Schoolcraft v. The City Of New York et al

Filing 497

DECLARATION in Support re: 495 MOTION to Preclude Expert Testimony.. Document filed by Christopher Broschart(Tax Id. 915354 in his official capacity), Kurt Duncan(Shield No. 2483, Individually), William Gough(Tax Id. 919124, Individually), Elise Hanlon(in her official capacity as a lieutenant with the New York City Fire Department), Shantel James(Shield No. 3004 in his official capacity), Theodore Lauterborn(Tax Id. 897840 in his official capacity), Michael Marino, Gerald Nelson(Assistant Chief Patrol Borough Brooklyn North, Tax Id. 912370 in his official capacity), Frederick Sawyer(Shield No. 2576 in his official capacity), The City Of New York. (Attachments: # 1 Exhibit A Expert Report, # 2 Exhibit B Deposition Excerpts, # 3 Exhibit C Deposition Excerpts, # 4 Exhibit D Under Seal, # 5 Exhibit E Email Chain Between Counsel, # 6 Exhibit F Justice Quarterly Article, # 7 Exhibit G Under Seal, # 8 Exhibit H Under Seal, # 9 Exhibit I Under Seal, # 10 Exhibit J Transcript, # 11 Exhibit PTX 66 Under Seal, # 12 Exhibit PTX 81 Part I, # 13 Exhibit PTX 81 Part II, # 14 Exhibit PTX 93 Arbitration Award, # 15 Exhibit PTX 159 Patrol Guide)(Scheiner, Alan)

EXHIBIT F

This article was downloaded by: [Molloy College], [Jhn Eterno]
On: 20 November 2014, At: 06:41
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Justice Quarterly
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/rjqy20

Police Manipulations of Crime
Reporting: Insiders’ Revelations
John A. Eterno, Arvind Verma & Eli B. Silverman
Published online: 17 Nov 2014.

To cite this article: John A. Eterno, Arvind Verma & Eli B. Silverman (2014): Police Manipulations of
Crime Reporting: Insiders’ Revelations, Justice Quarterly, DOI: 10.1080/07418825.2014.980838
To link to this article: http://dx.doi.org/10.1080/07418825.2014.980838

PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions

Justice Quarterly, 2014
http://dx.doi.org/10.1080/07418825.2014.980838

Police Manipulations of Crime Reporting:
Insiders’ Revelations

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

John A. Eterno, Arvind Verma and
Eli B. Silverman

Social scientists have theorized about the corruption of crime reports
(Bayley, 1983; Campbell, 1976). Yet, scant empirical research has examined
the impact of modern policing methods on the accuracy of crime reporting.
Our research uses an anonymous survey of 1,770 retired New York City
police officers examining retirees’ experiences with crime report manipulations across their years of retirement. This includes retirees from the
community policing as well as police performance management eras. We
subject the data to various statistical tests including tabular analysis,
graphical trends to visualize the data, MANOVA, and logistic regression to
explain report manipulations. Results indicate that the misuse of the
performance management system and pressures on officers from management are key explanations for manipulating crime reports. Individual
explanatory variables such as gender, educational status, rank, race, and

John A. Eterno is professor, associate dean and director of Graduate Studies in Criminal Justice at
Molloy College in Rockville Centre, New York. His research interests are democratic policing, police
management, international law enforcement, human trafficking, investigations, and law—especially
as it relates to policing. His publications include author/editor of four books, numerous book chapters
and reports, and examples of his scholarly articles can be found in Justice Research and Policy, The
Criminal Law Bulletin, Women and Criminal Justice, Police Practice and Research: An International
Journal, The International Journal of Police Science and Management, Policing: An International
Journal of Police Strategies and Management. Arvind Verma is on the faculty of the Dept of Criminal
Justice, Indiana University, Bloomington. His research interests are in Policing: Criminal Justice Policy Issues, Indian Police, Research Methods, Mathematical Modeling, and Agent-Based Simulation. His
recent publications include books titled Policing Muslim Communities: Comparative and International Context; The New Khaki: The Evolving Nature of Policing in India, Understanding the Police in
India and recent journal articles, “Validating Distance Decay through Agent Based Modeling” and
“Assessing the Role of Police in Containing Mob Violence.” Eli B. Silverman, PhD, Professor Emeritus,
John Jay College of Criminal Justice and Graduate Center, City University of New York, served with
the US Department of Justice and National Academy of Public Administration, Washington D. C and as
exchange professor, Police Staff College in England. His areas of interest include: police performance
management, community policing, policy analysis, stop and frisk, and compstat. Recent books: The
Crime Numbers Game: Management by Manipulation and NYPD Battles Crime: Innovations in Policing. Correspondence to: John A. Eterno, Molloy College, 1000 Hempstead Avenue, Rockville Center,
NY 11571-5002, USA. E-mail: [e-mail]

Ó 2014 Academy of Criminal Justice Sciences

2

ETERNO ET AL.

marital status had no effect. Our research supports Bayley’s and Campbell’s
theories. We recommend greater transparency to remedy this.

Keywords
statistics

COMPSTAT; performance management; NYPD; crime rates; crime

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

Introduction
Researchers, policy-makers, media, and scholars rely on crimes reported by
the police as a staple of their analyses. Numerous reports, scholarly papers,
media accounts, books, and other materials are grounded on crime reports.
The current study critically examines this excessive reliance on police reports
by focusing on the New York City Police Department (NYPD).
The NYPD’s substantial and significant impact on policing in the free world
is indisputable. Bayley (1994) calls the NYPD, the nation’s largest police force,
a “flagship” police agency as other departments follow NYPD’s lead. NYPD’s
performance management system, Compstat (short for compare statistics),
starkly epitomizes the NYPD’s leadership position. Compstat began in 1994
under the then Police Commissioner William Bratton (Silverman, 2001). Focusing on index crime and using computer mapping, managerial accountability,
data-driven reporting, and regular crime meetings, Compstat has been emulated throughout the United States and the free world. A recent survey by the
Police Executive Research Forum (PERF) indicates that Compstat is now ubiquitous. In 2011, the survey was sent to 326 of PERF’s member agencies with 166
responding. The survey indicated that 79% of the agencies utilized Compstat.
Further, over half (52%) indicated that they began using Compstat between
2006 and 2010 (Bureau of Justice Assistance, Police Executive Research Forum,
2013).
In addition, the sheer number of agencies now using computers for crime
analysis, crime mapping, and hotspot identification indicate the widespread
adoption of NYPD Compstat-like performance management systems. In the
United States, a nationwide study of police agencies in 2007 showed that 100%
of the agencies serving populations of 250,000 or more were using computers
for crime analysis and crime mapping. Further, 80% of those same agencies are
also using computers for hotspot identification (Reaves, 2010).
The bulk of New York City’s crime reduction has taken place since 1994
when Compstat—a police performance management system—and other police
reforms were introduced. The NYPD repeatedly publicizes an astounding
79.67% drop in index crime since 1990. New York City’s reported dramatic
crime decline has drawn widespread attention to the NYPD from professionals,
criminologists, law enforcement, politicians, the media, and the general public. Such a phenomenon is certainly important for social scientists to study and
understand.

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

3

Some attribute the decline almost exclusively to the practices of the police
and in particular to Compstat (Bratton, 1998; Henry, 2002; Kelling & Coles,
1998; Kelling & Sousa, 2001; Maple, 1999; Zimring, 2011). Some have labeled
these police activities as responsible for the “Guinness Book of World Records”
crime drop (Zimring, 2011). Other studies demonstrate less confidence in the
police’s role in the crime decline. (Bowling, 1999; Conklin, 2003; Harcourt,
2001; Karmen, 2000; Messner et al., 2007; Rosenfeld & Fornango, 2014;
Rosenfeld, Forango, & Rengifo, 2007).
Regardless of the explanation that any individual author favors, these studies often use official police reports as their underlying (data) support. However, earlier research on police reporting of crimes suggests such data is, at a
minimum, suspect. Marvell and Moody (1996), for example, demonstrate that
the number of crime reports is correlated with the number of officers—more
officers lead to more reports. In New York City, the number of officers has
been drastically reduced by thousands since its high point in 2000. This alone
may have a dramatic influence on reported crime in New York City, but is
rarely acknowledged by the media, politicians, and, more importantly, social
scientists. Thus, previous research on New York City tends to have an overreliance on crimes reported by the police.
Bayley (1983), moreover, lists 19 factors that affect the correspondence
between reported crime and actual crime incidence. Among these factors are:
organizational emphasis in policing, number of police, attentiveness of police,
consequences to the police of recording an offense, and willful repression by
police. He further writes, “In my view, unless there is evidence, not presumption, that these factors are negligible or stable between places or times being
compared, reported crime figures must not be used in evaluating the effectiveness of police performance.” (Bayley, 1983, p. 25) Thus, not only are there
well-known defects with the crime reporting system (see also Maltz, 1977;
McCleary, Nienstedt, & Erven, 1982; Mosher, Miethe, & Hart, 2011; Seidman &
Couzens, 1974), but using reported offenses to gauge police effectiveness can
be problematic.
The use of crime reports to measure police effectiveness may be especially questionable in New York City. For example, one important defect in
crime reporting that social scientists tend to overlook, downplay, or even
ignore in studying New York City is the possibility that an overly intense
focus on index crime by management could lead to the manipulation of
crime reports by lower ranking officers. Since Compstat meetings and reports
heavily focus on index crime, it is possible that this management priority will
distort official crime statistics (Eterno & Silverman, 2012). This postulate is
sensible if not axiomatic. As the eminent social psychologist Donald Campbell
(1976, p. 49) writes, “The more any quantitative social indicator is used for
social decision-making, the more subject it will be to corruption pressures
and the more apt it will be to distort and corrupt the social processes it is
intended to monitor.”

4

ETERNO ET AL.

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

Background Information
The setting for this study is New York City—a city with over 8 million people,
millions of tourists and commuters, as well as diverse cultures. In addition to
the city’s frequently cited large decrease in reported crime and associated
research, evidence and testimony from police unions, whistleblowers, hospital
data, a thwarted commission, and, most recently, an internal committee’s
findings provide compelling reasons to focus on New York City. Two police
unions (the Patrolman’s Benevolent Association and the Sergeant’s Benevolent
Association) joint press release on March 23 2004 maintained that crime
reports were downgraded. They state: “It [downgrading] is a truth that is
widely known by members of the department …” (“Unions Call”, 2004, n.p.)
Whistleblowers have provided audiotapes indicating that crime reports are
manipulated. Officers Adrian Schoolcraft, Adyl Polanco, Sergeant Robert
Borrelli, and others produced ample evidence of manipulation. Indeed, Officer
Schoolcraft’s allegations have since been confirmed by the department’s own
“secret” internal investigation exposed by a reporter (Rayman, 2012). One
example of the level of manipulation from Schoolcraft’s tapes is supervisors at
roll calls instructing entire shifts of officers not to take robbery reports if victims are not willing to return to the station house—clearly violating department training and procedures.
Hospital data, available through 2006 on these issues, raises serious questions regarding manipulation particularly with respect to assaults, thus providing additional reasons for studying the phenomenon of crime reporting
manipulation. New York City’s Department of Health and Mental Hygiene statistics indicate that in 1999 there were 25,181 visits to emergency rooms
throughout the City for assault. In 2006, there were 47,779 visits. This represents a 90% increase in visits for assault while the NYPD reported a substantial
decline. Even more a concern is visits to emergency rooms for firearms assaults
which skyrocketed as well. In 1999, there were 224 visits and in 2006, they
count 514 visits—a 129% increase (New York City Health and Hospitals
Corporation, 2011).
The NYPD’s lack of transparency (i.e. openness to non-partisan research and
requests from outside the agency) makes value-neutral scientific study difficult,
yet at the same time more urgent. NYPD has not allowed scrutiny of its written
crime reports even from official government bodies. For example, Mark Pomerantz, a former federal prosecutor who chaired New York City’s Commission to
Combat Corruption (a Mayoral agency) was denied permission by the police
department to view its crime reports (Rashbaum, 2005). His appeal to Mayor
Bloomberg was rebuffed resulting in Mr Pomerantz’s resignation.
Under mounting pressure from public disclosures of crime report manipulation, in January 2011, Commissioner Raymond Kelly appointed three former
federal prosecutors to review NYPD’s crime report auditing practices. The
report (released 2 years later than its promised date) does not examine written

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

5

complaints but merely reviews the auditing system. Even this report raises
serious questions. For example, the report indicates, “… the persistence of
‘egregious’ errors in certain precincts despite the pre-finalization review of
the complaint reports by supervisors may be construed to support the conclusion that complaint reports are not meaningfully or at least proficiently
reviewed at the precinct level—or, in the worst light, that the reviewing
supervisors are complicit in the downgrading …” (NYPD, 2013, p. 47)
Recent studies on New York City crime reveal how scientists are attempting
to cope with the issue of report manipulation. Greenberg (2014), for example,
simply chooses to disregard the possibility of police manipulating crime reports
by citing a non-peer-reviewed source. Another study using the National Crime
Victimization survey (NCVS) tangentially discusses this issue. Xie (2014) proposes that victims in New York City are confident in the police and report regularly. However, even if victims report to the police, this does not necessarily
mean the reports are properly classified or taken—victims would rarely know
this. Furthermore, Karmen (2014) reports that the NCVS has been plagued with
funding problems and that the authors of a major study on New York City who
compared NCVS with NYPD figures could not offer any comments with respect
to report manipulation by police (Langan & Durose, 2004).
Thus with numerous studies on the New York City crime drop often relying
on crimes recorded by the police as their foundation, the matter of report
manipulation is ripe for scientific inquiry. We hypothesize that perceived managerial pressures on lowering index crime via the Compstat crime management
system in New York City likely has a deleterious influence on the accuracy of
crime reporting by police. Testing this hypothesis is difficult, at best, since the
NYPD’s transparency is quite limited (Ortiz, 2011). The current study fills a
research gap by using unique primary data in the form of surveys of nearly
2,000 retired New York City officers of all ranks, thereby providing specific
data that demonstrates, explains, and compares these officers’ experiences
with crime report manipulation in different time periods.

Methodology
This combination of an array of evidence suggestive of crime report manipulation and the lack of NYPD transparency provides the backdrop for the construction of this study’s research method. While whistleblowers’ revelations and
newspaper accounts indicate some level of manipulation, there is little in the
form of scientific study on the issue. To that end, the heart of our method is
the use of an anonymous survey of retired officers of all ranks. The current
study incorporates a large sample of 1,962 retired officers who responded to
the survey. An e-mail-based survey was conducted in March 2012. The sampling
frame is registered officers on the NYPD active retiree’s database. Only retired
officers can register on the site which is maintained by the NYPD. The purpose
of the site is to offer information to retirees and to tap them in emergency

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

6

ETERNO ET AL.

situations as necessary. It is completely voluntary for retired members to
register. The site allows registered members to contact other registered retirees and the site supplies e-mail addresses and other contact information. We
had access to a retiree allowing us to download the contact information.
Additional information about the database can be found at the website http://
www.nypd2.org/retirement/home.html.
Using this database, we sent a request to participate in the survey—which
included a unique link to the questionnaire in its body—via e-mail to the 4,069
NYPD retirees with operational e-mail addresses (see e-mail verbiage in
Appendix). The survey was sent on March 26 2012. If the unique link did not
indicate a response, a follow-up was sent on April 3. A final follow-up was sent
only to links that indicated no responses to the first or second invitations on
April 16. In total, there were 1,962 responses. This indicates a return rate of
48.2%. Mail surveys commonly have response rates between 10 and 50%
(Neuman, 2000). E-mail surveys tend to have weaker response rates when we
examine more recent years (Sheehan, 2001). This study’s response rate, then,
is certainly good—especially for a contemporary e-mail survey.
We examine the data by the officers’ reported year of retirement. For some
analyses, we divide the sample into 3 key groups using 1981 as the earliest
year proceeding to 1993 which is the last year prior to the introduction of
NYPD’s Compstat management system. We then examine 1994 until 2001—the
Mayor Giuliani years. Lastly, we examined those who retired from 2002 until
2012—the Mayor Bloomberg years. This categorization emerged from trends
we saw in the data. We removed those who did not supply their year of retirement and those who retired before 1981. The officers who retired before 1981
had policing careers that focused on traditional and reactive policing. These
officers were from a different era. For example, most were trained pre-Knapp
Commission (a critically important commission to combat corruption at NYPD)
in the early 1970s. Another example of their differences is seen in the class of
1969. That class entered the streets during the Vietnam riots without police
academy training. As a final example, women made great strides in policing in
the 1980s. There were very few females before 1981. Thus, we removed 45
officers who retired before 1981—not enough for a separate analysis of this
era. In addition, we removed another 147 who did not supply a comprehensible
retirement year. Therefore, for the purposes of this study, our final sample
size was fairly large—1,770 retired officers. The sample size adds assurance
that the findings accurately reflect the experiences of actively retired officers.
Table 1 shows the demographic information of the sample. While our sampling frame examines actively retired officers, we also note that the trends in
the characteristics of the respondents are very similar to what is known about
NYPD. The number of retirees in each rank comports well with known rank distributions of NYPD. That is, police officers/detectives are the largest group
with descending counts as one goes up the ranks. Thus, as expected, there are
fewer sergeants compared to police officers/detectives and so forth. This
pyramidal structure clearly mimics the rank structure at NYPD. Education level

PO/Detective
60.1%
GED/HS only
7.7%
Pre 1994
2.1%
Pre 1994
65.9
White
<94
94–01
88.2 82.6
Single
6.7%

Sergeant
19.8%
Some College
44.2%
1994–2001
10.7%
1994–2001
58.8
Hispanic
>01
<94 94–01 >01
74.6 5.1 10.2
15.2
Married
79.4%

*Percentages may not add to 100 due to rounding.

%=
Marital

Race/ethnicity

Avg. Age

Females

Educ.

Rank

Variable

Table 1 Demographics of sample

Lieutenant
13.4%
Bachelor’s Deg.
29.4%
Post 2001
11.0%
Post 2001
50.1
Black
<94 94–01 >01
3.9 5.2
7.0
Divorced
9.1%

Capt. & up
5.9%
Some Grad.
6.6%
Total in sample
8.5% (n=149)
Mean of sample
56.7
Asian
<94 94–01 >01
0
.2
.7
Widow(er)
2.0%

Attributes
Other
.2%
Grad. Degree
11.3%
Missing
.7%
Missing
4.6%
Other
<94 94–01 >01
2.8 1.7
2.5
Other
1.2%

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

Sub-total n
<94 94–01
467 403
Missing
1.5%
n = 1,770

Missing
.6%
Missing
.7%

>01
846

POLICE MANIPULATIONS OF CRIME REPORTING
7

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

8

ETERNO ET AL.

is also as expected since 60 credits college or 2 years military has been a
requirement since the mid-1990s. Gender too is as expected since this study
examines earlier years where the NYPD had not yet hired many female officers. Also, the percentage of women officers increases in each of our groups
(1981–1993 [2.1% female]; 1994–2001 [10.7% female]; 2002–2012 [11.0%
female]]). Today, NYPD has nearly 35,000 officers and is 17% female (For 1st
Time, 2011). Only 30 years ago, there were 350 women officers (NYPD Press
Release, 2009). Lastly, race also mimics what is known about NYPD. NYPD
trends indicate that minorities are making great strides (El-Ghobashy, 2011).
Our sample shows whites decreasing as a percentage in our 3 categories pre1994, 1994–2001, and 2002 onward (88.2, 82.6, and 74.6%, respectively). Hispanics and blacks both show increasing trends with Hispanics having a greater
percentage. This again agrees with known trends (blacks 3.9, 5.2, and 7.0%;
Hispanics 5.1, 10.2, and 15.2% respectively). Importantly, we also note that
every officer in the study had considerable experience of working in different
units of NYPD including work in a patrol (precinct) assignment at one time in
their careers. Overall, the trends in the data are clear and completely in synch
with trends at NYPD.
The large sample size and the distributions that comport with what limited
information is available about NYPD personnel are indicators that social bias is
probably not an issue and that the sample is somewhat representative of
NYPD. We also note that all respondents elected to be on the database maintained by their former employer and volunteered to help their former
employer in emergencies. We supply this information for readers to consider
for themselves how this might influence the sampling frame and corresponding
interpretation of the data.
It is important to note that the probability of events being conflated (i.e.
officers reporting the same event) is highly unlikely. First, we ask officers to
report their personal experiences—not what they heard or learned about. Second, our sample includes officers who retired over various decades and who
worked in numerous and varied commands. The NYPD, for example, has 76
precincts, 12 Transit Districts, and 9 Housing Police Service Areas. This is not
to mention numerous other commands (detective commands, organized crime
control bureau, internal affairs, etc.) Also, officers with more seniority tend to
change assignments to more pleasant working environments. Additionally, the
open-ended responses to our last question show no evidence of events being
repeated. Simply put, it is exceedingly unlikely that officers reporting to us
were working together and reporting the same event. What we are capturing
in this survey are the officers’ personal and individual experiences while
employed at the NYPD.
In addition, anonymity is a key feature of our survey as it allows maximum
privacy and tends to produce more valid responses to sensitive questions such
as those in this survey (Babbie, 1989; Bradburn, 1983; Dillman, Smyth, &
Christianson, 2009; Neuman, 2000). As Mosher et al. (2011, p. 219) write, “…
the relative anonymity of the Internet may lead to higher response rates and

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

9

more truthful answers (Skitka & Sargis, 2006).” Furthermore, respondents have
the extra protection of being retired. They have far less to fear from repercussions by the NYPD (e.g. whistleblowers Schoolcraft, Polanco, and others were
punished for revealing their experiences). Since our sample retired from the
NYPD, they are in a far better position to reflect on their careers, share their
experiences and answer candidly.
The questionnaire was designed using focus groups including retired NYPD
officers of varied ranks, experienced researchers, and others familiar with
NYPD jargon. The instrument and the e-mail invitation were reviewed by the
Institutional Review Board affiliated with the principal researcher. Wording
was carefully designed to address any issues with respect to whether officers
actually observed the activities in the survey. Therefore, the wording “Based
on your experience, do you have any personal knowledge of …” was placed on
all pertinent questions. We do note, however, that such knowledge could be
from a third person. Also, respondents were instructed on the front page of
the survey that if they retired 1994 (the first year of Compstat) or after to
base their answers only on experiences in 1994 and after. Conversely, those
who retired before 1994 are being asked about their personal experiences in
the era when they worked.
Given that the entire sampling frame is retired, memory can be an issue.
We address this using techniques suggested in the scientific literature. We use
a very limited number of questions (Bradburn, Rips, & Shevell, 1987); a specific
cut time—1994—(Bradburn et al., 1987); a self-administration method that
allows maximum time to consider answers (Bradburn et al., 1987); analyze
whether the event occurred—not specific number of events—(i.e. our dependent variable is dichotomous) (Bradburn et al., 1987), less complex matrix type
questions for the key independent variables (Bourbonnais, Meyer, & Theriault,
1988), confirm the analysis with four separate dependent variables (i.e. questions) (Zelinski, 1999); use those employed for many years (Stewart, Tonascia,
& Matanoski, 1987 inform us that 5 or more years of employment leads to better recall—only 1 respondent had less than 5 years’ experience); and ask questions that are pertinent to the study and to respondents (Bourbonnais et al.,
1988).
We tested the reliability of the responses by estimating Cronbach’s α for
the 8 items measured on the Likert Scale of 1–10 [minimal to strong pressure].
The value was observed to be .802 that indicated a high degree of reliability in
the responses. Further, we also estimated the reliability by conducting a splithalf procedure. The Spearman Brown split half value was found to be .760
assuming unequal length of the two halves. Both of these values suggest that
the reliability of the items was high.
Our statistical analyses focus on three key areas. First, we compare and
contrast their experiences with crime report manipulation based on officers’
reported year of retirement. Second, we examine the perceptions of pressure
from management, including pressures to reduce index crime, across the
years. Third, we examine the extent to which, if at all, the perceived

10

ETERNO ET AL.

management pressures explain observed crime report manipulation. Where
appropriate, we also include various quotes from a general open ended question posed at the end of the questionnaire. These statements support the
quantitative analysis.

Data Analysis

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

Data Information
The second survey question was in the form of a matrix. It reads, “With
respect to the following criteria and based on your personal experience/knowledge, on a scale of 1–10 (with 1 being the least and 10 the most), how much
pressure did precinct (patrol) personnel receive from management/supervisors
to” decrease index crimes, decrease other crimes, downgrade index to nonindex crimes, maintain integrity in crime statistics, and to obey Constitutional
rules (see Appendix). The extent of those pressures (or lack thereof) helps to
reveal the experiences of those who worked in various years. For one variable,
downgrading index to non-index crime, little pressure is customarily expected
since the activity of downgrading is itself unethical and likely illegal (e.g. filing
false police reports, tampering with official documents). With regard to the
level of perceived pressure to report crime statistics accurately or to obey
Constitutional or departmental rules, we expect somewhat higher pressures to
remain fairly constant across our respondents’ retirement years. This is
because police ought to always report crime accurately and obey the law.
The coding of the independent variables is very straightforward. For the
pressure variable in the matrix question, we use a Likert scale of 1–10 with 1
being the least pressure and 10 being the most. We use the 1–10 scale for
familiarity and note that there is no zero point. We clearly mark “least pressure” for number 1, but without a zero point some pressure, however minimal,
might be inferred and readers should interpret the data with that in mind. For
gender 1 = Male and 2 = Female; education level 1 = High School, 2 = Some College, 3 = College Graduate, 4 = Some graduate/law school, 5 = Graduate or law
degree completed; marital status 1 = married 2 = not married; race 1 = White
2 = Non-White; rank 1 = police officer/detective, 2 = sergeant, 3 = lieutenant,
and 4 = captain or higher rank.
The survey also provided four specific dependent variables (based on questions 3–6, respectively) that gauge respondents’ knowledge/experiences of
crime report manipulation. The first was based on personal experience or knowledge that crime data was manipulated to make the crime numbers “look better
than they were.” A second variable measured respondent’s personal experience/knowledge that officers were “not taking complaint reports when they
should have.” A third variable focused on officers “changing words on complaint
reports to downgrade the crime.” These three questions ask respondents about
specific types of behaviors such as not taking reports and changing words to

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

11

downgrade reports. These responses were dichotomous and coded as “1” for
“yes” and “0” for “no.” A final question was a “catch-all” inquiring about “… any
unethical strategies to minimize/reduce/downgrade crime numbers ….” This
was asked in case anything about unethical strategies was missed.
The data is described to elicit respondents’ perspectives directly and clearly
as possible. Accordingly, we present descriptive statistics through a bivariate
crosstab; through visualization of the data and comparison of mean values. We
present statistical tests to show the direct relationships among different sets
of variables and retirement time periods. In addition, a multivariate logistic
regression model reveals more complex relationships among this set of variables. In order to supplement the analyses, we also offer a number of direct
quotes reflecting respondents’ NYPD experiences based on an open-ended final
question on the survey, “Please make any comments that you feel are important to understanding the NYPD based on your experience ….”

Crosstab
The first question reads, “are you confident that major crimes have declined
by 80% in New York since 1990s?” The majority of respondents were not confident in the accuracy of NYPD crime numbers widely cited by NYPD, the media
and social scientists. Indeed, 58.2% of the respondents stated “no” to this
question. This is not unexpected since some crime report manipulation has
been publicly acknowledged and some works on official statistics have indicated concerns with crime statistics (e.g. Mosher et al., 2011). However, as
we will show, the trends in our data are very important. Thus, it is not the
mere existence of manipulation but its extent, especially since 2002. Our
respondents often expressed strong feelings. For example, in the words of one
respondent who retired in 2007: “Crime is a constant. To believe that crime
has gone down 80% is a joke. I believe eventually if the Feds get involved the
truth will come out.”
We took two approaches to examining these trends. First, we created a
crosstab to analyze the relationship between the time officers retired and
whether or not they had personal knowledge of any instances in which crime
reports were changed to make crime numbers appear better. We recoded the
variable year of retirement into three groups based on time periods. The first
group consists of those who retired before Compstat began; 1993 was the cut
off year. The second worked during the early years of Compstat, and the third
in the most recent period. We found that there were 484 respondents who
retired between 1981 and 1993 and were designated as pre-Compstat category.
These officers worked during Mayors Edward Koch and/or Mayor David Dinkins.
The second group of 415 officers worked when Rudolph Giuliani was Mayor and
retired between the years 1994 and 2001. The third group of 871 respondents
consisted of officers who retired in the period from 2002 until 2012 during the
administration of Mayor Michael Bloomberg.

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

12

ETERNO ET AL.

To the specific question, “based upon your experience do you have personal
knowledge of any instance in which crime reports were changed to make crime
numbers look better than they were?”, 974 respondents stated that they were
not aware of manipulation of crime statistics. However, 78 stated that they
were aware of such manipulation 1–2 times; 309 stated they had knowledge of
this being done 3–4 times and 371 stated 5 or more times.
A bivariate crosstab analysis between period of retirement and this variable
of “personal knowledge of any instances in which crime reports were changed
to make crime numbers look better” is shown in Table 2.
In this table, we see that the majority of respondents who worked during the
Bloomberg years had personal knowledge of manipulation “to make crime
numbers look better” (55.5%). The previous two eras show a markedly smaller
proportion of officers observing manipulation. In the pre-Compstat era
[1981–1993], only 30.3% indicate personal knowledge of manipulation. While
there is a slight increase during the first Compstat era under Mayor Giuliani, it is
very similar to the previous era (34.5%). Thus, there is a striking increase in
knowledge of report manipulation during the Bloomberg era. Our measure of
association—γ = .337—supports this interpretation showing a strong relationship between knowledge of manipulation and the three periods of Compstat.
Although for the sake of brevity it is not necessary to provide the individual
crosstabs for other questions regarding personal knowledge of crime statistics
manipulation, it is important to note that this trend is exceedingly similar for
questions 4 (not taking complaint reports when they should have [γ = .288]) and
5 (changing words on complaint reports to downgrade crime [γ = .350]). Each of
these relationships showed similar associations and all were statistically significant (p < .0001) suggesting that knowledge of manipulation was related to the

Table 2 Crosstab: knowledge of crime manipulation and retirement category
Period 1
1981–1993
No
Yes

1–2 times
3–5 times
5+ times
n
Missing
N

Statistics
γ [ordinal by ordinal]
***p < .001.

Period 2
1994–2001

Period 3
2002–2012

327
(69.7%)
15
55
72
(30.3%)
469
15
484

264
(65.5%)
19
66
54
(34.5%)
403
12
415

383
(44.5%)
44
188
245
(55.5%)
860
11
871

.337***

N
974
78
309
371
1732
38
1770

POLICE MANIPULATIONS OF CRIME REPORTING

13

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

three periods of the study. There was greater knowledge of manipulation during
the last era of Mayor Bloomberg and Commissioner Kelly. As one respondent who
retired in 2005 offered, “A CO may ‘suggest’ that a burglary complaint was a
criminal trespass.” Another who retired in 2008 remarked: “A clear robbery was
split into its component parts-larceny and assault-two misdemeanors. I was told
that the c/w [complainant/witness] couldn’t ID so what’s the difference.” A
third respondent who retired in 2003 offered: “In some cases, larcenies became
‘lost property’ and values were skewed so they didn’t become grand larcenies.
Also complainants were told they must go to the stationhouse or to the precinct
of record to report a crime.”

Graphical Trends
A second way to demonstrate these trends in the data is to display officer
responses to key managerial pressures over time. To examine this, we estimated the mean values of yearly responses to the set of questions regarding
knowledge or experience of pressure to decrease index crime and to
downgrade various crimes. These were plotted according to the year of retirement beginning 1990. We note that downgrading is, by definition, suspect
activity. It involves officers attempting to place index crimes into non-index
crime categories.
Figure 1 clearly indicates the rising trend of pressure being exerted upon
the officers to reduce the index crime numbers and downgrade index crimes.
The trend lines embedded within the graph display the rising trend in pressure
for the two questions about index crimes. A second figure is used to show the
trends in pressures to obey the law and accurately report crime statistics.
Concurrently, departmental pressures to accurately report crime statistics and
to obey constitutional and legal rules show declining trends (see Figure 2).
However, for accurately reporting crime statistics there is some increase in

Figure 1

Trend in management pressures to decrease and downgrade index crime.

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

14

ETERNO ET AL.

Figure 2
rules.

Trend in management pressures to accurately report crime and obey legal

pressure initially at the beginning of Compstat but this dissipates over time.
Clearly, department managers were, during this latter period, sending out the
message that decreasing index crime rates were vastly important but, simultaneously conveying the message that maintaining accurate crime statistics and
following legal guidelines were not a major concern –a questionable combination, at best.
Comments from respondents support these trends. One respondent who
retired in 2005 commented: “Compstat instilled fear and pressure on COs to
lower index crime.” Another who retired in 2003 wrote: “Compstat was the big
thing for commanders keeping their positions and getting promoted. Changing
stats was not unusual.”

Comparison of Mean Values—Perceived Pressures by Retirement
Category
While the crosstabs and graphs offer an excellent visual representation of
responses, we further examined the data to pinpoint more precisely the pressures that respondents felt from management and/or supervisors. To do this,
we conducted a multivariate analysis of variance (MANOVA) between the three
categories of retirement years: pre-Compstat [1981–1993]; early Compstat1
[1994–2001], and most recent period [2002–2012] for the variables of knowledge or experience of pressures to manipulate crime data.
The MANOVA results were found to be significant at p-value = .0 for α .05
with Wilk’s lambda = .899 suggesting that mean differences for these set of
variables were real among the three time periods. Given the previous analysis,
this is not surprising. There are significant differences in pressure that officers
felt for decreasing index crime, decreasing other crime, downgrading index
crime, accurately reporting crime statistics, and obeying Constitutional guidelines (Tables 3 and 4).

POLICE MANIPULATIONS OF CRIME REPORTING

15

Table 3 Mean values of pressure for various items*
Retirement
category

Decrease index crime

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

Decrease other crime

Downgrade index crime to
non-index crime
Accurately report crime
statistics
Obey Legal/ Constitutional Rules

Retirement category
1
2
3

Mean

Std. dev.

N

1
2
3
1
2
3
1
2
3
1
2
3
1
2
3

Pressure type

4.46
5.45
6.36
4.36
4.9
5.43
3.77
4.16
5.63
5.25
5.75
5.28
6.35
6.59
5.95

2.929
3.051
2.98
2.677
2.692
2.781
2.943
3.05
3.308
3.007
2.861
2.888
3.041
2.859
2.854

410
371
810
410
371
810
410
371
810
410
371
810
410
371
810

Cut off years of retirement
1981–1993: Pre-Compstat
1994–2001: Post-Compstat1
2002–2011: Post-Compstat2

*Pressure measured on a Likert Scale 1–10 with 1 being the least.

Table 4 Manova results
Statistics

Value

F

Hypothesis df

Error df

Wilk’s Lamda
***p < .001

.899***

17.395

10

3190

Test of between subject effects
Dep. variable
Type III SS
Decrease index crime
1003.8
Decrease other crime
313.86

df
2
2

Mean S
501.91
156.93

F
56.41***
20.96***

Downgrade index to
Non-index crimes
Accurately report crimes
Obey rules

2
2
2

586.69
33.96
58.25

58.86***
3.99*
6.89**

1173.39
67.93
116.51

*p < .05; **p < .01; ***p < .001.

Importantly, an examination of the actual means and the post hoc results
[Games-Howell statistics] was very revealing. Recall all variables examining

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

16

ETERNO ET AL.

pressure are measured on a 1–10 Likert scale with 1 being the least pressure.
First, comparing the differences in means for reducing index crime and
decreasing other crime showed that every period had a statistically significant
difference showing markedly increasing pressure for later periods. That is, the
pressures to reduce index crime steadily increased with significant differences
in means for each time category pre-Compstat, early-Compstat, and most
recent period (M = 4.46, 5.45, and 6.36 respectively). The pressure to decrease
other crime also shows steady and significant increases over the same time
periods (M = 4.36, 4.90, and 5.43). So increasing pressure is consistent over
time for these two variables although the pressure to decrease other crime is
lower compared to decreasing index crime.
Interestingly, the pressure to downgrade index crime also steadily increases
over the time periods (M = 3.77, 4.16, and 5.63 respectively). However, the
difference between the periods 1981–1994 and 1994–2001 is not statistically
different for this variable of “downgrade index crime]. For 2002 onwards, however, the mean value significantly differs compared to the earlier two eras.
This strongly suggests that the Bloomberg era placed more pressure on officers
to downgrade than the previous two.
We now examine the stress placed upon officers regarding the importance
of accurately reporting crime statistics. For this variable, increased pressure
indicates that the agency is interested in accurate reporting. The pre-Compstat
period shows a low mean of 5.25. The department then increased pressure to
properly report crime when it first introduced Compstat (M = 5.75). This difference is statistically significant, though only barely at p = .048. So when the
department first introduced Compstat, it increased pressure to reduce index
crime but concurrently also increased pressure to ensure accuracy. However,
the increased pressure to accurately report does not hold for the most recent
Kelly/Bloomberg era but reverts back down to the mean value of 5.28. This
difference is also statistically significant with respect to the previous period of
1994–2001. However, for this era, as seen earlier, management did not concurrently reduce pressure to decrease or downgrade index crime. In sum, those
who retired in the Kelly/Bloomberg era felt significantly increased pressures to
decrease and downgrade crime and concurrently felt reduced pressures to
keep accurate crime statistics.
We see a similar effect with the variable measuring pressure to obey legal/
Constitutional rules. There is no statistical difference between pre-Compstat
[M = 6.35] and the post-Compstat category 1994–2001 [M = 6.59]. The more
recent Kelly/Bloomberg era, however, showed a significant difference between
the two previous eras [M = 5.95] with p-value being .001.

Logistic Regression Model
Four direct questions were asked regarding whether the respondents had personal
knowledge or experience of manipulating the crime reports. These were

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

17

dichotomous responses as “yes” or “no” and we examined their relationship with
the intensity of pressures to downgrade the crimes. Separate Logistic regression
models were developed for these four sets of dependent variable questions using
the independent variables of decreasing index crimes, downgrading index crimes,
reporting accurately, obeying the rules, and various individual explanatory variables. All the models had exceedingly similar results. For the sake of brevity, we
present the model for the specific question, “Based on your experience, do you
have personal knowledge of any instances in which crime reports were changed to
make crime numbers look better than they were” which is reported in Table 5.
This model was able to explain 39.0% the variance of the dependent variable
according to the Nagelkerke R squared. Furthermore, the classification table
showed that the model correctly predicted 75.0% of the responses falling in
the two groups. As may be seen, the variables “Decrease index crimes” and
“Downgrade index to non-index crimes” were good predictors of this knowledge/ experience of manipulation. It is also pertinent that for the variables
“accurately report crime statistics” and “obey rules” the slopes were negative.
This implies that these unethical strategies were predicted by a decrease in
the value of the variable. Significantly, the time variable “year of retirement”
variable was a good predictor too. As may be seen, the strongest predictor
was the “downgrade index to non-index crimes” with the “b value of .307 and
Table 5 Logit regression model
B
Decrease index crime
Downgrade index to
Non-index crimes
Accurately report crimes
Obey rules
Year retirement
Gender
Education
Marital status
Race
Rank
Constant
N = 1561 [missing 209]

Std. error

Wald stats

Exp (B)

.102***

.028

13.480

1.107

.307***
−.128***
−.051*
.038***
−.406
.082
.130
−.049
−.031
−77.672**

.025
.027
.026
.008
.234
.063
.157
.154
.076
16.978

149.957
22.230
30.855
20.551
3.004
1.668
.684
.101
.164
22.273

1.360
.880
.951
1.039
.666
1.085
1.139
.952
.970
.000

Dependent variable: Based on your experience, do you have personal knowledge of any
instances in which crime reports were changed to make crime numbers look better
than they were? Yes/No
Model coefficients
Chi-Square = 538.586***
df =10
Model Summary
−2 Log Likelihood =1612.666
Nagelkerke R2 = .390
*p < .05; **p < .01; *** p < .001.

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

18

ETERNO ET AL.

exp [b] of 1.36.” This implies that a unit change in the independent variable
“year of retirement category” affects the log of odds by a factor of .307. The
results are similar for the other specific questions about knowledge or experience of changing citizen complaints, not taking them or changing their words
to downgrade the crimes. The variables of gender, race, education, and rank
appear to have no impact upon the model, which suggests that pressures were
uniformly felt across the organization.
We tested the assumptions for the logit by examining the sample size, multicollinearity among the independent variables, and standard errors. The large
sample size n = 1538 ensured that we exceeded the minimum number of cases
per independent variable, which is suggested to be 10 (Hosmer & Lemeshow,
2000). We also tested for multicollinearity by examining the inter-correlations
among the independent variables used in logit. Except for 2 pairs with modest
correlations of .62 and .54, the remaining have very low correlations barely
reaching .2. Lastly, we examined the standard errors for the b coefficients of
every independent variable. All error values were less than 1.0 and hence
suggest that multicollinearity is not a problem (Hosmer & Lemeshow, 2000).

Discussion
It is widely known that there are weaknesses in the validity of official crime
data (e.g. Mosher et al., 2011). Bayley (1983, p. 24) directly states,
Changes in reported rates are a fraudulent measure of police effectiveness.
Although everyone who has looked into the reliability of reported crime figures
knows that they are unstable to the point of worthlessness, everyone, including
academics, continues to use them.

The current study strongly supports the validity of Bayley’s (1983) insights.
The analyses provide evidence regarding NYPD manipulation of crime statistics. We find a core set of variables that strongly predicts officers’ experiences
with report manipulation. Time of retirement is one of these variables. Time
of retirement is a good predictor of experience with report manipulation and,
as indicated by the multivariate model, has an independent effect from the
pressures on officers. One interpretation of the positive slope coefficient of
the time of retirement variable is that the management culture (community
policing in the 1980s, then Compstat in the 1990s, and continuing with even
more centralization in the 2000s) of the era has an influence on the officers’
knowledge of manipulation of crime reports—independent of pressures. The
slope coefficient is positive and significant indicating that officers had more
personal knowledge of manipulation of crime reports in recent years when the
performance management culture was at its apex. It may be that as reported
crime plummeted over 60% through 2001, it became more and more difficult
over time to sustain such decreases. It is possible that the Compstat system,
at first very productive, morphed into a culture of “gaming numbers” to keep

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

19

the decreases going. This suggests that the performance management
elements, initially healthy, must carefully be monitored. Specifically, the
top-down management style, commander morale, accountability goals or
expectations, centralization of bureaucracy, actions of management toward
subordinates, and other organizational effects are captured by this variable
(Eterno & Silverman, 2010; Weisburd, Mastrofski, McNally, Greenspan, & Willis,
2001). While future research will have to confirm this, the data certainly
reflect this interpretation. Comments from our respondents also clearly support this. Two pertinent examples from respondents who retired in 2001 and
2010, respectively, are: (1) Although Compstat started as a good project, it is
no longer that. (2) Compstat started out as a good thing and went bad.
Second, other explanations for officers’ experiences with crime report
manipulation are the various pressures they felt from management. Our
respondents indicate that the trend in pressures to decrease index crime numbers, decrease other crime and downgrade index to non-index crime have all
markedly increased. Nevertheless, the multivariate model suggests something
more complicated. Namely, these pressures have a separate influence regardless of the time period (management style). That is, as leadership in any era
exerts pressures to lower the crime rates, officers’ experiences and knowledge
of crime report manipulations are likely to increase. These increased pressures
were not due to chance alone and show reasonably high explanatory power
when predicting knowledge of crime report manipulation. Interestingly, the
level of pressure to accurately report and pressure to obey legal rules also had
independent and significant influences on knowledge of manipulation. Thus, in
these data, we see increasing pressures to keep crime numbers down and
decreasing pressures to accurately report crime and follow rules.
Additionally, the synchronicity of these variables across the data suggests
that the pressures on police are critically important to explaining the level of
knowledge of report manipulation. Our data suggest, then, that one important
factor to explaining knowledge of report manipulation or lack thereof is pressures from leadership (management), regardless of era. Given that the key
police leadership position (police commissioner) is selected by the mayor
(executive branch) in New York City (and in many departments), politics would
appear to be very important to explaining crime report manipulation.
We note that typical individual explanatory variables including gender, educational status, rank, race and marital status were not significant. Thus, individual explanatory variables played little role in explaining these behaviors. It
appears, then, that the powerful influences of the performance management
culture and the pressures on officers from leadership overwhelm individual
influences. This corresponds with research about the influence of police culture on police behavior (Drummond, 1976; Manning, 1977; Reuss-Ianni, 1983).
Comments from our respondents help confirm this interpretation. The following are a few excerpts from respondents who retired in 2002, 2003, and
2002 respectively: (1) “The pressure and humiliation attributed to Compstat
was too much for most commanding officers. The vast majority of them played

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

20

ETERNO ET AL.

the numbers game to some degree to avoid being embarrassed and degraded
in front of their peers.” (2) “Compstat was the big thing for commanders
keeping their positions and getting promoted. Changing stats was not unusual
we could see it in the squad [detectives] because we got copies of all uf61s
[complaint reports for crimes] and would see changes on a DD5 [a follow up
report that can change the classification] for a reclassification.” (3) “The
pressure from the advent of Compstat on was enormous. Commanding Officers
lived in mortal fear of being embarrassed or removed from command due to
poor performance at the podium. Otherwise ethical men were driven to cook
the books on major crimes to keep the Compstat gods appeased.”
These interpretations must await further study but the powerful explanatory
models and analyses strongly suggest our conclusions are, at a bare minimum,
logical. As with all social science, however, they are subject to limitations.
One limitation pertains to the sample. Since we do not have access to the
entire NYPD, our sampling frame is NYPD active retirees. We show trends and
other data indicating that the sample is reasonably representative but without
full access, this will necessarily be a limitation. Importantly, since active retirees volunteer to help the NYPD, this sample is more likely to be favorable to
the agency. Results ought to be interpreted considering this facet.
Another limitation is that the data are collected in a cross-sectional manner
(a one shot study). This may lead to some telescoping (i.e. respondents reporting on events that occurred in previous eras and bringing them to the more
recent time period). For the first period (1981–1993), telescoping is not a
problem since these respondents did not work during the Compstat periods
and, therefore, cannot report on events for which they were not present. To
address the telescoping issue (recalling incidents that occurred in the nonCompstat era by those who worked during both non-Compstat and Compstat),
the instructions stipulated responses only to those events that occurred in
1994 and after if respondents retired after 1994. This helps minimize any telescoping into the non-Compstat era.
We are confident that this paper accurately reflects police practices as they
occurred in New York City. Given the data and transparency issues we confronted, this line of research was the best possible option to pursue. We
expect that this study will motivate future research, replication, and provocative conversations among social scientists. Necessarily, clarification of the findings will have to await further inquiry.
This study indicates that Bayley’s (1983) warning that social scientists should
not use police reports as gauges of police success is profoundly accurate. If we
are to use crime reports for decision-making purposes (e.g. research, hiring of
officers, tactical use of officers, real estate values, etc.), far more scrutiny is
necessary. Transparency in policing would seem to be a prerequisite for this to
occur. That is, agencies would have to allow some outside bodies to closely
scrutinize their record keeping. One suggestion is professional accreditation
through bodies such as the Commission on Accreditation for Law Enforcement

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

21

Agencies (CALEA). While this is not a silver bullet, it is in our view a step in
the right direction—toward more transparency.
In sum, this study presents the results of a survey with nearly 2,000 retired
NYPD officers demonstrating the agency’s practices with respect to crime
report manipulation. Specifically, a top-down management style and managerial pressures help explain the knowledge of manipulation of index crime
reports by field officers. Experiences officers had with manipulation of crime
reports is likely a byproduct of over-reliance on crime numbers by police performance management systems as well as leadership pressures. The utility of
this study is not in questioning police crime reporting—that is well known.
Rather, it is in understanding the level of and explanations for that manipulation. We suggest more transparency in police agencies to help remedy the
situation.

References
Babbie, E. R. (1989). The practice of social research (5th ed.). Belmont, CA:
Wordsworth.
Bayley, D. H. (1983). Knowledge of the police. In M. Punch (Ed.), Control in the police
organization (pp. 18-35). Cambridge, MA: MIT Press.
Bayley, D. H. (1994). Police for the future. New York, NY: Oxford University Press.
Bernstein, D., & Isackson, N. (2014, May). The Truth about Chicago’s Crime Statistics.
Chicago Magazine. http://www.chicagomag.com/Chicago-Magazine/May-2014/Chi
cago-crime-rates/
Bourbonnais, R., Meyer, F., & Theriault, G. (1988). Validity of self reported work history. British Journal of Industrial Medicine, 45, 29-32.
Bowling, B. (1999). The rise and fall of New York murder: Zero tolerance or crack’s
decline? British Journal of Criminology, 39, 531-554.
Bradburn, N. M. (1983). Response effects. In P. H. Rossi, J. D. Wright, & A. B. Anderson
(Eds.), Handbook of survey research (pp. 289-318). New York, NY: Academic Press.
Bradburn, N. M., Rips, L. J., & Shevell, S. K. (1987). Answering autobiographical questions: The impact of memory and inference on surveys. Science, 236, 157-161.
Bratton, W. (1998). The turnaround: How America’s top cop reversed the crime epidemic. New York, NY: Random House.
Bureau of Justice Assistance, Police Executive Research Forum. (2013). Compstat: Its
origins, evolution, and future in law enforcement agencies. Retrieved November 13,
from http://policeforum.org/library/compstat/Compstat.pdf
Campbell, D. T. (1976). Assessing the impact of planned social change. Evaluation and
Program Planning, 2, 67-90. https://www.globalhivmeinfo.org/CapacityBuilding/
Occasional%20Papers/08%20Assessing%20the%20Impact%20of%20Planned%20Social%
20Change.pdf
Conklin, J. E. (2003). Why crime rates fell? Boston, MA: Allyn & Bacon.
Dillman, D. A., Smyth, J. D., & Melani Christian, L. (2009). Internet, mail and mixed
mode surveys: The tailored design method. Hoboken, NY: Wiley.
Drummond, D. S. (1976). Police culture. Beverly Hills, CA: Sage.
El-Ghobashy, T. (2011, January 7). Minorities gain in NYPD ranks. The Wall Street Journal. http://online.wsj.com/news/articles/SB10001424052748704415104576066302323
002420

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

22

ETERNO ET AL.

Eterno, J. A., & Silverman, E. B. (2010). Understanding police management: A typology
of the underside of compstat. Professional Issues in Criminal Justice, 5, 11-28.
Eterno, J. A., & Silverman, E. B. (2012). The crime numbers game: Management by
manipulation. Boca Raton, FL: CRC Press.
For 1st Time Ever, 2 Black Women in Command of NYPD Precinct. (2011, July 29). NBC
television [channel 4 news]. New York City, NY. http://www.nbcnewyork.com/news/
local/NYPD-Diversity-Black-Women-Cops-Bed-Stuy-81st-Precinct-126390033.html
Greenberg, D. F. (2014). Studying New York City’s crime decline: Methodological issues.
Justice Quarterly, 31, 154-188.
Harcourt, B. E. (2001). Illusion of order: The false promise of broken windows policing.
Cambridge, MA: Harvard University Press.
Henry, V. H. (2002). The compstat paradigm: Management accountability in policing,
business and the public sector. New York, NY: Looseleaf Law.
Hosmer, D. W., & Lemeshow, S. (2000). Applied logistic regression. Hoboken, NJ:
Wiley.
Karmen, A. (2000). New York murder mystery: The true story behind the crime crash
of the 1990s. New York, NY: NYU Press.
Karmen, A. (2014). Evaluating how the NYPD handles crime victims: Judgments based on
statistical performance measures. In Draft chapter for J. A. Eterno (Ed.), The
New York City police department: The impact of its policies and practices (pp. 47-70).
Boca Raton, FL: CRC.
Kelling, G. L., & Coles, C. M. (1998). Fixing broken windows: Restoring order and
reducing crime in our communities. New York, NY: Free Press.
Kelling, G. L., & Sousa, W. H. (2001). Do police matter? An analysis of the impact of
New York City’s police reforms (Civic Report No. 22). New York, NY: Manhattan
Institute.
Langan, P., & Durose, M. (2004). The remarkable drop in crime in New York City. Paper
delivered at the International Conference on Crime, Rome, December 3–5.
Maltz, M. (1977). Crime statistics: A historical perspective. Crime & Delinquency, 23,
32-40.
Manning, P. (1977). Police work. Cambridge, MA: MIT Press.
Maple, J. (1999). The crime fighter: How you can make your community crime free.
New York, NY: Doubleday.
Marvell, T. B., & Moody, C. E. (1996). Specification problems, police levels, and crime
rates. Criminology, 34, 609-646.
McCleary, R., Nienstedt, B. C., & Erven, J. M. (1982). Uniform crime reports as organizational outcomes. Social Problems, 29, 361-372.
Messner, S. F., Galea, S., Tardiff, K. J., Tracy, M., Bucciarelli, A., Piper, T., … Vlahov,
D. (2007). Policing, drugs, and the homicide decline in New York City in the 1990s.
Criminology, 45, 385-414.
Mosher, C. J., Miethe, T. D., & Hart, T. C. (2011). The mismeasure of crime. Los
Angeles, CA: Sage.
Neuman, W. L. (2000). Social Research Methods: Qualitative and Quantitative
Approaches (4th ed.). Boston, MA: Allyn and Bacon.
New York City Health and Hospitals Corporation. (2011). http://www.nyc.gov/html/
hhc/html/home/home.shtml
NYPD. (2013, April 8). The report of the crime reporting review committee to commissioner
Raymond W. Kelly Concerning Compstat Auditing. David Kelley and Sharon McCarthy.
http://www.nyc.gov/html/nypd/downloads/pdf/public_information/crime_reporting_
review_committee_final_report_2013.pdf
NYPD Press Release. (2009, April 15). NYPD celebrating women in policing. http://
www.nyc.gov/html/nypd/html/pr/pr_2009_ph07.shtml

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

POLICE MANIPULATIONS OF CRIME REPORTING

23

Ortiz, E. (2011, July 17). Many city agencies flout law by withholding public info from
press. amNewYork.com http://www.amny.com/urbanite-1.812039/many-city-agen
cies-flout-law-by-withholding-public-info-from-press-1.3031728
Rashbaum, W. K. (2005, April 22). Police corruption panel is losing its chairman. The New
York Times, New York, NY. http://query.nytimes.com/gst/fullpage.html?res=9D0CE7
DF1431F931A15757C0A9639C8B63http://query.nytimes.com/gst/fullpage.html?res=9D
0CE7DF1431F931A15757C0A9639C8B63
Rayman, G. (2012, March 7). The NYPD tapes confirmed. The Village Voice, New York,
NY. http://www.villagevoice.com/2012-03-07/news/the-nypd-tapes-confirmed/
Reaves, B. A. (2010, December). Local Police Departments 2007. Bureau of justice statistics, Table 22. http://bjs.ojp.usdoj.gov/content/pub/pdf/lpd07.pdf
Reuss-Ianni, E. (1983). Two cultures of policing. New Brunswick, NJ: Transaction Books.
Rosenfeld, R., Fornango, R., & Rengifo, A. F. (2007). The impact of order-maintenance
policing on New York City homicide and robbery rates: 1988–2001. Criminology, 45,
355-384.
Rosenfeld, R., & Fornango, R. (2014). The impact of police stops on precinct robbery
and burglary rates in New York City 2003-1010. Justice Quarterly, 31, 96-122.
Seidman, D., & Couzens, M. (1974). Getting the crime rate down: Political pressure and
crime reporting. Law & Society Review, 8, 457-494.
Sheehan, K. B. (2001). E-mail survey response rates: A review. Journal of ComputerMediated Communication, 6. doi:10.1111/j.1083-6101.2001.tb00117.x
Silverman, E. B. (2001). Epilogue. NYPD battles crime: Innovative strategies in policing.
Boston, MA: Northeastern University Press.
Skitka, L. J., & Sargis, E. G. (2006). The internet as psychological laboratory. Annual
Review of Psychology, 57, 529-555. Cited by Mosher, Miethe, and Hart (2011).
Stewart, W. F., Tonascia, J. A., & Matanoski, G. M. (1987). The validity of questionnaire-reported work history in live respondents. Journal of Occupational Medicine,
29, 795-800.
Unions Call for Crime Stat Audit and Crime Reporting Policy Change. (2004, March 23).
http://nycpba.org/archive/releases/04/pr040323-stats.html
Weisburd, D., Mastrofski, S. D., McNally, A. M., Greenspan, R., & Willis, J. J. (2001).
Compstat and organizational change: Findings from a national survey (Report submitted to the National Institute of Justice by the Police Foundation). Washington,
DC: Department of Justice.
Xie, M. (2014). Area differences and time trends in crime reporting: Comparing New
York with other metropolitan areas. Justice Quarterly, 31, 43-73.
Zelinski, E. M. (1999). Self-reported memory and performance in the oldest-old; comparisons between a representative and a convenience sample. The Gerontologist.
Special Issue, 39, 484.
Zimring, F. E. (2011). The city that became safe: New York’s lessons for urban crime
and its control (Studies in Crime and Public Policy). New York, NY: Oxford University
Press.

24

ETERNO ET AL.

Appendix

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

E-mail Invitation
__________________is the Principal Researcher on a study of the New York
City police department’s (NYCPD) management. It is requested that you take
part in a short survey that should take no more than 5 min to complete about
your personal experience as a New York City police officer. There are a total
of nineteen (19) questions. It is anonymous (there is no way for us to know
your identity) and voluntary. You indicate your voluntary consent by answering
the survey. The NYCPD has no involvement in this research. This is an independent, scientific study.
Your answers may influence policy decisions and results will be published in
press reports and scientific journals.
Please follow the link below to take the short survey. Let me thank you for
your help.

Link to survey
If you have any questions, feel free to contact me.

Contact information
This email and any attachments may contain confidential and privileged information. If you are not the intended recipient, please notify the sender immediately by return email, delete this message and destroy any copies. Any
dissemination or use of this information by a person other than the intended
recipient is unauthorized and may be illegal.

NYCPD Survey*
The Wording of the Directions
Directions: All responses are voluntary (you do not have to answer questions).
You may skip questions you feel unsure about or do not wish to answer. You
indicate your voluntary consent by answering the survey. The survey is anonymous (there is no way for us to know your identity).
For all questions, if you retired before 1994, base your answers on your
overall impressions over your entire career; if you retired in 1994 or after,
base your answers on experiences that occurred only in 1994 and after. Choose
the responses that best reflect your personal experiences.

POLICE MANIPULATIONS OF CRIME REPORTING

25

The Wording of the Questions
(1)
(2)

(3)

Downloaded by [Molloy College], [Jhn Eterno] at 06:42 20 November 2014

(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)

Based on your personal experience are you confident that major crime
has decreased 80% since 1990 in New York City?
With respect to the following criteria and based on your personal experience/knowledge, on a scale of 1 to 10 (with 1 being the least and 10 the
most), how much pressure did precinct (patrol) personnel receive from
management/supervisors to:
Based on your experience, do you have personal knowledge of any
instances in which crime reports were changed to make crime numbers
look better than they were?
Based on your experience, do you have personal knowledge of officers/
supervisors not taking complaint reports when they should have?
Based on your experience, do you have personal knowledge of officers/
supervisors changing words on complaint reports to downgrade the crime?
Based on your experience, do you have personal knowledge of any unethical strategies used to minimize/reduce/downgrade crime numbers
(reported crimes)?
What is your gender?
What is your current age?
What is the highest level of formal education you have completed?
What is your marital status?
What is your race/ethnicity?
What was your rank when you retired?
How many years of service did you do for the NYPD?
Do you have military experience?
What year did you retire?
Did you have a patrol (precinct) assignment after 1994?
Were you ever the commanding officer of a precinct?
Overall, how would you characterize your work experience in the following years?
Please make any comments that you feel are important to understanding
the NYPD based on your experiences. Please do not indicate your name or
other identifying information in your response.

*The above only shows the wording of the initial directions and questions for
respondents. It includes neither the formatting of the questionnaire nor the
response possibilities. Readers may request the full survey from the authors.