Police Use-of-Force Data Where We Are and Where We Should Be Going
Police Use-of-Force Data Where We Are and Where We Should Be Going
http://pqx.sagepub.com/
Published by:
http://www.sagepublications.com
On behalf of:
Police Executive Research Forum
Police Section of the Academy of Criminal Justice Sciences
Additional services and information for Police Quarterly can be found at:
Subscriptions: http://pqx.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://pqx.sagepub.com/content/2/1/57.refs.html
GEOFFREY P. ALPERT
University of South Carolina
MICHAEL R. SMITH
Virginia Commonwealth University
Support for this research has been provided, in part, by the National Institute of Justice grant number
98-IJ-CX-0018. Opinions stated in this article are those of the author and do not necessarily represent the
official position of the National Institute of Justice.
POLICE QUARTERLY Vol. 2 No. 1, March 1999 57
57-78
@ 2000 Sage Publications, Inc.
at the use of force and its justification, levels, and methods. Unfortunately, it
is difficult to explain, describe, or measure what level of force is necessary
and reasonable in all situations (Alpert & Smith, 1994). However, research
has been conducted on these various aspects of force, and a series of find-
ings and recommendations have been presented in the literature. What we
know about police use of force has been based upon research conducted in a
limited number of jurisdictions and by a limited number of methodologies.
What is missing from our thinking about police use of force is an under-
standing of the type of agencies from which the data have been collected
and an analysis of the relevance of the data upon which research findings
have been based.
The results of past research efforts have been presented in other places
and will not be reviewed here (see Geller & Toch,1995). This article will fo-
cus on the various methods used to collect use-of force information. Police
managers and researchers often have different interests and needs concern-
ing information on the organization and collect only the types of informa-
tion important to them. On one hand, managers are looking to improve orga-
nizational efficiency and effectiveness and do not always need a high level
of specificity. On the other hand, researchers may require data elements that
are different from managers, and they may require more detailed informa-
tion that can be used to explain variance in the level of force used or that can
help them describe or infer from the events under study. Police use of force
is an area of interest to both managers and researchers but one that suffers
from a lack of understanding, empirical information, and valid measures. In
1981, The U.S. Commission on Civil Rights noted,
Police officers possess awesome powers. They perform their duties under hazardous
conditions and with the vigilant public eye upon them. Police of6cers are permitted
only a small margin of error in judgment under conditions that impose high degrees of
physical and mental stress. Their general responsibility to preserve the peace and en-
force the law carries with it the power to arrest and to use force-even deadly force. It
is essential, therefore, that these sweeping powers be subject to constant scrutiny to
ensure that they are not abused.
The term used by the U.S. Commission, constant scrutiny, can be oper-
ationalized in many ways. In any case, scrutiny translates into a broad move-
ment of accountability and requires, at a minimum, that the actions taken by
officers are recorded and analyzed. The recording and analysis of officers’ use
of force is a necessary tool to control any excessive actions. Incredibly, al-
though police agencies authorize their officers to use force against citizens,
agencies nor always suitable to use as managerial or research data. This arti-
cle will begin with a brief description of the legal and policy standards that
govern police use of force. It will then turn to a discussion of use-of-force
information, the various ways this information is collected, and the biases
inherent in the various collection methods. Finally, we will present sugges-
tions for an improved data-collection strategy, which include a system that
has a supervisor document all aspects of police-citizen contacts where force
is used. Under this suggested system, the supervisor should receive input
from all involved parties, including the officer(s), suspect(s), and witness(es).
In 1989, the U.S. Supreme Court decided Graham v. Connor and set forth
the constitutional limitations on the use of force by police. The plaintiff in
Graham was a diabetic who was injured during a violent encounter with
Charlotte, North Carolina police officers. He sued the police department
and the officers involved alleging that his due process rights under the 14th
Amendment had been violated. The Supreme Court held that all claims of
excessive force by the police must be judged under the reasonableness stan-
dard from the 4th Amendment. Specifically, the Court held that police offi-
cers may use no more force than an objectively &dquo;reasonable officer&dquo; would
have used under the circumstances. The Court declared that the reasonable-
ness of an officer’s use of force depends upon the severity of the crime at
issue, whether the suspect poses an immediate threat to the safety of the offi-
cers or others, and whether the suspect is actively resisting arrest or attempt-
ing to flee.
Because the reasonableness of an officer’s actions depends upon the
unique facts and circumstances of each case, it is difficult to draw &dquo;bright
line&dquo; rules that separate legally reasonable force from illegally excessive
force. However, in an effort to provide their officers with guidance on the
use of force, many law enforcement agencies have incorporated force con-
tinua into their use of force policies. These continua present officers with a
series of escalating steps in the use of force that they are required to follow
whenever possible (Connor, 1991). Table 1 is an example of a common
USE-OF-FORCE INFORMATION
Police
departments that collect and maintain use-of-force information
generally compile it for administrative and bureaucratic purposes (Adams,
1995). Historically, agencies collected only limited information on the use
of deadly force. These data included information on shots fired purposely.
Many agencies collected information only on shots fired that hit their
intended targets. Some agencies included only that shots were fired, not the
number of shots, situational, environmental, personal data, or information
on precipitating events (Alpert & Fridell, 1992). As a result of questions and
complaints from citizens, the media, and the courts, agencies began to col-
lect more complete and comprehensive information on all shots fired. For
Data collected from agency records have drawbacks.... First of all, official data from
any governmental agency reflects the agency perspective and may or may not fully
represent reality. Relatedly, researchers who have studied force within individual de-
partments, most likely achieved access to the most progressive departments, which
were more amenable to scrutiny. Another problem is that information from depart-
mental records on force are frequently incomplete and variations across departments
may be so great that it could be inappropriate to assume that the data from an individ-
ual department, or several departments, is representative of law enforcement in the
United States. (p. 24)
The use of force against the suspect is justified by the men on the basis that they are
acting in the interests of the community. They feel that they are only using the meth-
ods of the suspect, and that in doing so they frequently apprehend men who are a dan-
ger to the people in the community.... It is likely that the individual men, faced with
specific situations, make their decisions in terms of a complex of factors that are not
easily disentangled but among which their desire for prestige, their feeling that they
are performing a community service, and their definition of the criminal as having ab-
rogated his rights as a citizen form a substantial basis for the use of force (Westley,
1970, pp. 129-130).
INCIDENT REPORTS
cer to control the suspect and any actions taken by the suspect to resist the
officer. Utilizing a combination of preselected categories and narrative, an
officer will provide a detailed description of what happened during the
encounter. It is important to include a sequential account of all actions taken
to control the suspect.
Suspect resisted by pushing me away and taking a swing and hitting me in the head. I
tackled the suspect who hit his head on the side of a chair as he fell to the ground. The
suspect was handcuffed and taken into custody without further incident. His injury
was treated at the scene by paramedics.
Other officers at the scene could write that they observed a scuffle and
saw the suspect and officer fall to the floor but were not in a position to see
how the injury occurred. This type of report could be read by a supervisor
and approved as reasonable or necessary force with an unfortunate but un-
avoidable injury. If, however, the suspect were able to provide his or her ver-
sion, it might be reported that the officer shoved the suspect, tackled him,
and hit him on the side of the head with his flashlight while they were on the
ground. Certainly, each party has a vested interest in the official or accepted
version of an incident. Therefore, it is important to have both parties tell
their story.
Although a citizen can file a complaint against an officer, the process may
be time consuming and difficult for the citizen. Collecting information from
both the officer and the suspect involved in use-of-force situations may pro-
vide the agency commanders with the most comprehensive understanding
of the incident. The options for this type of report include having the officer
provide a narrative and having the suspect also provide a narrative to the
officer. Obviously, the suspect would not feel completely comfortable with
and trusting of an officer with whom he has just been fighting. A good alter-
native is for the officer’s supervisor to complete a Supervisor’s Control of
Persons Report.
A SUPERVISOR’S CONTROL
OF PERSONS REPORT
Over the years, researchers have used a variety of methods to assess how
frequently and under what conditions police officers use force. The sections
that follow describe each of these methods and discuss their various
strengths and weaknesses.33
with using these types of records for research purposes. To begin with,
many police agencies do not routinely capture detailed information on the
use of force, if they collect the data at all. With more than 17,000 police
agencies in the United States and with the majority of those having 10 offi-
cers or fewer, there is wide variation in the availability of official records
involving use of force. If use-of-force incidents are recorded at all, there will
likely be significant variability among agencies in the thoroughness and
accuracy of the data collected.
Where use-of-force data are available from official records, they are also
likely to vary widely in their scope. For example, in Alpert and Dunham’s
study, data from Dade County was taken from use-of-force reports routinely
prepared following use of force incidents. Data from Oregon, on the other
hand, were drawn from a job task analysis instrument designed primarily to
identify the physical abilities required for police work (Farenholtz, Peak, &
Rhodes, 1995). Although the two sets of data shared some common charac-
teristics, they were, in many ways, noncomparable (Alpert & Dunham,
1998). Agency use-of-force data will vary according to the definitional cri-
teria established by each agency, for when a use-of-force report must be
completed, the forms themselves will use different definitions and classifi-
cations, and they will differ according to the variables for which informa-
tion is recorded. Given the current state of affairs, with no uniform system in
place for capturing use-of-force information, researchers are left with
site-specific data that has limited external validity.
CITIZEN SURVEYS
In 1996, for the first time, the Census Bureau, in conjunction with the
Bureau of Justice Statistics (BJS), conducted a nationally representative
survey of 6,421 persons older than the age of 11to determine the nature and
extent of their contacts with the police during the previous year (Greenfield,
Lanagan, & Smith, 1998). Among the questions asked of the respondents
was whether the police had used force against them. Of the persons sur-
veyed, only 14 reported that they had been the recipients of police use of
force. Although BJS extrapolated these cases to the population and esti-
mated that approximately 500,000 people had force used against them the
previous year, this figure contains a large margin of error because it was
based on so few responses. Acknowledging its limitations, the 500,000 per-
son figure represents less than 1 % of police-citizen contacts for the year.
As pointed out, the accuracy of the survey results regarding use of force
are questionable because of the small number of responses. In addition, the
survey asked respondents whether police used or threatened to use force
against them without specifying or defining what constitutes force. In sub-
sequent surveys, respondents were presented with a set of police actions
(officer pushed, hit, threatened you with a flashlight, etc.), and respondents
were asked to identify all that apply (Greenfield, Lanagan, & Smith, 1998).
This change helped researchers report more precisely the nature of the force
used by police against citizens.
Without a substantial increase in the sample size, however, the results
of the survey with respect to use of force will still contain a significant pos-
sibility of error. All indications are that police use of force, when measured
as a percentage of police-citizen contacts, is an exceedingly rare event
(Bayley & Garofalo, 1989; Greenfield, Lanagan, & Smith, 1998; Reiss,
1971). Measuring social phenomena with low baseline rates of occurrence
poses unique challenges for survey researchers using random sampling.
Moreover, as Adams (1995) points out, it is not clear whether police-citizen
contacts are the appropriate denominator to use when evaluating the preva-
lence of police use of force, or whether some other denominator, such as
arrests, would be more useful.
OFFICER SURVEYS
FIELD OBSERVATION
primary focus of the study, the researchers will gather data on use of force as
part of their overall research effort. When available, the data from this pro-
ject should yield valuable information on contemporary police use-of-force
practices.
~ if a policeman physically assaulted a citizen and then failed to make an arrest: proper
use involves an arrest;
~ if the citizen being arrested did not, by word or deed, resist the policeman: force
should be used only if necessary to make the arrest;
~ if the policeman, even though there was resistance to the arrest, could easily have re-
strained the citizen in other ways;
~ if a large number of policemen were present and could have assisted in subduing the
citizen in the station, in lockup, and in the interrogation rooms;
~ if an offender was handcuffed and made no attempt to flee or offer violent resistance;
and
~ if the citizen resisted arrest, but the use of force continued even after the citizen was
subdued.
In his reanalysis of the Reiss data, Friedrich (1980) used the judgment of
the original coders rather than the observers to gauge the extent of excessive
force. The coders labeled force as excessive if (a) it was not required to
make an arrest or (b) it was not required for an officer to defend himself. Of
the 1,565 observed encounters where police regarded citizens as offenders,
28 incidents (1.8%) involved the use of excessive force.
Determining how frequently police use excessive force largely depends
on how excessiveness is defined. For example, the figures reported by Reiss
(37 instances of excessive force) are probably higher than the figures
reported by Friedrich (1980) (28 instances of excessive force) because
Reiss’ observers used a broader definition of excessiveness than his coders.
As many researchers and police officials have noted, excessive force is diffi-
cult to define and depends upon the idiosyncrasies of a given police-citizen
encounter (Adams, 1995; Alpert & Smith, 1994; Klockars, 1995; Worden,
1995).
Because of the difficulties in precisely defining excessive force, the defi-
nitions used by both the observers and the coders in the Reiss study are
unsound and simply do not reflect the realities of police work or the current
legal and policy standards for excessive force. As Klockars (1995) points
out, police sometimes use force quite appropriately against citizens without
arresting them.6 Consequently, labeling force as excessive if it did not occur
in conjunction with an arrest (as the observers’ first criterion does) may
have overstated the prevalence of excessive force. However, the use of force
may have been required by an officer to defend him or herself (one of the cri-
teria used by the coders), yet may also have been more extreme than was
reasonably necessary under the circumstances. Under the prevailing legal
standard of reasonableness,’ such force would be excessive, yet it may not
have been labeled as such by Reiss’s coders.
Similar to the Reiss study, the police services study contains data on the
use of excessive force. Of the 59 recorded instances of force being used by
the police in that study, observers categorized 23 of the incidents as exces-
sive. In deciding whether the force used was reasonable or excessive, the
observers were apparently told to label force as excessive if the officer was
&dquo;kicking ass&dquo; (Worden, 1995). This definition is dubious on its face, thus
rendering highly suspect any conclusions that might be drawn from the
study about the prevalence of excessive force. In addition to their weak-
nesses in defining excessive force, both the Reiss study and the police ser-
vices study are quite dated, and neither contain a sufficient number of force
incidents upon which to draw reliable conclusions.
The field observation studies point to some obvious difficulties in mea-
suring how frequently police use excessive force. At some point, either
when the event is recorded by the observer or later when it is coded for the
purpose of analysis, a judgment must be made concerning whether the force
used in a particular encounter was excessive. This judgment is crucial, and
yet in the early observational studies, it was made by research assistants
who utilized highly suspect definitions of excessive force. This technique
CITIZEN COMPLAINTS
When police are sued for allegedly using excessive force, both the plain-
tiff and the defendant officer typically rely on expert witnesses to assist the
fact finder in determining whether the officer acted appropriately. The role
of the expert is to analyze the facts of the case and to provide an opinion as to
whether the force used by the officer was appropriate under the circum-
stances. Police excessive force cases, like those alleging medical malprac-
tice, involve potentially harmful activities conducted by persons with spe-
cialized training and expertise. Because the use of force is a unique police
activity that requires specialized knowledge and skill beyond the
experience of the typical juror, use-of-force experts play a vital role in help-
ing jurors (and even judges) reach a considered judgment regarding the rea-
sonableness and legality an officer’s conduct.
Prior studies that attempted to categorize police use of force as excessive
relied upon faulty definitions of excessive force applied by laypersons who
lacked the experience and training to make judgments about the appropri-
ateness of the force used. We propose that future researchers use a panel of
police use-of-force experts to review force incidents captured as part of a
research effort. These panels would perform a similar function to that of an
expert witness in a police use-of-force case. They would review the facts
surrounding use-of-force incidents, apply contemporary standards of
police conduct to those facts, and decide whether the officers acted appro-
priately in using the force.
Using the Miami-Dade model, for example, a panel could examine the
use of force reports and accompanying narratives prepared by field supervi-
sors to determine whether the force used during the encounters comported
with contemporary legal and policy standards regarding the use of force.
This approach should provide a much more accurate picture of how fre-
quently police officers use excessive force. The findings of these panels
could be used as part of a basic research effort on police use of force, as well
as by police agencies in identifying problems to be addressed through train-
CONCLUSION
Police have been granted the awesome power to use force to control sus-
pects and to maintain order. Unfortunately, many of those agencies whose
officers use force do not keep information on its application. This has led
researchers to base their work on a limited number of agencies that keep
proper data and that have allowed them to review and analyze them. Much
more could be learned about police use of force if agencies would collect
information, but it is subject to reporting bias and both careless and inten-
tional error. Certainly, the most appropriate method currently used by agen-
cies is similar to the model used by the Miami-Dade Police Department dis-
cussed above. This method of having a supervisor write a report based upon
interviews with participants and witnesses adds a level of oversight into the
reporting process. Although this method has its potential problems, it
allows for several versions of the incident as experienced and observed by
those involved. Reporting the suspect’s version of the encounter is a critical
component to a use-of-force management system.
Once an official version or official versions of the incident is (are) col-
lected, then an expert panel can review and analyze the decisions made by
the officer(s) based on training, prevailing legal standards, and policy. These
assessments should form the basis of the agencies’ control and management
of the use of force and should guide policies and training decisions.
Once agencies systematically begin to collect and analyze their use-of-
force data, researchers will be able to develop theoretical models of police
behavior, statistical models to test those theories, and practical information
to assist agencies in developing or altering their existing training regimens
and policies. Unless and until police agencies take the initiative to collect
this important information, police managers and researchers must continue
to rely on a limited number of agencies that are willing to have their records
opened and scrutinized.
NOTES
1. Many accreditation organizations require that law enforcement agencies collect
use-of-force information (see Commission for the Accreditation of Law Enforcement
Agencies [CALEA] standard 1.3.5a-d).
2. For example, the Miami-Dade Police Department has been criticized for crimes com-
mitted against tourists. Had the agency not maintained a separate category of tourist-related
crimes, it could not have responded properly to the criticism. Similarly, some agencies are
criticized for hate crimes and must respond. Interestingly, the most stinging criticisms are
directed at those agencies that record and maintain data on specific crime categories. The
best responses are also from those agencies that collect the data.
3. A detailed discussion of prior research efforts to measure police use and abuse of force,
as well as the methodological challenges faced by police use-of-force researchers, can be
found in Adams (1995).
4. This category contains entries for the following types of force: hands/arms, fist,
foot/leg, handgun, shotgun, radio, flashlight, PR-24 baton, K-9, chemical agent, special
weapon, and lateral neck restraint.
5. Categories include threaten, grab/hold, push/pull, strike/hit, bite, throw, discharge, and
restrain.
6. Simply stepping between two drunken bar patrons about to engage in a fistfight and
pushing each away from the other may technically qualify as an assault but would not neces-
sarily (or appropriately) involve an arrest.
7. The Supreme Court defines excessive force as more force than a reasonable officer
would use under the circumstances ( Graham v. Connor
, 1989).
REFERENCES
Adams, K. (1995). Measuring the prevalence of abuse of force. In W. Geller & H. Toch
(Eds.), And justice for all: Understanding and controlling police abuse of force (pp.
61-97). Washington, DC: Police Executive Research Forum.
Alpert, G., & Dunham, R. (1998). An analysis of police use offorce data. Final report. Wash-
ington, DC: National Institute of Justice.
Alpert, G., & Fridell, L. A. (1992). Police vehicles and firearms: Instruments of deadly force.
Prospect Heights, IL: Waveland.
Alpert, G., & Smith, W. (1994). Developing police policy: An evaluation of the control prin-
ciple. American Journal of Police, 13, 1-20.
Alpert, G., Smith, W. & Watters, D. (1992). Implications of the Rodney King beating. Crimi-
nal Law Bulletin, 28
, 469-479.
Bayley, D., & Garofalo, J. (1989). The management of violence by police patrol officers.
Criminology, 27
, 1-27.
Connor, G. (1991, March). Use of force continuum: Phase II. Law and Order, 30-32.
Croft, E., & Austin, B. (1987). Police use of force in Rochester and Syracuse, New York
1984 and 1985. Report to the New York State Commission on Criminal Justice and the
Use of Force (Vol. 3). Albany: New York State Commission on Criminal Justice and the
Use of Force.
Farenholtz, D., Peak, K., & Rhodes, E. (1995). Police officer’s physical abilities study.
Abbotsford, British Columbia, Canada: D.W.F. Consultants.
Friedrich, R. J. (1980). Police use of force: Individuals, situations, and organizations. Annals
of the American Academy of Political and Social Science, 452, 82-97.
Garner, J. (1995) Use of force by and against the police (95-IJ-CX-0066).
Garner, J., Buchanan, J., Schade, T., & Hepburn, J. (1996). Understanding the use of force by
and against the police. Final report. Washington, DC: National Institute of Justice.
Geller, W. A., & Toch, H. (Eds.). (1995). And justice for all: Understanding and controlling
police abuse . of force Washington, DC: Police Executive Research Forum.
Graham v. Connor, 490 U.S. 386 (1989).
Greenfield, L., Lanagan, P., & Smith, S. (1998). Police use of force: Collection of national
. Washington, DC: Bureau of Justice Statistics.
data
Hagan, F. (1997). Research methods in criminal justice and criminology (4th ed.). Boston:
Allyn and Bacon.
Hunt, J., & Manning, P. (1991). The social context of police lying. Symbolic Interaction, 14,
51-70.
Kappeler, V., Sluder, R., & Alpert, G. (1998). Forces of deviance: Understanding the dark
side of policing. Prospect Heights, IL: Waveland Press.
Klockars, C. (1995). A theory of excessive force and its control. In W. Geller & H. Toch
(Eds.), And justice for all: Understanding and controlling police abuse of force (pp. 11-
30). Washington, DC: Police Executive Research Forum.
Pate, A., & Fridell, L. (1993). Police use of force: Official reports, citizen complaints, and
legal consequences (Vol. 1). Washington, DC: Police Foundation.
Reiss, A. (1968). Police brutality—Answers to key questions. Trans-action, 5
(8), 10-19.
Reiss, A. (1971). The police and the public. New Haven: Yale University Press.
U.S. Commission on Civil Rights. (1981). Who is guarding the guardians: A report on police
practices. Washington, DC: Author.
Walker, S., & Graham, N. (1998). Citizen complaints in response to police misconduct: The
results of a victimization survey. Police Quarterly, 1 , 65-89.
Westley, W. A. (1970). Violence and the police. Cambridge, MA: The MIT Press.
Worden, R. (1995). The "causes" of police brutality: Theory and evidence on police use
of force. In W. Geller & H. Toch (Eds.), And justice for all: Understanding and control-
ling police abuse of force (pp. 31-60). Washington, DC: Police Executive Research
Forum.