the important 
responsibility of alerting other communities to new hazards, to 
evaluating measures taken in response to these hazards, and to 
disseminating the results achieved. This will require them to undertake 
a delicate political balancing act. They must act successfully to deal 




with the problem locally, while at the same time justifying the 
additional expenditure needed to benefit communities elsewhere in the 
region, the state, and the country. The desirability of encouraging 
police agencies to engage in such studies and the benefits to be 
realized beyond the jurisdiction conducting the study constitute strong 
arguments for a program of federal financial assistance for such 
efforts. 

Few police departments currently possess the technical capacity to 
evaluate and disseminate the results of their operations. If the 
required knowledge and skills exist in any degree, they are most likely 
to be found in departmental crime analysis units, and frequently not 
even there. This brings us to another important lesson of the project, 
which concerns the vital role of crime analysts in problem-oriented 
policing, and how their work relates to the role of police officers–on 
which much more attention has focused. 

As originally conceived, problem-oriented policing anticipated the 
heavy involvement of individuals trained in analysis (Goldstein, 1979, 
1990). But as problem-oriented policing has developed, most of the 
emphasis has been placed on the potential contribution of those who 
are depended on to carry out the daily, challenging work of policing– 
line police officers–but who also are at the lowest level in the police 
hierarchy. The Charlie One project leaned heavily on the contribution 
of the two officers, Cunius and Rost, and by any measure, their 
contribution–especially to the analysis–was far beyond what could 
normally be expected of line officers. They spent days sorting through 
and re-classifying theft reports relating to construction site theft. They 
repeatedly interviewed builders to collect data about their security 
practices and theft experiences. They constructed databases that made 
it possible to calculate the rates of theft for individual builders, initially 
by number of building permits issued and later by certificates of 
occupancy. They prepared graphs and tables setting out the results of 
these calculations so that relatively unsophisticated audiences could 
understand them. Above all they spent countless hours, in all kinds of 
weather, making and recording the results of thousands of 
compliance checks. Along the way, they absorbed many lessons about 
analyzing a problem, developed their own analytical skills, and grew in 
their capacity to present the findings in a coherent, persuasive manner. 

However, even with the support and guidance they received, the 
officers would readily acknowledge that they could not have 
undertaken much of the work that was carried out by the crime 
analysts, who created computer maps of the construction site thefts, 



obtained and processed data on building permits and certificates of 
occupancy, and designed the computerized data collection systems for 
the compliance checks. 

The crime analysis unit in the CMPD–in its size, the abilities of its 
staff, and its highly developed use of computerized databases, 
including advanced mapping–is among the strongest such units to be 
found in a police agency. It was possible, through the involvement of 
Clarke and Goldstein in this project, to build on those strengths and 
to give the analysts the additional knowledge and skills needed to carry 
out problem-oriented policing. Drawing on their experience, Clarke 
and Goldstein provided the theoretical background needed for guiding 
the project. They elaborated on the action research methodology of 
problem-oriented policing, which is too often simplistically captured in 
the four stages of the SARA model. And they demonstrated how this 
would play out in dealing with the problem of construction site theft. 
Sensitive to the traditional perimeters of crime analysis and drawing 
on their experiences in other contexts, Clarke and Goldstein were in a 
unique position to contribute further by drawing attention to the 
importance of: 

1. Having realistic expectations about the results of enforcement 
efforts 
2. Focusing the project on a highly specific problem, or form of 
crime, i.e., appliance theft 
3. Supplementing the most common form of hot spot mapping with 
carefully developed information about the environment being 
mapped (e.g. the stage of construction reached and the identity of 
the different builders operating in the various subdivisions) 
4. Relating the absolute numbers of appliance thefts to the number 
of vulnerable homes, thereby producing suitable measures of risk 
before reaching conclusions about trends or patterns in these 
thefts 
5. Using acquired data to compare security practices and risk of 
thefts among builders, and to engage builders in assuming some 
responsibility for solving the problem; expecting the builders to 
assume some responsibility for solving the problem and formally 
comparing security practices and risk of thefts among builders 
6. Assuring that, in the language of routine activity theory, there are 
capable guardians of vulnerable targets 
7. Being alert to the possibility of diffusion of benefits and not 
being deterred from preventive action by the threat of 
displacement 
8. Monitoring closely the process of implementation and 
9. Utilizing an evaluation design that would permit definitive 
conclusions about the value of the response 




Police officers charged with undertaking a problem-oriented project 
and crime analysts who have limited themselves to the traditional 
forms of crime analysis cannot be expected to have an intuitive grasp 
of these points, and they will need to call upon others for advice. We 
believe that the repository of the research skills required in problem-
oriented policing, at least for the larger police departments, should be 
the crime analysis unit. This does not diminish in any way the 
contribution that police officers–both line and management–can make 
in addressing problems. As Cunius and Rost demonstrated, in an 
exceptional example, their contribution is vital and they can be drawn 
more fully into the process. Police agencies should continue to 
encourage all of their personnel to think about their work in terms of 
the problems they handle and their effectiveness in dealing with them. 
But officer involvement by itself, absent some special analytical and 
research skills, will not be sufficient to realize the full potential in 
problem-oriented policing. 

Judged by the CMPD, crime analysts have the necessary background 
to profit from the appropriate level of training in problem-oriented 
policing, which should also include components on environmental 
criminology and program evaluation. Unfortunately, training courses 
of this kind are not readily available, and providing them will be a 
considerable challenge for police leaders and others, such as municipal 
governments, federal agencies, and universities that support advances 
in policing. 

More is needed than training, however, to turn crime analysts into 
effective resources for problem-oriented policing. Equally important is 
that they be brought more directly into the management of a police 
agency; that, to the extent they become involved in in-depth analysis 
of the effectiveness of their agency in dealing with specific problems, 
they be recognized as the equivalent of product researchers in the 
private sector. Chief executives in the business world lean heavily on 
those who are equipped to analyze the quality of the end product. 
Properly trained crime analysts, engaged in the systematic study of 
problems that the police handle, as contemplated in problem-oriented 
policing, should have direct access to the top police administrator, 
should be involved in management meetings, and should be routinely 
consulted for guidance on how to improve the effectiveness of police 
efforts. Fully developed, their unique contribution could go a long way 
toward increasing the effectiveness of the police and, as consequence, 
the professional status of the police in the community. 


Conclusion 

In reflecting on this project, a reader might be inclined to ask several 
basic questions. Why was so much attention given to such a discrete 
and, in relative terms, low volume crime? Why should it have taken 
more than two years and substantial resources to explore and address 
the problem? Was the involvement of the two authors and the 
backgrounds they were able to bring to the project warranted, and did 
their senior status distort the results of the experiment? Would the 
research methods utilized meet the tests of scientific inquiry? Didn't 
much of what was explored simply constitute common sense? Wasn't 
the solution implemented almost embarrassingly simple? Why make so 
much both of the experiment in addressing the theft of appliances 
and of the process by which it was carried out? 

The answers to these questions are to be found in the current state of 
policing–in the United States and elsewhere. Despite the enormous 
investment made in policing, little is known about the problems police 
handle on a daily basis. The people affected–citizens, businesses, 
industries–also know little about them. And our knowledge about the 
most effective and economical ways in which to address these 
problems is very limited and often primitive. While some progress has 
been realized in recent years, we are only at the beginning in testing 
the value of specific strategies for dealing with specific problems. We 
have not developed, within policing, a protocol, a staff, a 
methodology, and, most importantly, a way of thinking that leads to 
the systematic study of problems and the relative merit of different 
strategies for dealing with them. The body of research that is available, 
usually conducted outside police agencies, is often inaccessible to the 
police. So, given the primitive state of affairs, getting going requires a 
great deal of trial and error. It is not a neat process. Indeed, it is a 
rather messy process. It requires getting one's hands dirty. It requires 
digging for data and often making do with less than what is want. It 
requires correcting and often supplementing existing data. It exposes 
difficulties that were unanticipated; phenomena that cannot easily be 
measured; findings and results that become blurred. It requires being 
ready to retreat or change course, and start all over again. 

These are among the reasons the project took as much time, 
resources, and coaching as it did. And these are the reasons why we 
saw merit in recording all of the details of this relatively modest 
exploration. Through publication, others can build on these efforts– 
and thereby ultimately contribute to the critical need to build a strong 
commitment within policing to gain new insights into the problems 
the police are expected to handle, and to develop a sophisticated 
capacity within policing to conduct such inquiries. 


References  

Bichler, G. and R. V. Clarke. (1996) "Eliminating pay phone toll fraud 
at the Port Authority Bus Terminal in Manhattan." In R.V. Clarke 
(ed.), Preventing Mass Transit Crime. Crime Prevention Studies, 
Vol.6. Monsey, NY: Criminal Justice Press. 

Chaiken, J. M., M. W. Lawless, and K. A. Stevenson. (1974) The 
Impact of Police Activity on Crime: Robberies on the New York 
City Subway System. New York: The New York City Rand 
Institute. 

Clarke, R.V. (1997a). Problem-Oriented Policing and the Potential 
Contribution of Criminology. Unpublished Report to the National 
Institute of Justice (Grant # 95-IJ-CX-0021). 

Clarke, R.V., (ed.) (1997b). Situational Crime Prevention: Successful 
Case Studies. 2nd Ed. Albany, NY: Harrow and Heston. 

Clarke, R.V. (1998). "Defining Police Strategies: Problem Solving, 
Problem-Oriented Policing and Community-Oriented Policing." In 
Problem-Oriented Policing: Crime-Specific Problems and Critical 
Issues and Making POP Work, Volume 1. Washington, D.C.: 
Police Executive Research Forum. 

Clarke, R.V. (1999). Hot Products. Understanding, Anticipating and 
Reducing the Demand for Stolen Goods. Police Research Series. 
Paper 112. London: Home Office. 

Clarke, R.V., R. P. Cody, and M. Natarajan (1994). "Subway slugs: 
tracking displacement on the London Underground." British 
Journal of Criminology 34:122-38. 

Clarke, R.V and G. McGrath (1990). "Cash reduction and robbery 
prevention in Australian betting shops". Security Journal, 1: 160
163. 

Clarke, R.V., and D. Weisburd. (1994). "Diffusion of Crime Control 
Benefits: Observations on the Reverse of Displacement." In 
Crime Prevention Studies, vol. 2, edited by Ronald V. Clarke. 
Monsey, NY: Criminal Justice Press. 

Constructor (1999). "A billion here and a billion there. A guide to 
theft and vandalism control". November, Vol. LXXXI, No 11, 1-
4. Alexandria, VA: The Associated General Contractors (AGC) of 
America. 



Dayton, J. (1993) "Construction theft takes sharp rise". The Business 
Journal Serving Greater Sacramento. February, 22, Volume 9 (48): 
1-7. 

Eck, J. E. and W. Spelman (1988). Problem-Solving. Problem-Oriented 
Policing in Newport News. Police Executive Research 
Forum/National Institute of Justice. Washington, DC: Police 
Executive Research Forum. 

Farrington, D.P. and P.A. Langan. (1998). Crime and Justice in the 
United States and in England and Wales, 1981-96. NCJ 169284. 
Washington, DC: U.S. Department of Justice. 

Goldman, M.D. (1999) "Cure can cost more than construction site 
theft". Washington Business Journal. June 14: 1-3. 

Goldstein, H. (1979). "Improving Policing: A Problem-Oriented 
Approach". Crime and Delinquency (April): 234-58. 

Goldstein, H. (1990). Problem-Oriented Policing. New York: McGraw 
Hill. 

Goldstein, H. (1994a) "Examining the Current Status of Problem-
Oriented Policing and Thinking Through an Agenda for Research 
and Technical Support for the Concept." Unpublished 
memorandum addressed to Craig Uchida, Office of Community 
Oriented Policing Services in the United States Department of 
Justice, July 11, 1994. 

Goldstein, H. (1994b) "Suggestions for Inclusion in the National 
Institute of Justice Research Agenda Relating to the Police." 
Unpublished memorandum addressed to Jeremy Travis, Director 
of the National Institute of Justice, September 29, 1994. 

Goldstein, H. (1996a) Problem-Oriented Policing: The Rationale, the 
Concept, and Reflections on Its Implementation. London: Police 
Research Group, Home Office. 

Goldstein, H. (1996b) "Problem-Oriented Policing." An address to the 
summer conference of the Association of Chief Police Officers, 
Manchester, England, July 3, 1996. 

Hunter, R. and C. R. Jeffery (1997) "Preventing convenience store 
robbery through environmental design." In R.V. Clarke (ed.) 
Situational Crime Prevention: Successful Case Studies,2nd Ed. 
Albany, New York: Harrow and Heston. 



Laycock G. and N. Tilley (1995). "Implementing crime prevention." In 
M. Tonry and D.P. Farrington (eds.) Building a Safer Society: 
Approaches to Crime Prevention. Crime and Justice: A Review of 
Research, Vol.19. Chicago, IL: University of Chicago Press. p.535
584. 

Mcgreevy, S.L. (1999) "Theft and vandalism: how to protect yourself" 
CONTRACTOR.mag.com. The Newsmagazine of Mechanical 
Contracting. 11/99: 1-4. 

Read, T. and Tilley, N. (2000). Not Rocket Science? Problem-Solving 
and Crime Reduction. Crime Reduction Research Series Paper 6. 
London, UK: Home Office, Police Research Group. 

Scott, M.S. 2000. Problem-Oriented Policing. Reflections on the First 
20 Years. Washington, D.C.: U.S. Department of Justice, Office of 
Community Oriented Policing Services. 

Scott, M.S. and R.V. Clarke (2000) "A review of submissions for the 
Herman Goldstein Award for Excellence in Problem-oriented 
Policing" In: C. Sole Brito and E. E. Gatto (eds.) Problem-
oriented Policing: Crime-specific Problems, Critical Issues and 
Making POP Work, Volume 3. Washington, DC: Police Executive 
Research Forum. 2000. 

Sloan-Howitt, M. and G.L. Kelling (1997) "Subway graffiti in New 
York City: 'Gettin Up' vs. 'Meanin It and Cleanin It.'" In R.V. 
Clarke (ed.) Situational Crime Prevention: Successful Case Studies, 
2nd Ed. Albany, New York: Harrow and Heston. 

Snyder, R. G. (2000). "Construction theft: What you can do to prevent 
it". NC Builder Magazine. December, 1-2. 

Stanford Research Institute (1970). Reduction of Robbery and Assault 
of Bus Driver. Vol. III. Technological and Operational Methods. 
Stanford, Calif: Stanford Research Institute. 

Stewart, L. (1998). "Simple and sophisticated strategies to thwart 
thieves". Construction Equipment, August, Vol. 98(2): 60-70. 

Stewart, L. (2000). "Firms discount equipment lost to crime". 
Construction Equipment, April, Vol. 101(4): 66-72. 



Appendix 

Builder Compliance Rates and Burglaries Charlie One, May to October, 2000 

BUILDER 
1 
Compliance Checks ( 1 ) 
23 
Target appliance present ( 2 ) 
0 
Appliance Compliance ( 3 ) 
100.0% 
Homes completed (4 ) 
2 
Burglary of target appliance ( 5 ) 
0 
Burglary of any appliance (6 ) 
0 

2   121   0  100.0%   0  0  0 
3     1   0  100.0%   0  0  0 
4    13   0  100.0%   3  0  0 
5     6   0  100.0%   2  0  0 
6     1   0  100.0%   0  0  0 
7(H) 23   0  100.0%  10  0  0 
8(H)  2   0  100.0%   2  0  0 
9    27   1   96.3%   1  0  0 
10  295  17   94.2%  56  0  0 
11  377  25   93.4%  111 0  2 
12  130   9   93.1%  14  0  0 
13   21   2   90.5%   3  0  0 
14 1122  131  88.3%  108 3  9 
15   87  11   87.4%  58  0  0 
16  123  16   87.0%  25  0  0 
17   23   3   87.0%  20  0  0 
18  668  102  84.7%  117 2  4 
19(H) 6   1   83.3%   4  0  0 
20   24   4   83.3%  24  0  0 
21  487  118  75.8%  84  1  1 
22   58  17   70.7%   8  0  0 
23  137  48   65.0%  15  1  1 
24  583  207  64.5%  218 1  1 
25  120  51   57.5%  64  0  0 
26   18   8   55.6%   2  0  0 
27   17   8   52.9%  14  0  0 
28  222  112  49.5%  57  1  5 
29  100  51   49.0%  14  0  0 
30  141  73   48.2%  41  5  5 
31   43  26   39.5%   5  0  0 
32   33  20   39.4%  17  0  0 
33  319  196  38.6%  106 1  1 
34   96  59   38.5%  49  0  0 
35  186  116  37.6%  67  0  0 
36  164  104  36.6%  40  2  2 
37  118  75   36.4%  34  0  0 
38  105  69   34.3%  33  0  0 
39   62  41   33.9%  16  1  1 
40(H) 3   2   33.3%   7  0  0  
41  133  93   30.1%  16  0  0 
42   68  48   29.4%  17  0  1 
43  502  362  27.9%  108 2  2 
44   59  44   25.4%  55  0  0 
45   75  58   22.7%  11  2  2 
46   72  57   20.8%  15  0  0 
47  143  114  20.3%  28  2  2 
48   45  37   17.8%  12  0  0 
49   28  24   14.3%   6  0  0 
50   15  13   13.3%   0  0  0 
51  109  98   10.1%   6  0  0 
52   30  27   10.0%   2  1  1 
53  162  151   6.8%  13  3  3 
54  244  230   5.7%  18  2  2 
55  155  150   3.2%   2  0  0 
56   62  62    0.0%   0  0  0 
57    4   4    0.0%   0  0  0 
58    1   1    0.0%   1  0  0 
59   38  38    0.0%   1  0  0 
   
8,050  3,334  58.60%  1,762  30 45 


Notes to ANNEX 

(1 ) Number of checks made to determine if target appliances were 
present in a house. These are appliances that plug into an 
electrical outlet or any appliance that is scheduled to be hardwired, 
but has not yet been installed and is, therefore, like a plug-in 
appliance easily stolen. 

( 2 ) Number of times a target appliance was found present during a 
compliance check. 
( 3 ) The percentage of compliance checks when target appliances 
were not present within the home. 
( 4 ) A completed home is a home that has passed all inspections 
required by the building code and qualifies for a Certificate of 
Occupancy. 
( 5 ) Number of burglaries resulting in the theft of one or more 
targeted appliances. 
( 6 ) Number of burglaries resulting in the theft of one or more 
appliances, including hardwired appliances. 
( H ) indicates the builder uses hardwiring exclusively in installing all 
appliances. 



DOJ Seal

FOR MORE INFORMATION: 
U.S. Department of Justice 
Office of Community Oriented Policing Services 
1100 Vermont Avenue, NW 
Washington, D.C. 20530 

To obtain details on the COPS Office program, call the 
U.S. Department of Justice Response Center at 1.800.421.6770 
Visit the COPS internet web site by the address listed below. 


e12021822 Created Date: January 29, 2003 


www.cops.usdoj.gov