Professional Documents
Culture Documents
Identifying Situational Determinants of Police Use of Force
Identifying Situational Determinants of Police Use of Force
Reporting Descriptive Statistics for the Project Identifying Situational Determinants of Police
Use of Force: A Systematic Social Observation of Body Camera Footage in Newark, NJ
Victoria A. Sytsma
Department of Sociology
Queen’s University at Kingston
Eric L. Piza*
John Jay College of Criminal Justice
City University of New York
Vijay F. Chillar
School of Criminal Justice
Rutgers University
June, 2020
*Corresponding author: John Jay College of Criminal Justice, 524 West 59 th Street, Haaren Hall 636.15, New York,
NY 10019, epiza@jjay.cuny.edu, +1 347-850-7048
ABSTRACT
Results of a 2011 U.S. Department of Justice Civil Rights Division investigation into the Newark
Police Division (NPD) confirmed unconstitutional pattern of practice, which included excessive
use of physical force. This investigation resulted in a Department of Justice consent decree,
requiring the deployment body-worn cameras (BWC) and revised training on use of force in an
attempt to prevent unnecessary uses of physical force. This report capitalizes on the proliferation
of those BWCs to develop in-depth descriptions of police use of force events recorded on video
between December 2017 and December 2018. Such reporting addresses the need to produce
“strategic information products” that describe the context of use of force events, the suspect and
officer characteristics, and the characteristics of the use of force encounter to help inform police
strategies. Results indicate that suspects were commonly under the influence of drugs or alcohol,
less than 20% were in possession of a weapon, and a substantial proportion of use of force cases
involved domestic violence. Bystanders were commonly present during use of force events. Each
of these findings speak to the potential for volatility in many of use of force cases. Officers often
did not announce the presence of a BWC, which runs counter to NPD policy, but for those who
did, they did so shortly after arriving on scene. It was quite common for at least one officer to
display verbally antagonistic behavior, and to do so within less than 2 minutes of interacting with
the suspect. Displays of such behavior were typically short in duration and the force recipient was
almost exclusively the target of such behavior. That said those officers who permitted suspects to
speak for the purpose of expressing their views were consistent in their adherence to procedural
justice principles in that they did tend to address suspect concerns. While the median number of
minutes from the start of the force event to the time at which force was used generally increased
with the severity of the force type, threat of lethal force is an exception to this trend. Generally,
ii
most officers adhered to the NPD policy of using escalating verbal commands before resorting to
physical force. Such descriptive findings can provide practitioners with information products
iii
TABLE OF CONTENTS
ABSTRACT _____________________________________________________________ii
RESULTS _______________________________________________________________6
REFERENCES ___________________________________________________________51
iv
LIST OF TABLES
Table 1. Frequency and percent use of force events, by setting indicators. ____________7
Table 2. Descriptive statistics use of force time event indicators in minutes. ___________9
Table 3. Frequency and percent use of force events, by event indicators. _____________10
Table 5. Descriptive statistics number of suspects and number of bystanders on scene. ___11
Table 6. Frequency and percent use of force events, by type of bystanders and presence of victim
on scene. ________________________________________________________________11
Table 7. Descriptive statistics number of uninvolved bystanders and minutes until uninvolved
bystanders arrive, by time series. _____________________________________________12
Table 10. Frequency and percent use of force events, by suspect attributes. ____________16
Table 11. Frequency and percent use of force events, by suspect weapon possession. ____17
Table 12. Descriptive statistics minutes until suspect weapon discovered, by time series. _17
v
Table 13. Frequency and percent use of force events, by suspect conflict and flee attempt.
________________________________________________________________________19
Table 14. Descriptive statistics minutes until suspect attempted to flee, by time series. ___19
Table 15. Frequency and percent use of force events, by suspect displaying verbally antagonistic
behavior. ________________________________________________________________20
Table 16. Descriptive statistics minutes until suspect displayed verbally antagonistic behavior
and duration of behavior, by time series. _______________________________________20
Table 17. Frequency and percent suspect verbal antagonistic behavior target, by time series.
________________________________________________________________________21
Table 18. Frequency and percent use of force events, by suspect displaying physically
antagonistic behavior. ______________________________________________________22
Table 19. Descriptive statistics minutes until suspect displayed physically antagonistic behavior,
by time series. ____________________________________________________________22
Table 20. Frequency and percent suspect physical antagonistic behavior target, by time series.
________________________________________________________________________23
Table 21. Frequency and percent use of force events, by officer demographics. _________24
Table 22. Descriptive statistics number of officers on scene and minutes until officers arrive, by
time series. ______________________________________________________________25
Table 23. Frequency and percent use of force events, by officer announcing presence of BWC
and BWC falling off._______________________________________________________25
Table 24. Descriptive statistics minutes until officer announces presence of BWC and number of
officers wearing BWC. _____________________________________________________26
vi
Table 25. Frequency and percent use of force events, by procedural justice indicators. ___28
Table 26. Descriptive statistics minutes until suspect permitted to speak for the purpose of
expressing views and duration of suspect speaking, by time series. __________________28
Table 27. Frequency and percent suspects attempting to speak for the purpose of expressing
views, by officer allowing suspect to speak, officer addressing concerns, and by time series.
________________________________________________________________________29
Table 28. Frequency and percent use of force events, by officer displaying verbally antagonistic
behavior. ________________________________________________________________30
Table 29. Descriptive statistics minutes until officer displayed verbally antagonistic behavior and
duration of behavior, by time series. ___________________________________________30
Table 30. Frequency and percent suspect verbal antagonistic behavior target, by time series.
________________________________________________________________________31
Table 31. Frequency and percent use of force events, by officer who used force in the first
instance, final level of force and whether or not the force was captured on camera in the distance.
________________________________________________________________________32
Table 32. Frequency and percent use of force events, by officer giving calm command. __34
Table 33. Descriptive statistics minutes until officer calm command and duration of command,
by time series. ____________________________________________________________34
Table 34. Frequency and percent officer who gave calm command, by time series. ______35
Table 35. Frequency and percent suspect compliance with calm command, by time series.
________________________________________________________________________36
Table 36. Frequency and percent use of force event, by officer giving shout command. __37
vii
Table 37. Descriptive statistics minutes until officer shout command and duration of command,
by time series. ____________________________________________________________37
Table 38. Frequency and percent officer who gave shout command and suspect compliance with
shout command, by time series. ______________________________________________38
Table 39. Frequency and percent use of force events, by officer using soft, empty hand control.
________________________________________________________________________39
Table 40. Cross-tabulations use of force events with soft, empty hand control, by hard, empty
hand control and by threat of lethal force. ______________________________________39
Table 41. Descriptive statistics minutes until officer soft, empty hand control, by time series.
________________________________________________________________________40
Table 42. Frequency and percent officer who used soft, empty hand control and suspect
resistance of soft, empty hand control, by time series. _____________________________41
Table 43. Frequency and percent use of force events, by officer using hard, empty hand control.
________________________________________________________________________42
Table 44. Descriptive statistics minutes until officer hard, empty hand control, by time series.
________________________________________________________________________42
Table 45. Frequency and percent officer who used hard, empty, hard hand control and suspect
resistance of hard, empty hand control, by time series. ____________________________43
Table 46. Frequency and percent use of force events, by officer using blunt impact control.
________________________________________________________________________44
Table 47. Descriptive statistics minutes until officer blunt impact control, by time series. _44
Table 48. Frequency and percent officer who used blunt impact control and suspect resistance of
blunt impact control, by time series. ___________________________________________44
viii
Table 49. Frequency and percent use of force events, by officer using chemical control. __45
Table 50. Descriptive statistics minutes until officer chemical control, by time series.____45
Table 51. Frequency and percent officer who used chemical control and suspect resistance of
chemical control, by time series.______________________________________________45
Table 52. Frequency and percent use of force events, by officer using threat of lethal force.
________________________________________________________________________46
Table 53. Descriptive statistics minutes until officer threat of lethal force, by time series. _46
Table 54. Frequency and percent officer who used hard, blunt impact control and suspect
resistance of blunt impact control, by time series. ________________________________46
ix
INTRODUCTION AND STUDY OBJECTIVES
Police are defined by the coercive powers they are afforded by the state (Bolger, 2015). In
exchange for this coercive power, police are expected to exercise their ability to use force only
when necessary, and in a reasonable manner that is proportional to the threat (Ariel, Farrar, &
Sutherland, 2015; Reiss, 1968). However, the police and the public do not always agree on what
is an acceptable police practice (Reiss, 1968), and perceived excessive or unnecessary force runs
(see Murphy, Hinds, & Fleming, 2008). When the public views agents of the law unfavorably, the
ability of such agents to maintain social order is lost (Kirk & Matsuda, 2011), a concept known as
legal cynicism (Sampson & Bartusch, 1998, p. 778; see also Kirk & Papachristos, 2011).
Understanding how force erupts is essential to curbing it or at least ensuring force is used
only when necessary, and in a procedurally just manner. Prior research on police use of force at
the situational level has not provided detailed accounts of those actions that took place prior to the
“final-frame decision” (White, quoted in Lum, 2015, p. 8), nor has it provided detailed accounts
of how such final-frame decisions reach denouement. Rather, prior research on police use of force
at the situational level has focused on the impact of citizen demeanor as indicated by resistance on
force outcomes (Alpert, Dunham, & MacDonald, 2004; Garner, Maxwell, Heraux, 2006; Terrill,
2003). The proliferation of body-worn cameras (BWCs) provides opportunities to develop in-
depth descriptions of police-citizen encounters recorded on video. Such analysis fits into the
recently advanced video data analysis methodological frame, which involves the in-depth analysis
of video recorded in natural settings for the purpose of exploring situational dynamics and their
1
Makin et al. (2020) demonstrated the benefits BWCs can offer, specifically by enabling
interest. Researchers have recently leveraged such benefits by using BWC footage to study a range
2019), passenger aggression towards officers during citation events (Friis, Liebst, Philpot, &
encounters (Pollock, Scott, & Moore, 2020), and the duration of police use of physical force
(Willits & Makin, 2018). That said, as Rosenbaum points out in Lum (2015), “we lack details
about what specific behaviors lead to negative or positive outcomes, that is, who said what to
whom in what tone of voice with what nonverbal cues?” (p. 7).
Braga (2010, p. 177) speaks to the need for academics to provide practitioner stakeholders
with “strategic information products” and communicate their importance to police officials in a
range of venues for the purpose of informing policy and practice (Braga & Davis, 2014). Within
such an arrangement, strategic information products may differ from the experimental and quasi-
experimental evaluation research typically privileged in academia as outcomes represent only one
of many important pieces of information that police require (Greene, 2019). Sparrow (2011) noted
that exploratory and descriptive methods commonly employed in natural science inquiry may
better inform the design of police strategies than program evaluation efforts (also see Greene,
2019).
BWC footage provides researchers with details of the use of force event at the situational
level. We leverage these newfound research benefits by using BWCs to conduct a systematic social
observation (SSO) of police use of force events, and provide a detailed description of use of force
events. The study is an outgrowth of an applied partnership between the research team and the
2
Newark Police Division (NPD) , and the use of force event dimensions described in this
‘information product’ include the context of the event, the suspect and officer characteristics, and
STUDY SETTING
Newark is the largest municipality in the state of New Jersey, with a population of over
280,000 according to the most recently available figures from the U.S. Census Bureau. Blacks
comprise nearly 50% of the population with about 36% of residents identifying as Latino. In 2018,
Newark had a violent crime rate of 733 per 100,000 residents, which is larger than the national
average for cities with populations of 250,000 and over (451.5 per 100,000) according to the
In May 2011, the U.S. Department of Justice Civil Rights Division opened an investigation
into the NPD after receiving allegations of civils rights violations by the agency. Results of the
excessive use of physical force (U.S. Department of Justice Civil Rights Division, 2014). This
introduced in the 1994 Violent Crime Control and Law Enforcement Act that has become a popular
remedy for agency-driven violations of federally protected rights (Powell, Meitl, & Worrall, 2017).
The consent decree required the NPD to adopt reforms to address the violations, including the
deployment BWC cameras and revised training on use of force in an attempt to prevent
RESEARCH METHODS
3
The present study is a SSO video data analysis of police use of force cases captured on
BWC by the NPD. SSO is a systematic method of data collection developed by Reiss (1968, 1971)
wherein data collection is independent of the phenomena being observed. Video data analysis,
involving the analysis of pre-existing video footage, provides an innovative venue for SSO
Similar to Mastrofski et al.’s (2010) description of an event as a social construct, the unit
of analysis for this study is use of force events which includes at least one instance of police use
of physical force. Use of force events include a period of time preceding and following the use of
force incident(s), beginning when the officers are first visibly seen interacting with any involved
parties (e.g., suspects, bystanders, or victims). The exception to this is cases in which the video
footage begins after police had already begun interacting with involved parties. There are 36 such
cases in this study. The end of the use of force event can be described as the time at which full
suspect compliance is secured, making the likelihood of physical force minimal. This may include
the period following an arrest, the time at which suspects were secured within a patrol car, or the
The sampling frame for this study is all closed (i.e., resolved in the courts by plea, trial, or
dismissal) and inactive (not under active investigation by internal affairs units) use of force events,
not involving victims, and recorded by BWCs between December 2017 and December 2018
(N=122). Thirty events were excluded from the analysis for a variety of reasons. The most common
reason for exclusion (n=18) is that the use of force event was not actually captured by BWC. In
these cases, the BWC-equipped officer(s) arrived on-scene after force had been applied. In 5
additional cases use of force occurred after arrest processing—at either a police precinct or
hospital. Such settings fall outside of the public areas of interest for this study. Six cases were
4
excluded either because the use of force only constituted the application of handcuffs (n=3) or
because the internal affairs unit was actively investigating the cases (n=3), thus restricting access
to the video from NPD servers pending the conclusion of the investigation. Finally, a single
incident was incorrectly tagged as two events in the BWC database. The final sample consists of
91 cases.
Cases were coded for a variety of variables related to context of the event, including
presence of bystanders, suspect and officer characteristics, various actions of the parties involved,
and the characteristics of the use of force encounter. It took approximately 300 hours to review
and code all use of force events. Due to technological limitations preventing coders from viewing
the video footage remotely, a single coder was responsible for data collection and tests for intra-
coder reliability were conducted. To test reliability, ten percent of cases were randomly selected
and recoded six months after the original coding commenced. Kappa coefficients confirmed the
reliability of all coding for this study, with all coefficients >0.60 (see Landis & Koch, 1977).1
Several items captured here are time-variant. For each time-variant item a variable was
created measuring the number of minutes from the start of the event to the time at which the action
of interest occurred. For some time-variant items a variable was created measuring the duration of
the action in minutes. Additionally, while some time-variant indicators represent isolated actions
(e.g., an officer announced the presence of a BWC at the 1:05 minute mark of the footage), some
indicators fit within a series of the same or similar actions which can occur once, or multiple times
beginning at different time-points (e.g., an officer delivered multiple commands to a suspect during
the encounter). A series is conceptualized as the same or similar action occurring consecutively
for a period of time. The series ends when the action changes, or it shifts in severity or intent. For
5
example, if an officer calmly instructs a suspect to open their backpack using similar language for
1 minute, and does not change the instruction or tone during the entire minute, that would be
considered a 1-minute long series of calm commands. That series of calm commands would end
when the instruction changes (e.g., from “open your backpack” to “turn out your pockets”) or the
tone changes from calm to a shouting command. For cases with multiple series of actions we
measured the number of minutes from the end of one series to the start of the next; or the in absence
of duration (e.g., force actions that occur within seconds or fractions of seconds), we measured the
number of minutes from the start of one series to the start of the next.
To ensure reliability for those variables that are time-variant, start and end points were
assigned to the 5-second intervals closest to the actual start and end time points. For example, a
use of force event beginning between 3:00 and 3:04 was assigned a start time of 3:00 and a use of
force event beginning between 3:05 to 3:09 was assigned a start time of 3:05. Preliminary coding
demonstrated that assigning time points at the 1-second interval resulted in unreliable measures.
Analytical Framework
Frequency tables and cross-tabulation tables are used to present counts and percent of
categorical variables, and measures of central tendency are used to describe discreet and
RESULTS
Setting. Of the 91 use of force events included in the sample, most (61.5%) occurred during
the night hours (see Table 1). ‘Nighttime’ is conceptualized as the time which darkness has
provided a cover and flashlights or other illuminating devices are required by officers for the
6
purpose of identifying suspects, bystanders, potential weapons, and the scene around them. The
majority of events took place both outdoors (80.22%), and in public spaces (79.12%). Public
spaces are those areas that are open for any persons to occupy, such as sidewalks, public parks,
and common spaces in shopping areas. Additionally, most (65.93%) events occurred on public
streets, with some cases occurring in private areas of single family homes (7.69%) and in common
areas of apartment building residences (7.69%). Five use of force events (5.49%) occurred at gas
stations and 4 events (4.4%) took place in parking lots. These results are not surprising as people
tend to congregate in public spaces, creating more opportunity or need for a police presence and/or
Time of Day
Frequency Percent Cumulative Percent
Daytime 35 38.46 38.46
Nighttime 56 61.54 100
Total 91 100
Location
Frequency Percent Cumulative Percent
Indoor 18 19.78 19.78
Outdoor 73 80.22 100
Total 91 100
Area Use Type
Frequency Percent Cumulative Percent
Public 72 79.12 79.12
Private 19 20.88 100
Total 91 100
Classification of Area
Frequency Percent Cumulative Percent
Public street 60 65.93 65.93
Single family home, private area 7 7.69 73.63
Single family home, publicly visible area 2 2.2 75.82
Apartment building residence, private area 4 4.4 80.22
Apartment building residence, common area 7 7.69 87.91
Public park/commons 1 1.1 89.01
Corner store/small market 0 0 89.01
Liquor establishment 0 0 89.01
Parking lot 4 4.4 93.41
Transit location 0 0 93.41
Retail facility 0 0 93.41
Restaurant 1 1.1 94.51
Vacant space 0 0 94.51
Gas station 5 5.49 100
Total 91 100
7
Event. Time in minutes until event resolution is operationalized as the number of minutes
into the video when the officers are first visibly seen interacting with any involved parties to the
time the event is resolved. Event resolution refers to the time a natural break occurs in the event.
This may be different from the time at which the video ends, and may include the time following
an arrest, time at which officers are seen leaving the scene, or the time at which the video turns off
before any of the aforementioned is observed. There is quite a bit of variability in length of use of
force events. The average use of force event in this sample lasts just under 11 minutes, with a
median length of 6.58 minutes, and a range of 1 to 72.08 minutes (see Table 2). In 36 cases officers
are first visibly seen interacting with involved parties immediately when the video begins,
indicating that the event was already in progress when the camera was activated. When those cases
are removed (table not shown), the average use of force event in this sample lasts 12.25 minutes
The median length of time from the start of the force event to the time the suspect is
handcuffed is fairly short (2.58 minutes; mean=6.54), with a range of zero to 69.08 minutes. The
median length of time from the start of the force event to the time the suspect is put into a police
vehicle is 6.67 minutes (mean=10.64), and the median length of time until the suspect is put into
an ambulance is 33.5 minutes. Both the police vehicle and ambulance variables have fairly wide
Police use of force may be organized into a hierarchy of actions that increase in severity.
This conceptualization is commonly referred to as the use of force continuum, with the least severe
actions, such as officer presence and verbalization at the bottom of the hierarchy, and the most
severe actions, such as threat of lethal force or use of lethal force at the top of the hierarchy. Our
data demonstrate that it took a median of 2 minutes (mean=6.77) from the start of the force event
8
to the time at which the event’s highest level of use of force is used. These results indicate how
dynamic and rapid these events can be. In one case, the final level of force was already in progress
when officers are first visibly seen interacting with any involved parties on the scene; however, in
one case it took well over an hour (68.75 minutes) before the highest level of force of the event
was observed. Again, in 36 cases officers are first visibly seen interacting with involved parties
immediately when the video begins, indicating that the event was already in progress when the
camera was activated. When those cases are removed, the median time until the event’s highest
level of force is used is 2.41 minutes, with an average length of 8.14 minutes.
In more than half (56.04%) of use of force events under study the officer was directed to
the scene by a dispatcher (compared to 43.96% that were not) and 58.24% of police/citizen
contacts are in response to a citizen call for service, rather than a proactive police action (compared
to 41.76% proactive responses) (see Table 3). While not mutually exclusive, proactive actions and
non-dispatcher cases differ slightly in that officers sometimes respond to a situation as a result of
being flagged down on the street from a civilian seeking service, rather than through a dispatcher
call for service. In those instances, the case is not considered proactive, but it is also not a case of
an officer being directed to a scene by a dispatcher. Overall, results suggest that a large proportion
of police citizen interactions ending in force stem from contacts occurring apropos.
In one out of 91 cases an indication of gun violence was observed (e.g., gun shot victim,
gun shot shells, officers’ indication of associated gun violence), and 25.27% of use of force events
9
(n=23) involved domestic violence. Of those 23 domestic violence-related events, 17 or 73.91%
included a spouse on scene (see Table 4). Results around spousal presence may suggest that when
both involved parties of a domestic violence incident are present on scene, the situation is inflamed.
Table 4. Cross-tabulation use of force events involving domestic violence, by spousal involvement.
Domestic Violence
Domestic Violence with Spouse No Yes Total
No 68 6 74
100% 26.09% 81.32%
Yes 0 17 17
0% 73.91% 18.68%
Total 68 23 91
100% 100% 100%
People. While the number of suspects on scene was typically 1, there were as many as 5
suspects on scene throughout a use of force event (see Table 5). The median number of bystanders
on scene during a use of force event is 2, however there were often no bystanders and the maximum
number of bystanders observed is 24. Of the 64 use of force events in which bystanders were
present, the most common type of bystander on scene was involved, non-antagonistic/neutral or
helpful bystanders, which are present exclusively (no other types of bystanders present) in 39.06%
10
of use of force events (see Table 6). These are bystanders who insert themselves into the event but
are not behaving in an antagonistic manner. Involved, antagonistic bystanders were present
exclusively during 12.5% of use of force events, and uninvolved bystanders were exclusively
present in 10.94% of use of force events. In 10.94% of cases both uninvolved and involved,
antagonistic bystanders were present, in 10.94% of cases both involved, antagonistic bystanders,
and involved, non-antagonistic/neutral or helpful bystanders were present. Victim(s) were on scene
during 31.87% of use of force events. As with the presence of a spouse in domestic violence calls,
Table 6. Frequency and percent use of force events, by type of bystanders and presence of victim on scene.
All types of bystanders tended to arrive on scene very quickly, although this finding can
be expected given that most use of force incidents occurred in public spaces. Those cases in which
uninvolved bystanders were present, the median number of minutes from when the officers are
first visibly seen interacting with any involved parties to the time at which uninvolved bystanders
arrive on scene is .83 minutes (mean=2.96), with a range of zero to 26.67 minutes (see Table 7).
11
The mean number of uninvolved bystanders on scene at this time-point is 2 (mean=4.22), with a
range of 1 to 23. Cases in which there is more than one time of arrival for uninvolved bystanders
(n=7), the median number of minutes from the time at which the first uninvolved bystanders arrive
on scene until the next uninvolved bystanders arrive on scene is 2.84 minutes and the median
number of observable uninvolved bystanders increases from 2 to 4. In one case there were 3
different arrival times for uninvolved bystanders entering the scene, and the median number of
minutes from the time at which the second uninvolved bystanders arrived on scene until the next
Table 7. Descriptive statistics number of uninvolved bystanders and minutes until uninvolved bystanders arrive, by
time series.
For cases in which involved, antagonistic bystanders were present, the median number of
minutes from when the officers are first visibly seen interacting with any involved parties to the
time at which involved bystanders arrive on scene is .79 minutes (mean = 4.61), with a range of
zero to 62.75 minutes (see Table 8). While the median number of involved, antagonistic bystanders
at this time is 1.5, one case had 11 bystanders of this type on scene. Of the 26 cases in which an
antagonistic bystander was present during this time-frame, in 9 cases involved antagonistic
bystanders filming the use of force incident, with as many as 5 bystanders filming at a time. Cases
in which there was more than one time of arrival for involved, antagonistic bystanders (n=6), the
median number of minutes from the time at which the first arrived on scene until the next arrived
on scene is 1.21 minutes. Of the 6 cases in which there is a second arrival time for involved,
12
antagonistic bystanders, bystanders were filming the incident in two cases. Filming appears to be
Table 8. Descriptive statistics number of involved, antagonistic bystanders, minutes until involved, antagonistic
bystanders arrive, and number of involved, antagonistic bystanders filming the incident, by time series.
Those cases in which involved, non-antagonistic or helpful bystanders were present, the
median number of minutes from when the officers are first visibly seen interacting with any
involved parties to the time at which the bystanders arrive on scene is .5 minutes (mean = 2.27),
with a range of zero to 26.67 minutes (see Table 9). While the median number of involved, non-
antagonistic bystanders at this time is 1, one case had 16 bystanders of this type on scene. Of the
43 cases in which a non-antagonistic bystander was present during this time-frame, one case
involved a non-antagonistic bystander filming the use of force event. Cases in which there is more
than one time of arrival for involved, non-antagonistic bystanders (n=9), the median number of
minutes from the time at which the first arrived on scene until the next arrived on scene is 2.25
minutes. The median number of involved, non-antagonistic bystanders on scene during this second
time period is 2 and of the 9 cases of this nature, 1 includes an involved, non-antagonistic bystander
filming the use of force incident. There are two cases with more than two arrival times for involved,
non-antagonistic bystanders. See Table 9 for descriptive information regarding these two cases.
13
Table 9. Descriptive statistics number of involved, non-antagonistic/neutral or helpful bystanders, minutes until
involved, non-antagonistic/neutral or helpful bystanders arrive, and number of involved, non-antagonistic/neutral
or helpful bystanders filming the incident, by time series.
Suspect attributes. As mentioned, use of force cases most often involved only one suspect
on scene (see Table 5 above). In the majority of cases (92.31%), the first suspect that the police
interacted with was the person who had force used upon them (see Table 10). Most suspects
(71.43%) are men and most (70.33%) are Black. Suspect median age is 28 years. Results around
age and gender are consistent with known offending populations and the ethnicity results are
consistent with the demographic make-up of the city of Newark. In most cases (96.7%) the suspect
who had force used upon them spoke fluent English and of the two cases in which the suspect did
not speak fluent English, in at least one case an officer on scene spoke to the suspect in their native
tongue. In 9.89% of cases the suspect spoke with an accent. In 25.27% of the use of force cases
under study there is evidence that the suspect who had force used upon them was under the
influence of drugs or alcohol. Such evidence includes the suspect displaying slurred speech,
difficulty standing or walking, and verbal indication of intoxication from bystanders. These results
suggest both that substance use may result in suspects becoming agitated and difficult to manage,
and that police officers may be ill-equipped to deescalate situations involving intoxicated subjects.
14
In 3 cases there is evidence that the suspect suffered from mental health issues. Such evidence
indications from bystanders. The small number of force recipients with mental health issues is
encouraging as it indicates that officers have either been successful deescalating situations with
such subjects, or they rarely come into contact with subjects with mental health issues.
15
Table 10. Frequency and percent use of force events, by suspect attributes.2
Force Recipient
Frequency Percent Cumulative Percent
Primary suspect: Suspect that was first contacted by officers 84 92.31 92.31
Secondary suspect: Suspect not first contacted by officer but on scene from event start 3 3.3 95.6
Back-up suspect: Suspect that arrived on scene after the event already began (not bystander) 1 1.1 96.7
Don't know 3 3.3 100
Total 91 100
Suspect Gender
Frequency Percent Cumulative Percent
Woman 15 16.48 16.48
Man 65 71.43 87.91
Don’t know 11 12.09 100
Total 91 100
Suspect Ethnicity
Frequency Percent Cumulative Percent
White 2 2.20 2.20
Black 64 70.33 72.53
Hispanic 14 15.38 87.91
Don’t know 11 12.09 100
Total 91 100
Suspect Age
N Median Mean Std. Dev. Min Max
79 28 29.39 10.21 15 61
Suspect Fluent in English
Frequency Percent Cumulative Percent
No 2 2.2 2.2
Yes 88 96.7 98.9
Don't know 1 1.1 100
Total 91 100
Officer Speaks Suspect's Language (non-English)
Frequency Percent Cumulative Percent
No 0 0 0
Yes 1 50 50
Don't know 1 50 100
Total 2 100
Suspect Speaks with Accent
Frequency Percent Cumulative Percent
No 81 89.01 89.01
Yes 9 9.89 98.9
Don't know 1 1.1 100
Total 91 100
Evidence Suspect Under Influence of Drugs or Alcohol
Frequency Percent Cumulative Percent
No 67 73.63 73.63
Yes 23 25.27 98.9
Don't know 1 1.1 100
Total 91 100
Evidence Suspect Suffers from Mental Health Issues
Frequency Percent Cumulative Percent
No 87 95.6 95.6
Yes 3 3.3 98.9
Don't know 1 1.1 100
Total 91 100
2 Suspect gender, ethnicity and age data came from NPD administrative data.
16
Suspect weapon possession. In 18.68% of the use of force cases under study the suspect
was found to be in possession of a weapon (n=17) and in 47.06% of weapon possession cases that
weapon is a firearm (n=8) (see Table 11). In 35.29% of weapon possession cases the weapon is a
knife (n=6) and in 17.65% of weapon possession cases the weapon is a blunt object (n=3). These
results may be an indication that officers are implicitly or explicitly more prepared to or willing to
use force when they are aware of suspect weapon possession, regardless of threat level. In the
majority of weapon possession cases (70.59%), the weapon was found after force was used. This
does not mean the officer(s) were not aware of the presence of a weapon prior to using force. In
15 weapon possession cases the median number of minutes from the start of the event to the time
at which the weapon is located is 1.42 minutes, with a maximum time of 11.58 minutes (see Table
12). In one case a second weapon was discovered and that weapon was discovered 30 seconds
Table 11. Frequency and percent use of force events, by suspect weapon possession.
Table 12. Descriptive statistics minutes until suspect weapon discovered, by time series.
17
Suspect actions. In 28.57% of use of force cases the force recipient came into conflict with
any persons at the scene besides the police officer(s) (see Table 13). Examples of such conflict
include a physical altercation (punching, shoving, pinning to the wall or ground), or a verbal
altercation (shouting, heated argument). Of those 26 cases in which the suspect came into conflict
with someone on scene, in 84.62% of cases the suspect came into conflict with the initial crime
victim/complainant (n=22). Results around suspect conflicts indicate that many force recipients
may have posed an immediate danger to those around them prior to use of the force, as opposed to
In 32 of 91 cases (35.16%), the suspect attempted to flee the scene and in 46.88% of those
cases (n=16), the flee attempt was by way of running. In 40.63% (n=13) of flee attempt cases the
flee attempt was by way of walking, and in 4 instances the flee attempt was by way of driving.
Suspects who attempt to flee do so very quickly into the interaction. In those cases in which the
suspect attempted to flee, the median number of minutes from the start of the force event to the
time at which the suspect attempted to flee is .29, with a minimum value of 0 minutes, and a
maximum of 63 minutes (see Table 14). In two cases the suspect attempted to flee more than once,
and the median number of minutes from the first flee attempt to the second is 1.42
18
Table 13. Frequency and percent use of force events, by suspect conflict and flee attempt.
Table 14. Descriptive statistics minutes until suspect attempted to flee, by time series.
The force recipient displayed verbally antagonistic behavior toward anyone on scene in
52.75% of the use of force cases under study (see Table 15). These results suggest that such
behavior is so common that officers should expect to encounter it frequently and be prepared to
navigate it.
In those 48 cases in which the suspect displayed verbally antagonistic behavior, the median
number of minutes from the start of the use of force event to the time at which the suspect displayed
verbally antagonistic behavior is 1.88, with a minimum value of .08 minutes, and a maximum of
35.75 minutes (see Table 16). Such a display of verbally antagonistic behavior lasted a median of
1.33 minutes, but in one instance such a display of behavior has a duration of 45.75 minutes.
In 24 cases the suspect displayed verbally antagonistic behavior twice, and the median number of
minutes from the first display of such behavior to the second display is 1.8. The median length of
19
time such behavior went on for is 4.83 minutes, with a maximum duration of 49.17 minutes. In 11
cases the suspect displayed verbally antagonistic behavior on three occasions and the median
number of minutes from the second display of such behavior to the third display is 1.83. The
median duration of such behavior for this third time period is 10.5 minutes, with a maximum
duration of 52.09 minutes. There are some cases in which the suspect continued to display verbally
antagonistic behavior on more than three occasions and generally there is a positive correlation
between the number of times a suspect displays such behavior and the length of time the behavior
goes on for. Verbally antagonistic behavior is most often directed toward police officers, and in
Table 15. Frequency and percent use of force events, by suspect displaying verbally antagonistic behavior.
Table 16. Descriptive statistics minutes until suspect displayed verbally antagonistic behavior and duration of
behavior, by time series.
20
Table 17. Frequency and percent suspect verbal antagonistic behavior target, by time series.
The force recipient displayed physically antagonistic behavior toward someone on scene
in 39.56% of the use of force cases under study (see Table 18). In those 36 cases in which the
21
suspect displayed physically antagonistic behavior, the median number of minutes from when the
officers are first visibly seen interacting with any involved parties to the time at which the suspect
displayed physically antagonistic behavior is 2.8, with a minimum value of .0, and a maximum of
62.67 minutes (see Table 19). In 86.11% of those cases, the behavior was directed toward police
(see Table 20). In 8 cases the suspect displayed physically antagonistic behavior on two occasions,
and the median number of minutes from the first display of such behavior to the second display is
1.13. All of these cases directed such behavior toward police. Several cases displayed physically
antagonistic behavior on more than 2 occasions. See Tables 19-20 for more detailed information.
As with suspect verbal antagonism, these results suggest that physical antagonism is common
Table 18. Frequency and percent use of force events, by suspect displaying physically antagonistic behavior.
Table 19. Descriptive statistics minutes until suspect displayed physically antagonistic behavior, by time series.
22
Table 20. Frequency and percent suspect physical antagonistic behavior target, by time series.
Officer Characteristics
Officer demographics. Nearly all officers (86.81%) are men, which is consistent with the
profession of policing in general, 47.25% are Hispanic, and there is an equal number of White and
Black officers (n=18), indicating that Blacks are underrepresented in the NPD compared to
23
Table 21. Frequency and percent use of force events, by officer demographics.3
Officer Gender
Frequency Percent Cumulative Percent
Woman 1 1.10 1.10
Man 79 86.81 87.91
Don’t know 11 12.09 100
Total 91 100
Officer Ethnicity
Frequency Percent Cumulative Percent
White 18 19.78 19.78
Black 18 19.78 39.56
Hispanic 43 47.25 86.81
Don’t know 12 13.19 100
Total 91 100
Officer Age
N Median Mean Std. Dev. Min Max
79 29 31.95 7.26 21 54
Number of officers. The median number of officers on scene at the outset is 2, with a
maximum value of 9. In 67 of 91 cases more officers arrived on scene as time progressed (see
Table 22). In those 67 cases, the median number of minutes from the time the initial officers arrived
on scene to the next officer(s) arrival is 2 minutes, with a maximum time of 32.42 minutes. The
median number of officers on scene at this second time point is 4. In 31 cases there is a third time
of officer arrival, with the median time at which this occurs being 1.33 minutes after the arrival of
the previous group. The median number of officers on scene by this third point of arrival is 5, with
a maximum value of 9. In 13 cases there is a 4 th time of arrival for officers, and the median time
at which this occurs is 1.25 minutes after the previous group. The median number of officers on
scene at time period 4 is 6, with a maximum of 9 officers on scene. There is one case in which
officers continued to arrive on scene across 6 different time points, with a total of 13 officers being
on scene at last tally. These results demonstrate that back-up officers tend to arrive on scene very
3 Officer gender, ethnicity and age data came from NPD administrative data.
24
Table 22. Descriptive statistics number of officers on scene and minutes until officers arrive, by time series.
Officer interactions with BWCs. In 38 cases (41.76%) an officer announced the presence
of a BWC (see Table 23). The median length of time from the point of the video at which officers
are first visibly seen interacting with any involved parties on the scene to the time at which an
officer announces the presence of a BWC is 1.69 minutes (mean=5.2), with a maximum time of
53.25 minutes, and the median number of officers seen wearing a BWC during a given event is 4
officers (including the officer whose footage was primarily used for data collection) (see Table
24). While results indicate officers do not inform citizens of BWCs in most cases, officers
informing citizens did so quickly. In 34.07% of use of force cases the at least one BWC fell off at
Table 23. Frequency and percent use of force events, by officer announcing presence of BWC and BWC falling off.
25
Table 24. Descriptive statistics minutes until officer announces presence of BWC and number of officers wearing
BWC.
Procedural justice. In 53.86% of use of force cases an officer provided a reason to the
suspect as to why the officer was on the scene (in 18.68% of cases it is unknown as to whether or
not a reason was given), and in 27.47% of cases no reason was given (see Table 25). In 61.54% of
use of force cases an officer explained to the suspect why they were being detained. In 23.08% of
cases no such explanation was provided. Results suggest there are a large number of cases in which
information sharing with suspects was not observed. In 65.93% of use of force cases the suspect
attempted to speak for the purposes of expressing their views at least once during the event, and
median number of minutes from when the officers are first visibly seen interacting with any
involved parties to the time at which the suspect attempted to speak for the first time is .96 minutes,
with a minimum value of .0, and a maximum of 62.67 minutes (see Table 26). Suspects first
attempt to speak for a median duration of 2.79 minutes, and a maximum of 44.01 minutes; and in
76.67% of cases an officer allowed the suspect to express their views (see Table 27). In 75% of
use of force cases in which a suspect attempted to speak for the purpose of expressing their views,
a police officer addressed the suspect’s concerns by making attempts to answer any questions
Of the 60 cases in which a suspect attempted to speak for the purpose of expressing views,
in 24 cases the suspect made a second attempt to speak. The median number of minutes from the
time the suspect stopped attempting to speak in the first instance to the time they began again is
1.25 minutes, with a maximum value of 38 minutes—meaning one suspect did not try a second
time to express their views verbally for 38 minutes from the time they stopped expressing their
26
views the first time (see Table 26). During this second attempt at expressing their view, the median
number of minutes suspects attempted to speak is 5.29, with one suspect attempting to express
their views for 58.83 minutes. Of these 24 cases, 17 were permitted to speak and all 17 had their
Of the 24 cases in which suspects attempted to speak on two occasions for the purpose of
expressing their views, 12 attempted a 3rd time to speak. The median number of minutes from the
time the suspect stopped attempted to speak in the second instance to the time they began again is
1.25 minutes and this group of suspects attempted to express their views for a median of 7.55
minutes (see Table 26). One suspect attempted to speak for 59.08 minutes straight. Of those 12
who attempted to speak on a third occasion, all were permitted and 10 had their concerns addressed
by police (see Table 27). Several suspects continued to attempt to speak for the purpose of
expressing their views (see Tables 26-27 for details). Results suggest that with additional attempts
to express views, suspects become more persistent or committed to sharing their perspective and
police may be less open to allowing suspects to express their views. That said most officers
permitted suspects to speak at the suspect’s first attempt, and those officers who permit suspects
to speak are consistent in their adherence to procedural justice principles in that they address
suspects’ concerns.
27
Table 25. Frequency and percent use of force events, by procedural justice indicators.
Table 26. Descriptive statistics minutes until suspect permitted to speak for the purpose of expressing views and
duration of suspect speaking, by time series.
28
Table 27. Frequency and percent suspects attempting to speak for the purpose of expressing views, by officer
allowing suspect to speak, officer addressing concerns, and by time series.
29
In 38.46% (n=35) of use of force cases at least one officer displayed verbally antagonistic
behavior (see Table 28). The median number of minutes from when the officers are first visibly
seen interacting with any involved parties to the time at which an officer displayed verbally
antagonistic behavior is 1.42 minutes, with a minimum value of .08, and a maximum of 17.25
minutes (see Table 29). The median number of minutes such behavior occurred is .66, representing
more than a few words expressed by officers. In one case an officer displayed verbally antagonistic
behavior for 43.91 minutes. The force recipient was almost exclusively the target of such behavior
(see Table 30) during this time. In 10 cases there was a second instance of verbally antagonistic
behavior displayed by an officer, which occurred a median of 1.04 minutes after the first display
of verbally antagonistic behavior and lasted for a median of 1.75 minutes—although in one case
the behavior occurred for a duration of 47.34 minutes (see Table 29). The target of such behavior
was almost exclusively the force recipient (see Table 30). In 3 cases there was a third instance of
Table 28. Frequency and percent use of force events, by officer displaying verbally antagonistic behavior.
Table 29. Descriptive statistics minutes until officer displayed verbally antagonistic behavior and duration of
behavior, by time series.
30
Table 30. Frequency and percent suspect verbal antagonistic behavior target, by time series.
Who? What? Where? In 40.66% of use of force cases under study (n=37), the primary
officer used force the first time (in cases with multiple instances) or only time it was applied (see
Table 31). Primary officer is conceptualized as the officer who made initial contact with the
suspect. In 16.48% of cases both the primary and secondary (officer that did not make initial
contact with suspect, but was on scene for the entirety of the event) officers applied force the first
time (in cases with multiple instances) or only time it was applied. In a notable portion of cases
(13.19%) force was first applied a back-up officer. The back-up officer is the officer who arrived
on scene after the event had already begun. In some case (8.79%) both the primary and secondary
officers applied force in the first (or only) instance of force and in a smaller proportion of cases
31
(6.59%) it was the secondary officer who applied force. In 10 cases (10.99%) it is unknown which
In 50.55% of cases the highest level of force used during the force event is hard, empty
hand control, and in 36.26% of cases the highest level is soft, empty hand control (see Table 31).
In 7.69% of use of force cases under study the highest level of force was threat of lethal force and
in 5.49% of cases it was use of a chemical device (i.e., pepper spray). There are zero instances of
In most cases (92.31%) the video footage of the force event(s) was captured in close
proximity to the actual event, but in some cases (7.69% or n=7) the video footage depicts the force
event to be in the distance (see Table 31). ‘Distance’ is conceptualized as being out of direct view
of the BWC or in the periphery. For example, an event occurred across a street from the filming
location. The event was observed, but some contextual factors were lost due to the distal nature of
the event.
Table 31. Frequency and percent use of force events, by officer who used force in the first instance, final level of
force and whether or not the force was captured on camera in the distance.
32
Calm command. In 78.02% of use of force cases under study an officer gave the force
recipient a calm command (see Table 32). A calm command is a non-threatening, verbal command.
Examples of a calm, non-threatening verbal command include, “Let me see your identification”
and “Open your backpack”. That said the focus is not on the words used, but on the tone with
which the command is delivered. The median number of minutes from when the officers are first
visibly seen interacting with any involved parties to the time at which the first calm command is
delivered by an officer to the suspect is 1.16 minutes, with a minimum value of 0 (meaning a calm
command was in-progress when the video began), and a maximum of 67.25 minutes (see Table
33). This first series of calm commands lasted a median of .83 minutes, with one case having a
maximum calm command duration of 46.83 minutes. In other words, in one case an officer
continued to repeat the same set of instructions, in the same tone for over ¾ of an hour. The
minimum duration of value of zero seen in Table 32 indicates there is at least one case in which
the calm command series last for less than 1/100 if 1 minute or around half a second (an example
might be a command such as, “be quiet”). The most common officer to deliver a calm command
in the first series is the primary officer, but there is some variability in officer; with various officer
types and combinations of multiple officers each delivering calm commands in 5-7 cases (see
Table 34). Suspects complied with the calm command in 23.94% of cases (see Table 35). In other
In many cases there were several series of calm commands. In 37 cases there were two
series of commands and the median number of minutes from the end of the first series to the
beginning of the second is 1.75, with a maximum value of 18.83 (see Table 33). The median
duration of this second series of commands is 3.25 minutes and the primary officer is the most
33
likely person to give such commands at this time (see Table 33). In 35.14% of cases with a second
series of calm commands the suspect complied with the command (see Table 35).
In 10 cases there was a third series of calm commands given. The median number of
minutes from the end of the second series to the beginning of the third series is 1.58, with a
maximum value of 4.42 (see Table 33). The median duration of this third series of commands is
7.25 minutes and the primary officer is the most likely person to give such commands at this time
(see Tables 33-34). In 40% of cases with a third series of calm commands the suspect complied
with the command (see Table 35). There are 5 cases with more than 3 series of commands (see
Tables 33-35 for details). Results around calm commands suggest that many officers are quite
committed to the directive itself, as well as they are committed to remaining calm in their delivery
of a directive.
Table 32. Frequency and percent use of force events, by officer giving calm command.
Table 33. Descriptive statistics minutes until officer calm command and duration of command, by time series.
34
Table 34. Frequency and percent officer who gave calm command, by time series.
35
Table 35. Frequency and percent suspect compliance with calm command, by time series.
Shout command. In 53.85% of use of force cases under study an officer gave the force
recipient a shout (i.e. increased volume) command (see Table 36). Again, the focus is not on the
words used, but on the tone with which the command is delivered. The median number of minutes
from when the officers are first visibly seen interacting with any involved parties to the time at
which the first shout command is delivered by an officer to the suspect is 1.59 minutes, with a
maximum of 52.08 minutes (see Table 37). This first series of shout commands lasted a median of
.79 minutes, with one case having a maximum shout command duration of 43.92 minutes. The
most common officer to deliver a shout command in the first series is the primary officer (50%),
but the back-up officer is also a common source (16.67%) (see Table 38). Suspects complied with
36
In 10 cases a second series of shout commands took place and the median number if
minutes from the end of the first series to the beginning of the second is 2.09, with a maximum
value of 5.67 (see Table 37). The median duration of this second series of commands is 2.88
minutes and the primary officer is the most likely person to give such commands at this time (see
Table 38). Again, results indicate many officers are quite committed to the directive itself, as well
as they are committed to the verbal delivery of the directive. In 40% of cases with a second series
of shout commands the suspect complied with the command (see Table 38). There are 2 cases with
Table 36. Frequency and percent use of force event, by officer giving shout command.
Table 37. Descriptive statistics minutes until officer shout command and duration of command, by time series.
37
Table 38. Frequency and percent officer who gave shout command and suspect compliance with shout command,
by time series.
Soft-hand control. In the majority of cases (85.71%) soft, empty hand control was used
(see Table 39). Examples of soft, empty hand control include grabs, and holds. There are 13 cases
(14.29%) in which soft, empty hand control was not used. When cross-tabulations were run it can
be seen that of all the cases in which soft, empty hand control was not used, hard empty hand
control was used (see Table 40). Examples of hard, empty hand control include punches and kicks.
38
Additionally, of the 13 cases in which soft, empty hand control was not used, threat of lethal force
was used in 6 cases. This result may be an indication that those 6 cases in which the force
continuum was not followed, either the threat to the public or officers was immediately high.
Table 39. Frequency and percent use of force events, by officer using soft, empty hand control.
Table 40. Cross-tabulations use of force events with soft, empty hand control, by hard, empty hand control and by
threat of lethal force.
The median number of minutes from when the officers are first visibly seen interacting
with any involved parties to the time at which soft, empty hand control used on a suspect by an
officer is 1.63 minutes, with a minimum of zero (meaning use of force was in progress when the
video footage began) and maximum of 67.25 minutes (see Table 41). These results demonstrate
that it takes longer, on average, for soft force to be applied than it does for commands to be
delivered. The most common officer to use soft, empty hand force in the first instance of such
force is the primary officer (43.59%), with the combination of the primary and secondary officers
being fairly common as well (16.67%) (see Table 42). In 82.05% of soft, empty hand force cases
39
the suspect showed resistance (see Table 42). Resistance can be conceptualized as repeated failure
In 26 cases soft, empty hand control was used a second time and the median number of
minutes from the time of the start of the first use of soft, empty hand control to the second is 1
minute, with a minimum of .16 and a maximum of 5.58 (see Table 41). These results indicate that
once soft, empty hand controlled is used once, it is used again quite soon after. The most common
officer to use soft, empty hand force in the second instance of such force is the primary officer
(42.31%), with the combination of the primary and secondary officers being fairly common
(19.23%), and the combination of primary and back-up officers also occurring (15.38% or n=5)
(see Table 42). In the majority cases in which soft, empty hand force was used a second time (24
of 26), the suspect showed resistance. In 5 cases a third instance of soft, empty hand control was
observed and in all 5 instance the suspect displayed resistance (see Tables 41-42).
Table 41. Descriptive statistics minutes until officer soft, empty hand control, by time series.
40
Table 42. Frequency and percent officer who used soft, empty hand control and suspect resistance of soft, empty
hand control, by time series.
Hard empty hand control. Hard, empty hand control was used in 63.74% of use of force
cases under study (see Table 43). The median number of minutes from when the officers are first
visibly seen interacting with any involved parties to the time at which hard, empty hand control
41
used on a suspect by an officer is 2 minutes, with a minimum of zero and maximum of 67.83
minutes (see Table 44). Again, these results suggest that is takes, on average, longer for officers
to apply hard hand control, then it does to deliver commands or apply soft hand control. The most
common officer to use hard, empty hand force in the first instance of such force is the primary
officer (39.66%), with the combination of the primary and secondary officers being fairly common
as well (24.14%) (see Table 45). In 75.86% of hard, empty hand force cases the suspect displayed
resistance toward being detained. In 4 cases hard, empty hand control was used a second time and
Table 43. Frequency and percent use of force events, by officer using hard, empty hand control.
Table 44. Descriptive statistics minutes until officer hard, empty hand control, by time series.
42
Table 45. Frequency and percent officer who used hard, empty, hard hand control and suspect resistance of hard,
empty hand control, by time series.
Blunt impact control. There was one instance of blunt impact control, which was the use
of a flashlight to immobilize (see Table 46). In this one case the force occurred after 18.08 minutes
from the time at which the officers are first visibly seen interacting with any involved parties (see
Table 47). The blunt object was used by the primary officer and the suspect resisted this attempt
43
Table 46. Frequency and percent use of force events, by officer using blunt impact control.
Table 47. Descriptive statistics minutes until officer blunt impact control, by time series.
Table 48. Frequency and percent officer who used blunt impact control and suspect resistance of blunt impact
control, by time series.
Chemical control. Chemical control (i.e., pepper spray) was used in 5 cases (5.49% of
cases) and the median number of minutes from the time at which the officers are first visibly seen
interacting with any involved parties until chemical control was used is 2.16 minutes, with a
maximum time of 17 minutes (see Tables 49-50). This is longer, on average, then it takes to apply
other types of control or deliver commands. The type of officer who used chemical control differs
for each case and in all 5 cases the suspect resisted attempts to detain them (see Table 51). In one
instance chemical control was used twice and it was deployed .25 minutes after the first use of
chemical control (see Table 50). The suspect did not resist the second time (see Table 51).
44
Table 49. Frequency and percent use of force events, by officer using chemical control.
Chemical Control
Frequency Percent Cumulative Percent
No 86 94.51 94.51
Yes 5 5.49 100
Total 91 100
Table 50. Descriptive statistics minutes until officer chemical control, by time series.
Table 51. Frequency and percent officer who used chemical control and suspect resistance of chemical control, by
time series.
Threat of Lethal Force. The threat of lethal force was used in 8 use of force cases (8.79%)
and the median number of minutes from the time at which the officers are first visibly seen
interacting with any involved parties until threat of lethal force is .21 minutes, with a minimum of
zero and a maximum of 1.42 minutes (see Tables 52-53). As mentioned, in most of these instances
soft hand control was not also applied. Generally, suspects in these cases posed an immediate
45
threat to the public or the police, although there were rare instances of officers neglecting to move
through the use of force continuum when they could have done so. When those cases with an event
that was in-progress at the time the video footage began are removed, 5 cases with threat of lethal
force remain. In those 5 cases the median number of minutes until the threat is made decreases to
.09. In most cases (5/8 or 62.5%) the threat of lethal force was made by the primary officer and in
2 cases the suspect resisted detention (see Table 54). In no cases was there a second threat of lethal
force.
Table 52. Frequency and percent use of force events, by officer using threat of lethal force.
Table 53. Descriptive statistics minutes until officer threat of lethal force, by time series.
Table 54. Frequency and percent officer who used hard, blunt impact control and suspect resistance of blunt
impact control, by time series.
This report builds upon the work of prior research on police use of force by capitalizing on
46
on video by NPD. Such reporting addresses the need to produce “strategic information products”
(see Braga, 2010; Sparrow, 2011; and Greene, 2019) around which specific party characteristics
and actions can be observed during use of force events to help inform police strategies. The use of
force event dimensions described in this “information product’” include the context of the event,
the suspect and officer characteristics, and the characteristics of the use of force encounter.
These data indicate that police use of force cases in Newark, NJ typically occur outdoors
and in public spaces. There is a great deal of variability in the length of use of force event from
when the officers are first visibly seen interacting with any involved parties to the time the event
is resolved. With that said the situation can escalate fairly quickly and even when event in-progress
cases are removed (those cases in which the camera was activated after the suspect and officers
began interacting), the median time until the event’s highest level of force is used is 2.41 minutes.
Victim(s) were on scene during around a third of use of force events and it is common for
suspects to have an altercation with the crime victim or complainant. It is also not unusual for the
suspect to attempt to flee the scene. A flee attempt can happen very quickly (within 30 seconds of
the interaction). In a quarter of cases there is evidence that the suspect who had force used upon
them was under the influence of drugs or alcohol. These results suggest both that substance use
may result in suspects becoming agitated and difficult to manage, and that police officers may be
ill-equipped to deescalate situations involving intoxicated subjects. In less than 20% of cases the
suspect was found to be in possession of a weapon. Of the weapons cases, around half involved a
firearm, indicating that officers may be implicitly or explicitly more prepared to or willing to use
force when they are aware of suspect weapon possession, regardless of threat level. Further, a
substantial proportion of use of force cases involve domestic violence. Each of these findings speak
to the potential for volatility in many of use of force cases, and results around spousal presence
47
may suggest the situation is inflamed when both involved parties of a domestic violence incident
are present on scene,. While the background characteristics of the force recipients are unknown in
the present study, the chronic domestic abuser population does overlap with the general criminal
population and can be a threat to the community and to their intimate partner victims (Scott,
While proactive policing strategies can produce short term crime-control gains, there tend
to be racial and income disparities in the application of such strategies (Weisburd et al., 2019;
Svensson & Saharso, 20214). The present study suggests that a substantial portion (41.76%) of
use of force cases occurred during a proactive policing response and overall results suggest that a
large proportion of police citizen interactions ending in force stem from contacts occurring
apropos.
Bystanders were present in most cases, with the most common type of bystander on scene
antagonistic bystanders were present, such bystanders arrived within less than one minute, and
many were filming the event. In 6 cases a second group of antagonistic bystanders arrived.
Certainly the addition of antagonistic bystanders may exacerbate an already stressful situation for
both suspects and police, and the potential for increased visibility of police practices due to the
proliferation of cellular telephones with filming capabilities can contribute to further the police
culture of suspicion of the public (see Sierra‐Arévalo, 2019). Even non-antagonistic bystanders
may exacerbate the police-citizen interaction; however, filming appears to be an activity that is
There were typically 2 officers on scene, but in a substantial number of cases more officers
arrived on scene at multiple arrival times within a few minutes of one another. Many officers did
48
not announce the presence of a BWC; but for those who did, the announcement came quickly. That
said, it is possible that compliance with NPD policy on announcing the presence of a camera has
improved since 2018, the final year of our study period. Finally, it is so common for BWCs to fall
off officers during interactions (34.07%) that addressing this hardware limitation should be a top
In many cases no reason was given to the suspect as to why the officer was on the scene
and no explanation was provided to the suspect as to why they were being detained. It was also
quite common for at least one officer to display verbally antagonistic behavior, and to do so within
less than 2 minutes of interacting with the suspect. These displays of such behavior were typically
short in duration and the force recipient is almost exclusively the target of such behavior. These
abrasive practices could compound the legitimacy-eroding effects of the use of force incident—
both at the individual and community level—to foster legal cynicism (Kirk & Papachristos, 2011).
Such community-level legal cynicism is associated with low arrest rates and low engagement in
collective efficacy (Kirk & Matsuda, 2011). With additional attempts to express views, suspects
become more persistent or committed to sharing their perspective and police may be less open to
allowing suspects to express their views. That said most officers permitted suspects to speak at the
suspect’s first attempt, and those officers who permit suspects to speak are consistent in their
In most (78.02%) cases an officer gave the force recipient a calm command and in just over
half of cases an officer gave a shout command. Results around calm commands suggest that many
officers are quite committed to the directive itself, as well as they are committed to the verbal
delivery of the directive. In half of the cases under study the highest level of force used during the
force event is hard, empty hand control, and in another 36.26% the highest level is soft, empty
49
hand control. The most common officer to use soft, empty hand force in the first instance of such
force was often someone other than the primary officer—this is true for hard had control as well.
Generally, suspects showed resistance to this type of force, as well as to hard, empty hand control.
In many cases in which soft, empty hand control was not used, threat of lethal force was used.
While the median number of minutes from when the officers are first visibly seen interacting with
any involved parties to the time at which force is used generally increases with the severity of the
force type, threat of lethal force is an exception to this trend. When the force continuum was not
followed, it may be that the threat to the public or officers was immediately high. Generally, most
officers adhered to the NPD policy of using escalating verbal commands before resorting to
physical force.
The most notable limitation of the present study is that, as Terrill et al. (2016, p. 67) point
out, BWCs “rely on officer activation, which does not occur equally for all citizen encounters and
fails to capture events preceding activation, which is a problematic gap.” To control for this, we
have presented measures of central tendency for both the full sample, as well as a sub-sample with
those cases which were in-progress at the time of camera activation removed. Results differ very
little from the full to sub-samples. Despite this limitation, the descriptive findings presented here
50
REFERENCES
Alpert, G.P., Dunham, R.G., & MacDonald, J.M. (2004). Interactive police-citizen encounters
that result in force. Police Quarterly, 7(4), 475-488.
Ariel, B., Farrar, W.A., & Sutherland, A. (2015). The effect of police body-worn cameras on use
of force and citizens’ complaints against the police: A randomized controlled trial.
Journal of Quantitative Criminology, 31, 509-535.
Bolger, P. (2015). Just Following Orders: A Meta-Analysis of the Correlates of American Police
Officer Use of Force Decisions. American Journal of Criminal Justice, 40(3), 466–492.
Braga, A. (2010). Setting a higher standard for the evaluation of problem-oriented policing
initiatives. Criminology & Public Policy, 9(1), 173-182.
Braga, A. A., & Davis, E. F. (2014). Implementing science in police agencies: The embedded
research model. Policing, A Journal of Policy and Practice, 8(4), 294–306.
Department of Justice Civil Rights Division. (2014). Investigation of the Newark Police
Department. United States Attorney’s Office, District of New Jersey.
Friis, C. B., Liebst, L. S., Philpot, R., & Lindegaard, M. R. (2020). Ticket inspectors in action:
Body-worn camera analysis of aggressive and nonaggressive passenger encounters.
Psychology of Violence, DOI: 10.1037/vio0000276.
Garner, J.H., Maxwell, C.D., Heraux, C.G. (2002) Characteristics associated with the prevalence
and severity of force used by the police. Justice Quarterly, 194, 705–746
Greene, J. (2019). Which evidence? What knowledge? Broadening information about the polcie
and their interventions. In Weisburd, D. & Braga, A. (eds.) Police Innovation: Contrasting
Perspectives. pp. 457-484. Cambridge Univeristy Press: Cambridge, UK.
Hough, M. and Roberts, J.V. (2004), Confidence in Justice: An International Review, King’s
College, London.
Kirk, D.S. & Matsuda, M. (2011). Legal cynicism, collective efficacy, and the ecology of arrest.
Criminology, 49(2), 433-472.
Kirk, D.S. & Papachristos, A.V. (2011). Cultural mechanisms and the persistence of
neighborhood violence. American Journal of Sociology, 116, 1190–233.
Landis, J.R. & Koch, G.G. (1977). The measurement of observer agreement for categorical
data. Biometrics, 33, 159-174.
51
environment? Translational Criminology Magazine. Fairfax, VA: Center for Evidence-
Based Crime Policy, George Mason University.
Makin, D. A., Willits, D. W., & Brooks, R. (2020). Systematic social event modeling : a
methodology for analyzing body-worn camera footage body-worn camera footage.
International Journal of Social Research Methodology, DOI:
10.1080/13645579.2020.1766775
Makin, D. A., Willits, D. W., Koslicki, W., Brooks, R., Dietrich, B. J., & Bailey, R. L. (2019).
Contextual determinants of observed negative emotional states in police–community
interactions. Criminal Justice and Behavior, 46(2), 301–318.
Mastrofski, S.D., Parks, R.B., & McCluskey, J.D. (2010). Systematic social observation in
criminology. In Handbook of quantitative criminology, p. 225-247. New York, NY:
Springer.
Mazerolle, L., Bennett, S., Davis, J., Sargeant, E., & Manning, M. (2013). Legitimacy in
policing: A systematic review. Campbell systematic reviews. Oslo: The Campbell
Collaboration.
Murphy, K., Hinds, L., & Fleming, J. (2008). Encouraging public cooperation and support for
police. Policing and Society, 18(2), 136-155.
Nassauer, A., & Legewie, N.M. (2018). Video data analysis: A methodological frame for a novel
research trend. Sociological Methods & Research, 1-40.
Pollock, W., Scott, S., & Moore, E. (2020). Using problem-oriented policing to address police
problems through the study of body-worn camera footage. The Police Journal: Theory,
Practice and Principles, 1–18.
Powell, Z. A., Meitl, M. B., & Worrall, J. L. (2017). Police consent decrees and section 1983
civil rights litigation. Criminology & Public Policy, 16(2), 575–605.
Prokos, A., & Padavic, I. (2002). ‘There oughtta be a law against bitches’: Masculinity lessons in
police academy training. Gender, work and organization, 9(4), 439-459.
Putnam, R.D. (1995). Bowling alone: America’s declining social capital. Journal of Democracy,
6(1), 65-78.
52
Sampson, R. J., & Bartusch, D. J. (1998). Legal cynicism and (subcultural?) tolerance of
deviance: The neighborhood context of racial differences. Law & Society Review, 32(4),
777-804.
Scott, K., Heslop, L., Kelly, T., & Wiggins, K. (2015). Intervening to prevent repeat offending
among moderate- to high-risk domestic violence offenders: A second-responder program
for men. International Journal of Offender Therapy and Comparative Criminology, 59(3),
273-294.
Svensson, J.S. & Saharso, S. (2015). Proactive policing and equal treatment of ethnic-minority
youths. Policing and Society, 25(4), 393-408.
Terrill, W. (2003). Police use of force and suspect resistance: The micro-process of the police
suspect encounter. Police Quarterly, 6, 51-83.
Terrill, W. & Mastrofski, S.D. (2002). Situational and officer-based determinants of police
coercion. Justice Quarterly, 192, 215–248.
Terrill, W., Paoline, E.A., & Gau, J.M. (2016). “Three pillars of police legitimacy: Procedural
justice, use of force, and occupational culture,” The Politics of Policing: Between Force
and Legitimacy (Sociology of Crime, Law and Deviance, Vol. 21), Emerald Group
Publishing Limited, pp. 59-76.
Weisburd, D. et al. (2019). Proactive policing: A summary of the Report of the National
Academies of Sciences, Engineering, and Medicine. Asian Journal of Criminology, 14,
145-177
Willits, D. W., & Makin, D. A. (2018). Show me what happened: Analyzing use of force through
analysis of body-worn camera footage. Journal of Research in Crime and Delinquency,
55(1), 51-77.
Wood, G., Tyler, T.R., & Papachristos, A.V. (2020). Procedural justice training reduces police
use of force and complaints against officers. Proceedings of the National Academy of
Sciences, 117(18), 9815-9821.
53
Court Cases Cited
54