Police Use of Force in the United States: Non-Lethal Use of Force Analysis
Police Use
of Force in the United States:
Non-Lethal
Use of Force Analysis
January 1, 1992
– December 31, 2007
David Carrier, Kelly Feeney, Cuvee Zhou
Data Analysis report for:
[CJ 591] Applied Analytical Methods
May 2021
---------------------
Variables List & Data Report
Independent
Variables:
· First
Officer Years of Experience
· First
Suspect Race
· First
Suspect Age
· Suspect
Had Weapons
· Call
Types (recoding)
· Suspect
Intoxication (Alcohol/Drug Impairment)
· Maximum
of Suspect Behavior
· First
Officer Race
Dependent
Variables:
· OC
Used in Incident
· CED
Used in Incident
Report:
Taylor, B., Koper, C.
(2013, October 29) Evaluation of Less-Lethal Technologies on Police
Use-of-Force Outcomes in 13 Sites in the United States, 1992-2007. U.S
Department of Justice, Office of Justice Programs. Doi:
10.3886/ICPSR27561.v1
Introduction
In recent years, police use of force has come into question throughout
the United States. Because of a rapid expansion in available technology for law
enforcement and the public, pictures, videos, speculation, and politics have
provided multiple lenses towards the matter. This report uses empirical data from
the U.S. Department of Justice for the years 1992 to 2007 to answer specific
questions regarding police use of force. The outcome of this report will be
providing the answers that are necessary to determine potential problems within
policing practice and/or training.
Los Angeles, 1991, the LAPD used batons to
beat Rodney King, an African American male, during his arrest after a high-speed
chase. King was evading arrest for drunk driving and driving while intoxicated
when the police finally stopped him and pulled him from the car. An uninvolved
citizen caught the incident on video and shared it to the media. The footage
sparked a national outrage, and police use of force has been on watch ever
since. The acquittal of all four officers involved prompted the notorious 1991
LA Riots and led many to believe the criminal justice system was foundationally
corrupt[1].
Since then, body camera and citizen footage of similar incidents with police
have routinely been shared to question both lethal and non-lethal use of force
tactics. In 2020, a Minneapolis police officer was arrested and charged with
the murder of George Floyd, an African American male, after bystander cell
phone footage recorded the officer putting his knee on Floyd’s neck during the
arrest. This move was initially proclaimed by the Minneapolis Police Department
as an advanced tactic used to subdue an uncooperative suspect. In court,
however, the Minneapolis police use of force training instructor testified that
it was not a trained neck restraint tactic, though it is commonly used in police
departments throughout the U.S.[2]
The lack of transparency in training and officer accountability once again led
to widespread protests and rioting across the nation. The aim of the protests
was that the element of race played a role in the use of force during basic
traffic stops. Notably during these protests was the use of non-lethal weapons
such as bean bag rifles, tear gas, tasers, and pepper spray. While these are
commonly used as non-lethal crowd-control and arresting weapons, there were multiple
reports of serious injury and even death[3].
This was not the first time non-lethal force used by police had caused a
fatality. In 2019, George Zapantis, an African American male with Bipolar
Disorder, was wielding a sword when he suffered a cardiac arrest and died after
an NYPD officer tased him four times[4].
Similar incidents have occurred in the past, all with their own unique factors
that detail the use of force with non-lethal weapons.
Understanding the history and context of service
weapons is important to analyzing the data and conclusions within this report. Non-lethal
use of force has been relied on by law enforcement for as long as policing has
existed in America. In the mid-19th Century, police officers in New
York and Boston relied on various types of weapons, mostly wooden clubs, until
around the late 1800’s[5].
At this time, criminals became better-armed, and departments issued their
officers a service firearm weapon to keep par. Since lethal force had been
introduced to the array of weapons police officers use, departments needed to
determine specifically when each item should be used. According to federal law,
lethal force can be used specifically in these five scenarios[6]:
1. Self-Defense:
necessary when one reasonably believes himself/herself to be in imminent danger
of death or serious bodily harm
2. Serious
offenses against persons: necessary to prevent
the commission of a serious offense against a person(s) in circumstances
presenting an imminent danger of death or serious bodily harm
3. Nuclear
weapons or nuclear explosive devices: necessary to
prevent the theft, sabotage, or unauthorized control of a nuclear weapon or
nuclear explosive device
4. Special
nuclear material: necessary to prevent the theft, sabotage,
or unauthorized control of special nuclear material
5. Apprehension:
necessary to prevent the escape of a person believed to have A) committed an
offense of the nature in reasons 1-4 or B) is escaping by use of a weapon or
explosive or whom otherwise indicates that he/she poses a significant threat of
death or serious bodily harm
If
feasible, a warning such as an order to halt should be given before a shot is
fired, though it is not required. All other instances where force may be
needed, but lethal force may not be used, require the use of non-lethal weapons.
Specific policy regarding non-lethal use of force weapons varies by department
throughout the U.S. Although this report cannot generalize for the sake of
validity, an example of a common policy can be examined in the Rye Brook Police
Department Non-Lethal Use of Force reference summary from the New York State
Accreditation Standards:
“A police officer, in the course
of effecting or attempting to effect an arrest, or
preventing or attempting to prevent
the escape from custody of a person whom he
reasonably believes to have committed
an offense, may use non-deadly force to the
extent he reasonably believes such to
be necessary to effect the arrest, or to prevent
the escape from custody, or to defend
himself, or a third person from what he
reasonably believes to be the use or
imminent use of physical force.”[7]
In
summary, for officers during a traffic stop or in responding to a call for
service, non-lethal weapons may be administered in the apprehension of a
suspect attempting to resist or flee custody. The decision between lethal and
non-lethal force lies within the officer’s discretion of the department policy
regarding use of force. That discretion generally accounts for whether the
safety of the officer, suspect, or a third-party individual(s) is being
threatened. However, the discretion of officers is, frankly, unmeasurable to a
significant degree. Therefore, the contents of this report going forward will
include only prior research and data that is able to be analyzed in theory and
statistics.
Specific non-lethal service weapons have
been designated to law enforcement across the country. Though not every
department has the same weapons or guidelines for using them, the most common
non-lethal weapons provided to police officers are conducted electrical devices
(CED) and oleoresin capsicum (OC). These two weapons, commonly known as the
“taser” and “pepper spray” are frequently used in confrontations, suicide
interventions, riots, and prison disturbances[8].
A CED in law enforcement patrol is the officer’s taser, a less-lethal single
shot weapon designed to temporarily incapacitate a suspect using an electrical
current[9].
It is a hand-held weapon similar in shape and size to a pistol but is normally
bright yellow in color. The taser can either be applied directly onto an
individual or projected from a distance. Other CED’s, more frequently used for
riots and prison control, include electric shock shields and stun batons. All
work by sending small doses of electrical currents to the nervous system,
producing a “shock” effect that temporary immobilizes the brain from sending
signals throughout the body. An OC in law enforcement patrol is the officers mace/pepper
spray, formally named OC for its scientific origin (oleoresin capsicum). Time
for a quick chemistry lesson! Oleoresin capsicum is an oily resin derivative
from capsicums (solanaceous nightshade species plants). The capsaicinoids are
distilled, dried, and compounded together. The final oleoresin contains several
branched-chain alkyl vanillylamides, in addition to capsaicin, the major
component in OC[10]. Capsaicinoids
cause dermatitis as well as nasal, ocular, pulmonary, and gastrointestinal
effects in humans. OC gained popularity in the 1990s as a defensive weapon for
civilians and law enforcement agencies because they produce immediate,
temporary immobilization and incapacitation when sprayed directly into the face
or eyes[11].
OC produces significant eye irritation with watering, pain, and temporary
blindness. And a fun fact, capsaicin
comes from the same plant genus as chili peppers, yum! The advanced technology
of non-lethal weapons for policing have become an element of significance in
scenarios where they are needed.
There are two initial questions that need
to be identified with prior research evaluation when discussing non-lethal use
of force weapons: do administrative policies regarding lethal and non-lethal
force affect its usage; and, do factors of the scenario in which they are being
used play into its use? In 2018, a research study within the Chicago Police
Department was conducted to determine if policies administering non-lethal
weapons worked to replace lethal force. The CPD changed its taser policy in
2010, issuing the weapons to all patrol officers instead of restricting the use
to sergeants. Researchers wanted to determine if this change affected the type
of force employed by the officer, along with other outcomes. Initial results
found the policy change led to a large increase in the use of tasers. After a
re-training period, substitution from other types of force (both greater and
lesser) increased. The research concluded that while police injuries fell, the
injury rates of civilians were not affected, and there was no evidence that CED
led to a reduction in police use of lethal force[12].
A similar study published in the Justice Quarterly from 2016 accounted
for the impact of administrative policy on non-lethal force. These researchers
were also interested in whether policy changed the use of force, specifically
analyzing the results from restrictive versus permissive policy environments in
3,340 use of force incidents from three U.S. agencies. In the Chicago Police
Department study, the CPD had to re-train its officers after reviewing high
levels of non-lethal use of force post-policy implementation. The Justice
Quarterly study backs this summary, stating consistent results showing that
officers working with the most restrictive policy framework used force less
readily than officers who operated within more permissive policy environments[13].
These prior research conclusions significantly strengthen the argument that
administrative policy plays a key role in the use of force.
The second paragraph of this introductory
section gave real-life scenarios as an example to the different variables
involved in an incident where force was used. Prior research evaluation is
necessary to make a hypothesis regarding these variables on use of force. Specifically,
one of the independent variables in this report is the suspect’s race. An
Empirical Analysis of Racial Differences in Police Use of Force, a research
study by Roland Fryer, discusses the potential relationship between race and
use of force. Lethal use of force saw no racial differences in both raw data
and the data when other variables were considered. On non-lethal use of force,
African Americans and Hispanics were more than fifty percent more likely to
experience some form of force in the interaction with police[14].
The author notes that this disparity can be reduced, but not fully explained,
by adding controls that account for important context and civilian behavior
regarding the interaction. This includes other variables like the ones used in
this report. Expanding on this, a similar research study in the British
Journal of Criminology investigated the differences in use of force
regarding the officer race, which this report also analyzed. The findings
revealed that white officers were more coercive toward black suspects, and
tended to use non-lethal force more frequently, but black officers’ force usage
is unaffected by suspect race, with equal amount of non-lethal force used[15].
Conversely, officer race did not predict resistance among white or black
suspects. The analysis by Fryer and the BJC provide substantial evidence
towards one or multiple variables having an influence in use of force
incidents.
The analysis of data within this report,
paired with previous empirical research review, should enhance the reader’s
understanding of the different factors that go into use of force. Whether those
factors have an impact on the administration of force is the guiding question
of the report and will be the salient discussion point going forward. This
report analyzes multiple variables involved in police use of force incidents during
routine traffic stops and calls for service, hypothesizing that certain characteristics
of these events increase the likelihood of use of force involving conducted
electrical devices (CED) and oleoresin capsicum (OC). The analysis hypothesizes
that younger suspects experience involvement in less-than-lethal use of force
than older suspects, and suspects who display threatening and aggressive
behavior towards law enforcement have force used against them at higher rates.
---------------------
Data Source
The initial study
the data originated from aimed to examine how law enforcement agencies managed
use of force. Specifically, the study sought to assist agencies in establishing
guidelines regarding Conducted Energy Device (CED) deployment to ensure safety for
officers and members of the public. The data included thirteen law enforcement
agencies from around the United States. Seven agencies that did utilize CED and
maintained related policies were matched to six agencies that did not. For
those agencies that did utilize CED, data was collected two years prior to
deployment and two years after deployment. For the agencies that did not
utilize CED, data was collected for four years over synonymous time periods.
The researchers collected data at the incident-level for 17,965 incidents where
force was used by law enforcement officers. The data received included
incidents that occurred between 1992 and 2007. Researchers collaborated with
the law enforcement agencies to obtain the use of force data available through
the agency’s own database. They also examined the FBI’s Uniform Crime Report
system and the 2000 US Census to gather crime and demographic data on the
cities associated with the participating agencies. The study of origin happened
in three parts. Part One incorporated all available use of force data from the
participating agencies and focused on the individual officers and suspects
involved in the use of force. Part Two scrutinized the chosen law enforcement
agencies along with the crime and demographic data. Part Three encompassed the
same information from Part One but employed only the data in the focal time
periods and derived variables that concerned all officers and suspects involved
in a use of force incident. For confidentiality purposes, researchers masked
some of the information to prevent recognition of specific officers, suspects,
bystanders, or incidents (Taylor & Koper, 2013).
Taylor, B., Koper, C.
(2013, October 29) Evaluation of Less-Lethal Technologies on Police
Use-of-Force Outcomes in 13 Sites in the United States, 1992-2007. U.S
Department of Justice, Office of Justice Programs. Doi: 10.3886/ICPSR27561.v1
https://www.icpsr.umich.edu/web/ICPSR/studies/27561
Variable Measurement &
Descriptive Statistics
One of the primary dependent variables this analysis
examined centered on deployment of CED during the use of force incident among
any officer present. The measure of this variable was nominal, with respondents
giving a ‘yes’ or ‘no’ answer. 719 (4.0% of all collected incidents) cases did
not provide data. For the incidents that did provide CED use data, 92.3%
(n=15,918) of incidents reported no use of CED while only 7.7% (n=1,328) of
incidents did involve use of CED.
Table 1:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
CED Used in Incident |
|||
|
No |
15,918 |
92.3 |
|
|
Yes |
1,328 |
7.7 |
|
|
Missing |
(719) |
(4.0) |
The secondary dependent variable for this analysis
focused on use of OC spray by an officer during the use of force incident. This
measure was nominal and included a dichotomous response of ‘yes’ or ‘no’. 719
(4.0% of all collected incidents)did not provide data. Of those that had data
available, 85.8% (n=14,792) of incidents did not involve use of OC spray and
14.2% (n=2,454) did entail deployment of OC spray.
Table
2:Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
OC Spray Used in Incident |
|||
|
No |
14,792 |
85.8 |
|
|
Yes |
2,454 |
14.2 |
|
|
Missing |
(719) |
(4.0) |
The type of call which initiated suspect/officer
contact was also examined as part of the analysis. Call Type fell into nine
different categories: Pedestrian Contact, Traffic Stop, Robbery, Burglary,
Under the Influence of Drugs or Alcohol, Domestic Dispute, Emotionally
Disturbed Person/Mental Illness, Assault (excluding domestic), and Other. For
this analysis, some of the categories were combined for ease of use. Robbery
and Burglary were grouped into one category, as well as Alcohol/Drug and Mental
Illness. 8,828 incidents (49.1% of all gathered information) did not report
data about call type. The most frequent was Other, which included 68.9%
(n=6,298) of the calls; nothing further was specified. Assault (excluding
domestic) made up 6.6% (N=599) of the calls, followed by Domestic Dispute which
comprised 5.7% (n=520) of the calls. Traffic Stops encompassed 5.6% (n=509) of
calls. The combined Alcohol/Drug and Mental Illness category included 8.9%
(n=815) of calls. Only a combined 3.1% (n=280) of calls were due to
Robbery/Burglary, and finally 1.3% (n=116) due to pedestrian contact.
Table 3:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
Call Type |
|||
|
Pedestrian Contact |
116 |
1.3 |
|
|
Traffic Stop |
509 |
5.6 |
|
|
Robbery/Burglary |
280 |
3.1 |
|
|
Alcohol/Drug & Mental
Illness |
815 |
8.9 |
|
|
Domestic Dispute |
520 |
5.7 |
|
|
Assault (excluding domestic) |
599 |
6.6 |
|
|
Other |
6,298 |
68.9 |
|
|
Missing |
(8,828) |
(49.1) |
The analysis examined whether the suspect(s) possessed
a weapon. This nominally measured variable required a ‘yes’ or ‘no’ response.
7,404 incidents (41.2%) were missing this data. Most of the suspect(s) did not
possess a weapon (n=8,506, 80.5%) while 19.5% (n=2,055) did possess some sort
of weapon.
Table 4: Descriptive
Statistics
|
Variable |
Measurement |
N |
Percentage |
|
Suspect(s) Had Weapon |
|||
|
No |
8,506 |
80.5 |
|
|
Yes |
2,055 |
19.5 |
|
|
Missing |
(7,404) |
(41.2) |
This analysis explored the race of the first suspect
involved in the use of force. This variable was nominally measured and included
responses White (not Hispanic), Black or African American (not Hispanic),
Hispanic or Latino, and Other. Of all incidents, 1,819 (10.1%) did not include
data about suspects' race. Black/African American suspects held the highest
frequency (42.9%, n=6,926), followed by White suspects (33.4%, n=5,388), then
Hispanics/Latinos (21.4%, n=3,453). A small percentage (2.3%, n=379) of
suspects identified as Other.
Table 5:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
First Suspect Race |
|||
|
White (not of Hispanic
descent) |
5,388 |
33.4 |
|
|
Black or African American
(not of Hispanic descent) |
6,926 |
42.9 |
|
|
Hispanic or Latino |
3,453 |
21.4 |
|
|
Other Race |
379 |
2.3 |
|
|
Missing |
(1,819) |
(10.1) |
Another variable of scrutiny involved years of
experience among the first officer involved. The ordinal responses included
less than one year of experience, one to two years of experience, three to four
years, five to six years, seven to eight years, nine to ten years, and more
than ten years of experience. Of all the incidents, approximately half (50.4%,
n=9,060) did not include information about the officer's experience. The
highest frequency of officers involved in use of force incidents had more than
ten years of experience (21.3%, n=1,896) followed by officers with only one to
two year’s experience (20.0%, n=1,784) and officers with three to four year’s
experience (19.9%, n=1,772). 12.4% (n=1,102) of officers had seven to eight
year’s experience, 6.8% (n=605) had nine to ten years, and only 2.7% (n=237)
had less than one year of experience. Many officers (59.9%) involved in uses of
force incidents had less than seven years of experience. The mean fell into the
4.22 ordinal category, with a standard deviation of 1.872.
Table 6:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
Mean |
Standard Deviation |
|
First Officer Year’s
Experience |
4.22 |
1.872 |
|||
|
<1 Year Experience |
237 |
2.7 |
|||
|
1-2 Year’s Experience |
1,784 |
20.0 |
|||
|
3-4 Year’s Experience |
1,772 |
19.9 |
|||
|
5-6 Year’s Experience |
1,509 |
16.9 |
|||
|
7-8 Year’s Experience |
1,102 |
12.4 |
|||
|
9-10 Year’s Experience |
605 |
6.8 |
|||
|
>10 Year’s Experience |
1,896 |
21.3 |
|||
|
Missing |
(9,060) |
(50.4) |
Alcohol/Drug Impairment of the first suspect involved
in the use of force was a variable of interest. This dichotomous and nominally
measured variable consisted of ‘yes’ or ‘no’ answers. 4,467 incidents (24.9%)
from the data collection had missing data and 787 (4.4%) unknown. Of the
incidents that did provide data, 65.9% (n=8,378) of suspects did not have
impaired by drugs or alcohol and 34.1% (4,333) did have some form of
impairment.
Table 7:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
Alcohol/Drug Impairment -
Suspect 1 |
|||
|
No |
8,378 |
65.9 |
|
|
Yes |
4,333 |
34.1 |
|
|
Don’t Know |
787 |
(4.4) |
|
|
Missing |
(4,467) |
(24.9) |
The age of the first suspect involved in the use of
force was ordinally measured, with categories ranging 17 or younger, 18-20,
21-25, 26-30, 31-40, 41-50, 51-60, and 61 or older. This variable missed 3,988 pieces
of data (22.2%). The youngest group made up 7.1% (n=988) and the oldest group
made up 1.0% (n=135). The 18-20-year-old age group comprised 18.1% (n=2,527).
Both the 21-25 and 26-30-year-old age groups comprised 15.2% (n=2,124 &
n=2,122, respectively). The 31-40-year-old made up 26.7% (n=3,732),
41-50-year-old 13.2% (n=1,850), and 51-60-year-old 3.6% (n=499). The mean fell
into the 2.95 ordinal category, with a standard deviation of 1.668.
Table 8:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
Mean |
Standard Deviation |
|
First Suspect Age |
2.95 |
1.668 |
|||
|
17or Younger |
988 |
7.1 |
|||
|
18-20 |
2,527 |
18.1 |
|||
|
21-25 |
2,124 |
15.2 |
|||
|
26-30 |
2,122 |
15.2 |
|||
|
31-40 |
3,732 |
26.7 |
|||
|
41-50 |
1,850 |
13.2 |
|||
|
51-60 |
499 |
3.6 |
|||
|
61 or Older |
135 |
1.0 |
|||
|
Missing |
(3,988) |
(22.2) |
A variable of interest involved suspect behavior. This
nominal variable included answers such as No Resistive Behavior Recorded,
Passive, Fleeing, Verbal, Mild Aggression, Severe Aggression, Deadly Force, and
Other. A total of 11,633 (64.8%) incidents did not report on this variable.
Severe Aggression was reported in 23.9% of incidents (n=1,516), followed
closely by No Resistant Behavior Recorded (22%, n=1,392) and Other (21.1%,
n=1,336). Fleeing included 15.1% (n=956) of incidents, Mild Aggression 7.6%
(n=481), Passive 6.2% (n=392), and Deadly Force just 1.4% (n=91).
Table 9:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
Maximum of Suspect Behavior |
|||
|
No Resistive Behavior
Recorded |
1,392 |
22.0 |
|
|
Passive |
392 |
6.2 |
|
|
Fleeing |
956 |
15.1 |
|
|
Verbal |
168 |
2.7 |
|
|
Mild Aggression |
481 |
7.6 |
|
|
Severe Aggression |
1,516 |
23.9 |
|
|
Deadly Force |
91 |
1.4 |
|
|
Other |
1,336 |
21.1 |
|
|
Missing |
(11,633) |
(64.8) |
A variable that explored officer demographics included
race of the first officer involved in the use of force. This variable was
nominally measured and included response options White, Black or African
American, Hispanic, or Latino, or Other Race. 40.6% (n=7,286) of data for this
variable was not reported. White officers had the highest frequency of use of
force involvement (80.8%, n=8,625), followed by Black/African American (10.3%,
n=1,105), Hispanic/Latino (5.8%, n=621), and lastly Other (3.1%, n=328).
Table 10:
Descriptive Statistics
|
Variable |
Measurement |
N |
Percentage |
|
First Officer Race |
|||
|
White, not Hispanic descent |
8,625 |
80.8 |
|
|
Black or African American,
no Hispanic descent |
1,105 |
10.3 |
|
|
Hispanic or Latino |
621 |
5.8 |
|
|
Other Race |
328 |
3.1 |
|
|
Missing |
(7,286) |
(40.6) |
---------------------
Bivariate Analyses and Results
First Suspect Race
Only 8.5% of suspects had force utilizing CED used
against them. 9.0% of Black/African American suspects were involved in ED use,
8.3% of Hispanic/Latino suspects, 8.3% of the first suspect’s that identified
as Other racially, and 8.0% of White suspects were involved in CED use of force
during their encounter. A cross-tabulation test indicates race of the first
suspect was not statistically significant at the .05 level (Pearson’s
chi-square=3.57, df=3, p=0.312). A total of 15.2% of first suspects were
involved of OC Spray use of force. 16.5% of Black/African American suspects
were involved in these incidents, 15.3% of White suspects, 13.7% of suspects
that identified as Other racially, and 12.7% of Hispanic/Latino suspects. A
cross-tabulation determined this statistically significant at the .05 level
(Pearson’s chi-square=25.8, df=3, p<.001).
Table 11:
Less-Than-Lethal Use of Force by First Suspect Race
|
Use of Force Type |
First Suspect Race |
||||
|
White |
Black/African American |
Hispanic/Latino |
Other |
Total |
|
|
CED |
8.0% (n=405) |
9.0% (n=595) |
8.3% (n=282) |
8.3% (n=31) |
8.5% (n=1313) |
|
OC Spray |
15.3% (n=775) |
16.5% (n=1094) |
12.7% (n=429) |
13.7% (n=51) |
15.2% (n=2349) |
First Suspect Age
Just 4.8% of first suspects encountered had CED used
against them. 5.7% of suspects 26 to 30 years old were involved in CED use,
5.7% of suspects 41 to 50 years old, 5.6% of suspects 31 to 40 years old, 5.2%
of suspects 61 or older, 4.8% of suspects between 21 to 25 years old, 4.8% of
suspects between 51 and 60 years old, 3.2% of suspects between the ages of
eighteen to twenty years old, and 1.9% of suspects seventeen years old or
younger were involved in CED use of force. A cross-tabulation test indicates age
of the first suspect statistically significant at the .05 level (Pearson’s
chi-square=43.81, df=7, p=<.001. A total of 15.6% of first suspects police
encountered had OC Spray used against them. 18.2% of suspects 21 to 25 years of
age were involved in OC Spray use, 16.6% of suspects 31 to 40 years old, 16.3%
of suspects 26 to 30 years old, 15.2% of suspects between 41 and 50 years old,
14.2% of suspects ages eighteen to twenty years old, 12.9% of suspects between
51 and 60, 11.6% of suspects aged 17 years or younger, and just 6.7% of
suspects aged 61 or older had OC Spray used on them. Age of the first suspect
involved proved significant via cross-tabulation at the .05 level (Pearson’s
chi-square=41.85, df=7, p<.001).
Table 12:
Less-Than-Lethal Use of Force by First Suspect Age
|
Use of Force Type |
First Suspect Age |
||||||||
|
17 or Younger |
18-20 |
21-25 |
26-30 |
31-40 |
41-50 |
51-60 |
61 or Older |
Total |
|
|
CED |
1.9% (n=19) |
3.2% (n=81) |
4.8% (n=101) |
5.7% (n=121) |
5.6% (n=207) |
5.7% (n=105) |
4.8% (n=24) |
5.2% (n=7) |
4.8% (n=665) |
|
OC Spray |
11.6% (n=114) |
14.2% (n=358) |
18.2% (n=385) |
16.3% (n=345) |
16.6% (n=620) |
15.2% (n=280) |
12.9% (n=61) |
6.7% (n=9) |
15.6% (n=2175) |
Maximum Suspect
Behavior
A total of 17.4% of suspects police encountered had
CED used on them. 32.4% of suspects fleeing were involved in CED use, 26.8% of
suspects who displayed severe aggression, 26.7% of suspects who acted against
officers with deadly force, 15% of suspects who had no resistive behavior
recorded, 11.2% of suspects who displayed mild aggression, 7.1% of verbal, 6.2%
of suspects whose behavior was categorized as Other, and 1.8% of passive
suspects had CED used against them. A test of cross-tabulation determined
maximum suspect behavior was significant at the .05 level (Pearson’s
chi-square=460.27, df=7, p<.001). Also, a total of 12.3% of suspects had OC
Spray used on them by police. 27.2% of suspects who displayed mild aggression
were involved in OC Spray deployment, 25.0% of verbal, 19.0% of suspects who
displayed severe aggression, 19.0% of suspects who acted with deadly force,
13.6% suspects who displayed Other behavior, 7.5% of passive suspects, 5.6% of
fleeing suspects, and 3.2% of suspects who had no resistive behavior recorded.
A cross-tabulation proved suspect behavior significant at the .05 level
(Pearson’s chi-square=343.35, df=7, p<.001).
Table 13:
Less-Than-Lethal Use of Force by Maximum Suspect Behavior
|
Use of Force Type |
Maximum Suspect Behavior |
||||||||
|
No Resistive Behavior
Recorded |
Passive |
Fleeing |
Verbal |
Mild Agg |
Severe Agg |
Deadly Force |
Other |
Total |
|
|
CED |
15% (n=209) |
1.8% (n=7) |
32.4% (n=307) |
7.1% (n=12) |
11.2% (n=82) |
26.8% (n=53) |
26.7% (n=406) |
6.2% (n=24) |
17.4% (n=1100) |
|
OC Spray |
3.2% (n=44) |
7.5% (n=29) |
5.6% (n=53) |
25.0% (n=42) |
27.2% (n=181) |
19.0% (n=129) |
12.2% (n=288) |
13.6% (n=11) |
12.3% (n=777) |
Suspect(s) Had
Weapon
A mere 10.9% of suspects who engaged with police were
involved in use of force utilizing CED. 11.9% of suspects who did not possess a
weapon had CED used against them compared to 6.6% of suspects who did possess a
weapon. Whether or not the suspect had a weapon proved significant via
cross-tabulation at the .05 level (Pearson’s chi-square=4.36, df=1, p<.001).
An overall 15.8% of suspects had OC Spray deployed on them during their
encounters with police. 20.5% of suspects who did possess a weapon were
involved in OC Spray use, compared to 14.6% of suspects who did not possess a
weapon. This factor proved significant through cross-tabulation at the .05
level (Pearson’s chi-square=42.44, df=1,p<.001).
Table 14: Less-Than-Lethal
Use of Force by Suspect(s) Had Weapon
|
Use of Force Type |
Suspect(s) Had Weapon |
||
|
Yes |
No |
Total |
|
|
CED |
6.6% (n=135) |
11.9% (n=1011) |
10.9% (n=1146) |
|
OC Spray |
20.5% (n=421) |
14.6% (n=1246) |
15.8% (n=1667) |
Alcohol/Drug
Impairment - Suspect One
Only 8.3% of suspects had CED deployed on them during
their police encounter. 14.7% of suspects with alcohol or drug impairment were
involved in CED, compared to 5.0% of suspects who did not have alcohol or drug
impairment. This factor proved significant through cross-tabulation (Pearson’s
chi-square=352.04, df=1, p<.001). A total of 15.4% of suspects had OC Spray
used on them during their interaction with police. 15.7% of suspects who were
not impaired alcohol or drug impairment were involved in OC Spray use compared
to 15.0% of suspects who did have an impairment caused by alcohol or drugs. A
cross-tabulation determined alcohol/drug impairment not significant at the .05
level (Pearson’s chi-square=1.08, df=1, p=0.299).
Table 15: Less-Than-Lethal
Use of Force by Alcohol/Drug Impairment-Suspect One
|
Use of Force Type |
Alcohol/Drug Impairment |
||
|
Yes |
No |
Total |
|
|
CED |
14.7% (n=633) |
5.0% (n=416) |
8.3% (n=1049) |
|
OC Spray |
15.0% (n=646) |
15.7% (n=1310) |
15.4% (n=1956) |
First Officer Race
Just 11.4% of first officers utilized CED. 16.1% of
Hispanic/Latino officers deployed CED, 11.3% of White officers, 10.1% of
officers who identified as Other racially, and 9.6% of Black/African American
officers deployed CED. A cross-tabulation indicated race of officers
significant at the .05 level (Pearson’s chi-square=17.48, df=3, p<.001). A
sum of 16.0% of officers utilized CO Spray. 16.7% of White officers deployed OC
Spray, 14.3% of officers identified as Other race, 13.4% of Hispanic/Latino
officers, and 12.2% of Black/African American officers. A test of
cross-tabulation proved officer race significant at the .05 level (Pearson’s
chi-square=18.050, df=3, p<.001).
Table 16:
Less-Than-Lethal Use of Force by First Officer Race
|
Use of Force Type |
First Officer Race |
||||
|
White |
Black/African American |
Hispanic/Latino |
Other |
Total |
|
|
CED |
11.3% (n=906) |
9.6% (n=102) |
16.1% (n=100) |
10.1% (n=29) |
11.4% (n=1137) |
|
OC Spray |
16.7% (n=1335) |
12.2% (n=130) |
13.4% (n=83) |
14.3% (n=41) |
16.0% (n=1589) |
First Officer
Year’s Experience
An overall 10.7% of first officers involved deployed
CED. 16.9% of officers with less than one year of experience used CED, 12.9 of
officers with five to six years of experience, 12.7% of officers with seven to
eight year’s experience, 11.2 of officers with three to four years of
experience, 9.7% of officers with nine to ten year’s experience, 9.0% of
officers with one to year’s experience, and 8.3% of officers with more than ten
year’s experience utilized CED. Officer years of experience proved significant
for use of CED via a test of cross-tabulation (Pearson’s chi-square=37.03,
df=6, p<.001). A total of 16.0% of first officers deployed OC Spray when
encountering suspects. 9.2% of officers with less than one year experience used
OC Spray, 14.3% of officers with one to two year’s experience, 16.8% of
officers with three to four years, 15.9% of officers with five to six years,
15.6% of officers with seven to eight years, 15.4% of officers with nine to ten
years, and 18.1% of officers with more than ten years of experience deployed OC
Spray. How many years of experience an officer had proved significant through a
cross-tabulation at the .05 level (Pearson’s chi-square=17.02, df=6,
p=.009).
Table 17:
Less-Than-Lethal Use of Force by First Officer Year’s Experience
|
Use of Force Type |
First Officer Years of Experience |
|||||||
|
<1 Year |
1-2 Years |
3-4 Years |
5-6 Years |
7-8 Years |
9-10 Years |
>10 Years |
Total |
|
|
CED |
16.9% (n=35) |
9.0% (n=141) |
11.2% (n=179) |
12.9% (n=182) |
12.7% (n=137) |
9.7% (n=56) |
8.3% (n=147) |
10.7% (n=877) |
|
OC Spray |
9.2% (n=19) |
14.3% (n=225) |
16.8% (n=269) |
15.9% (n=224) |
15.6% (n=168) |
15.4% (n=89) |
18.1% (n=321) |
16.0% (n=1315) |
Call Type
Only 2.8% of calls resulted in officer deployment of
CED. 13.6% of assault calls resulted in CED use, 7.2% of robbery/burglary
calls, 5.2% of calls for alcohol/drug or mental illness, 1.7% of all other call
types, 0.8% of traffic stops, 0.2% of domestic disturbances, and 0.0% of
pedestrian contacts. A test of cross-tabulation proved call type significant at
the .05 level (Pearson’s chi-square=342.7, df=6, p<.001). A total of 17.1%
of calls resulted in OC Spray being deployed by officers. 21.7% of domestic disturbance
calls resulted in use of OC Spray, 21.7% of domestic disturbance calls,
19.3% of calls for alcohol/drug or mental illness, 19.0% of assaults, 19.0% of
pedestrian contacts, 16.9% of all other calls, 13.0% of traffic stops, and 8.6%
of calls for robbery/burglary. This proved significant at the .05 level through
cross-tabulation (Pearson’s chi-square=32.80, df=6, p<.001).
Table 18:
Less-Than-Lethal Use of Force by Type of Call
|
Use of Force Type |
Call Type |
||||||||
|
Pedestrian Contact |
Traffic Stop |
Robbery/Burglary |
Alcohol/Drug or Mental
Illness |
Domestic Disturb |
Assault |
Other |
Total |
||
|
CED |
0.0% (n=0) |
0.8% (n=4) |
7.2% (n=20) |
5.2% (n=42) |
0.2% (n=1) |
13.6% (n=81) |
1.7% (n=109) |
2.8% (n=257) |
|
|
OC Spray |
19.0% (n=22) |
13.0% (n=66) |
8.6% (n=24) |
19.3% (n=156) |
21.7% (n=113) |
19.0% (n=113) |
16.9% (n=1063) |
17.1% (n=1557) |
|
---------------------
Multivariate Regression
Analyses & Results
The regression analysis showed that race of the
officer involved in OC Spray use did not always prove significant. Compared to
White officers, Black/African American officers had an odds of .556 less likely
to use OC Spray. This proved significant at the .05 level (b=-.587, t=16.465,
p<.001). Hispanic/Latino officers had an odds of using OC Spray .911 times
less than White officers, which was not significant (b=-.093, t=.260,
sig=.610). Officers who identified as Other racially had an odds of utilizing
OC Spray .565 times more compared to White officers, also not significant
(b=.125, t=.331, p=.565). There was a positive relationship between officer
years of experience and OC Spray use. For each ordinal category increase,
typically two-year increments, the likelihood of deployment of OC Spray
increased by about 1.027. The relationship between officer years of experience
and use of OC Spray was not statistically significant at the .05 level (b=.027,
t=1.722, sig=.189). When compared to White suspects, Black/African American
suspects were 1.300 times more likely to have OC Spray deployed against them
during police interactions. This proved significant at the .05 level of
analysis (b=.263, t=10.130, p=.001). Hispanic/Latino suspects were .800 times
less likely to have OC Spray used on them. This was not significant (b=-.223,
t=2.893, p=.089). Suspects who identified as Other racially were also .970
times less likely to have OC Spray deployed against them. This was again not
significant at the .05 level (b=-.031, t=.013, p=.970). There was a positive
relationship between age of the first suspect involved and OC Spray deployment.
For every increase in age category, the odds of OC Spray use increased by
1.012. This did not prove significant at the .05 level (b=.012, t=.273,
p=.601). Holding no resistant behavior by the suspect recorded constant, those
suspects who displayed passive behavior towards officers had an increase in
odds of 14.568 of having OC Spray deployed on them. Passive behavior was
significant at the .05 level (b=2.679, t=5.221, p=.022). Suspects
who fled from officers had a decreased odds of .688. Fleeing did not prove
significant (b=-.307, t=.161, p=.688). Suspects who engaged in verbal
confrontations with officers had an odds of 11.344 more likely than suspects
who did not display resistant behavior to have OC Spray deployed. This did
prove significant (b=2.429, t=15.680, p<.001). Suspects who engaged in mild
aggressive behavior towards officers had a 10.180 increased odds of OC Spray
deployment. Mild aggression proved significant at the .05 level (b=2.320,
t=33.569, p<.001). Suspects who displayed severe aggression had a 7.077
increased odds of OC Spray deployment. This was also significant (b=1.957,
t=34.940, p<.001). Suspects who displayed deadly force against officers had
an increased odds of being 2.093 more likely to have OC Spray used against
them. Deadly force used by the suspect was not significant at the .05 level
(b=.738, t=.917, p=.338). Suspect behavior categorized as ‘other’ had a 7.932 increased
odds of OC Spray use compared to suspects who did not display resistant
behavior. This was significant at the .05 level (b=2.071, t=36.409, p<.001).
For those suspects who did possess a weapon, the odds of having OC Spray used
against them was 1.264 more likely than suspects who did not possess a weapon.
This proved significant at the .05 level (b=.234, t=7.030, p=.008). Suspects
who had drug or alcohol impairment had a decreased odds of .552 for OC Spray
deployment compared to those suspects who were not alcohol or drug impaired.
This also proved significant at the .05 level (b=-.595, t=4.507, p=.034). When
controlling for pedestrian contacts, traffic stops had a 4.251 increased odds
of resulting OC Spray deployment. This proved significant at the .05 level
(b=1.447, t=17.958, p<.001). Robbery/burglary calls had an odds of 4.334
more likely to result in OC Spray than pedestrian contacts. This did not prove
significant at the .05 level (b=1.466, t=3.692, p=.055). Calls for alcohol/drug
or mental illness were 9.670 times more likely to result in OC Spray
deployment. This proved significant at the .05 level (b=2.269, t=55.988,
p<.001). Calls for domestic disputes also had increased odds of 7.507 and
proved significant (b=2.016, t=40.992, p<.001). Calls for assault had an
increased odds of 12.836 of resulting OC Spray use. This was significant, as
well (b=2.552, t=32.237, p<.001). All other calls types showed a 5.051
increased odds of OC Spray use compared to pedestrian contacts. This was
significant at the .05 level (b=1.620, t=32.382, p<.001).
Table 19:
Regression Analysis of OC Spray Use
|
Variables |
b |
exp(b) |
s.e |
t |
sig |
|
Black/African American
Officer |
-.587*** |
.556 |
.145 |
16.465 |
<.001 |
|
Hispanic/Latino Officer |
-.093 |
.911 |
.183 |
.260 |
.610 |
|
Other Race Officer |
.125 |
1.133 |
.217 |
.331 |
.565 |
|
First Officer Year’s
Experience |
.027 |
.189 |
.020 |
1.722 |
.189 |
|
Black/African American
Suspect |
.263*** |
1.300 |
.083 |
10.130 |
.001 |
|
Hispanic/Latino Suspect |
-.233 |
.800 |
.131 |
2.893 |
.089 |
|
Other Race Suspect |
-.031 |
.970 |
.271 |
.013 |
.909 |
|
First Suspect Age |
.012 |
1.012 |
.022 |
.273 |
.601 |
|
Passive |
2.679* |
14.568 |
1.172 |
5.221 |
.022 |
|
Fleeing |
-.307 |
.736 |
.765 |
.161 |
.688 |
|
Verbal |
2.429*** |
11.344 |
.613 |
15.680 |
<.001 |
|
Mild Aggression |
2.320*** |
10.180 |
.400 |
33.569 |
<.001 |
|
Severe Aggression |
1.957*** |
7.077 |
.331 |
34.940 |
<.001 |
|
Deadly Force |
.738 |
2.093 |
.771 |
.917 |
.338 |
|
Other Behavior |
2.071*** |
7.932 |
.343 |
36.409 |
<.001 |
|
Suspect(s) Had Weapon |
.234** |
1.264 |
.088 |
7.030 |
.008 |
|
Alcohol/Drug Impairment |
-.595* |
.552 |
.280 |
4.507 |
.034 |
|
Traffic Stop |
1.447*** |
4.251 |
.341 |
17.958 |
<.001 |
|
Robbery/Burglary |
1.466 |
4.334 |
.763 |
3.692 |
.055 |
|
Alcohol/Drug or Mental
Illness |
2.269*** |
9.670 |
.303 |
55.988 |
<.001 |
|
Domestic Dispute |
2.016*** |
7.507 |
.315 |
40.992 |
<.001 |
|
Assault |
2.552*** |
12.836 |
.450 |
32.237 |
<.001 |
|
Other Calls |
1.620*** |
5.051 |
.285 |
32.382 |
<.001 |
Note: ***=p<.001,
**=p<.01, *=p<.05
For the regression analysis pertaining to CED use, the
Call Type variable was omitted from the final analysis. The main purpose of
this overall analysis was to examine the characteristics of officers and
suspects involved in use of force incidents, not necessarily what initiated the
law enforcement contact. The regression analysis showed that as officer years
of experience increased, the odds of CED use also increased by 1.083. Years of
officer experience did not prove significant at the .05 level (b=.080, t=1.832,
p=.176). Compared to White officers, Black/African American officers had a
decreased odds of .514 related to CED deployment. This was not significant at
the .05 level (b=-.665, t=2.854, p=.091). Hispanic/Latino officers had a
decreased odds of .565 compared to White officers when it came to CED use. This
was also not significant (b=-.571, t=2.176, p=.140). Officers who identified as
‘other’ racially had a 1.278 increased odds of CED use compared to White
officers. This did not prove significant at the .05 level (b=.245, t=.080,
p=.777). When controlling for White suspects, Black/African American suspects
had an increased odds of 1.100 times more likely to be involved in incidents
where CED was deployed. This did not prove significant at the .05 level
(b=.096, t=.126, p=.723). Hispanic suspects also showed an increased odds of
1.887 compared to White suspects. This was statistically significant at the .05
level (b=.635, t=4.800, p=.028). Suspects who identified as ‘other’ racially
had an increased odds of 2.449 times more likely to have CED used against them.
This was also not significant at the .05 level (b=.896, t=2.155, p=.142). There
was a positive relation between suspect age and CED deployment. For each
ordinal category increase in age, there was a 1.043 increase in the likelihood
of being involved in an incident where CED was deployed. Suspect age did not
prove significant at the .05 level (b=.042, t=.271 p=.603). For suspect
behavior towards officers, only those behaviors that showed high instances were
included in the regression analysis. No resistance, passive, verbal, and deadly
force were used as the control. Suspects who fled from officers had an increase
in odds of 1.043 times more likely to be involved in CED use compared to the
control group. Fleeing proved significant at the .05 level (b=.1.573, t=21.564,
p<.001). Suspects that displayed mild aggression towards officers had a
decrease in odds .434 times less likely to experience CED deployment. Mild
aggression was not significant at the .05 level (b=-.835 t=1.181, p=.277).
Suspects who were severely aggressive towards officers also had an increase in
odds 1.524 more likely to be involved in CED incidents than the control group.
Severe aggression did not prove significant at the .05 level (b=.421,
t=2.912, p=.088). Suspects noted to possess a weapon had a decreased
likelihood of being involved in CED use of force .738 times more likely than
suspects who did not possess a weapon. Weapon possession by the suspect did not
prove significant at the .05 level (b=-.304, t=.433, p=.510). Suspects who were
impaired by alcohol or drugs also showed a 1081.049 increase in odds more
likely to have CED used against them compared to suspects who were not impaired
by drugs or alcohol. This did prove significant at the .05 level (b=6.986,
t=47.107, p<.001). Important to note, this variable was left off the bar
graph due to extremely high odds coefficient, which would have impacted the
ability to view the other variables.
Table 20:
Regression Analysis of CED Use
|
Variables |
b |
exp(b) |
s.e |
t |
sig |
|
First Officer Year’s
Experience |
.080 |
1.083 |
.059 |
1.832 |
.176 |
|
Black/African American
Officer |
-.665 |
.514 |
.394 |
2.854 |
.091 |
|
Hispanic/Latino Officer |
-.571 |
.565 |
.387 |
2.176 |
.140 |
|
Other Race Officer |
.245 |
1.278 |
.865 |
.080 |
.777 |
|
Black/African American
Suspect |
.096 |
1.100 |
.269 |
.126 |
.723 |
|
Hispanic/Latino Suspect |
.635* |
1.887 |
.290 |
4.800 |
.028 |
|
Other Race Suspect |
.896 |
2.449 |
.610 |
2.155 |
.142 |
|
First Suspect Age |
.042 |
1.043 |
.081 |
.271 |
.603 |
|
Suspect(s) Had Weapon |
-.304 |
.738 |
.462 |
.433 |
.510 |
|
Alcohol/Drug Impairment |
6.986*** |
1081.049 |
1.018 |
47.107 |
<.001 |
|
Fleeing |
1.573*** |
4.819 |
.339 |
21.564 |
<.001 |
|
Mild Aggression |
-.835 |
.434 |
.768 |
1.181 |
.277 |
|
Severe Aggression |
.421 |
1.524 |
.247 |
2.912 |
.088 |
Note: ***=p<.001,
**=p<.01, *=p<.05
---------------------
Discussion
Overall, race proved to be a significant
determinant factor through bivariate, cross-tabulation analysis. However, when
accounting for other factors, officer race proved less significant in
determining less-than-lethal weapon deployment. In fact, only Black/African American
officers compared to White officers in OC Spray was significant after taking
other variables into consideration. This finding mirrored the prior research
literature this analysis based the hypothesis on. As previously mentioned,
research revealed Black/African American officers used force equally across
suspect race compared to their White counterparts. This analysis found
Black/African American and Hispanic/Latino officers were less likely to deploy
less-than-lethal force. Race of the suspect also proved less significant
through a multivariate analysis than bivariate. Race of the suspect only
matters for Black/African American suspects and OC Spray deployment, and
Hispanic/Latino suspects for CED deployment. This finding also supported our
hypothesis that non-White suspects would have higher instances in use of force
incidents utilizing less-than-lethal weapon deployment. As prior data found,
people of color experience use of force at higher rates than their White
counterparts during law enforcement interactions.
The most consistent predictor of
less-than-lethal weapon deployment of officers was shown to be the suspect’s
behavior during the incident. Through both bivariate and multivariate analyses,
suspects who displayed aggression towards officers were more likely to have
less-than-lethal weapons used against them, particularly OC Spray. The
differences in suspect behavior and less-than-lethal weapon options could also
be related to higher levels of aggression meant a higher return of force from
officers. Deadly force was not significant through the multivariate analysis
for OC Spray. This finding could be that if a suspect appears to have intent to
seriously maim law enforcement or bystanders, officers return with lethal
force. OC Spray is not a lethal option. A similar conclusion can be made for
weapon possession of the suspect. This analysis only examined whether the
suspect possessed a weapon, not what that weapon was (i.e. club, firearm,
etc.). The degree of lethality of the weapon could have an influence on the type of force used by
officers. Although there was no prior literature linked to behavior, this
analysis’s data supported the hypothesis that aggressive behavior exhibited by
the suspect increased the likelihood of use of force. One would hope this is also
true in the practice of policing. The use of force officers used on suspects
should equal that which is deserved based on the suspects’ actions and not some
other arbitrary factor.
The significance of years of experience an
officer had also varied when considering other factors. Years of experience
proved significant for both OC Spray and CED use through bivariate analysis;
however, through a multivariate analysis, years of experience was significant
for neither OC Spray nor CED deployment. This finding showed other factors, not
officer years of experience, influenced whether less-than-lethal weapons were
utilized in use of force incidents. No research was linked to experience to
base a hypothesis on. A reason for the insignificant finding could be related
to the possibility that as officers get older, they stop working in the field
as much, which eliminates their opportunity to have interactions with suspects.
A similar conclusion can be made about suspect age and the influence of other
factors. A bivariate analysis showed age significant for both OC Spray and CED
use. The multiple regression analysis proved other factors were more important.
Even looking at the results of the cross-tabulation, the percentages of
less-than-lethal weapon engagement across age groups showed little variance.
This analysis’s findings of suspect age did not support the original
hypothesis. This analysis had previously hypothesized that younger suspects
would have higher instances of less-than-lethal force compared to older suspects;
as suspect age increased, the chances of being involved in use of force
decreased. The data supported age did not matter as much as other influential
characteristics surrounding use of force incidents.
Ultimately, there were several missing
data points missing to conduct a complete analysis. Since the study of origin
relied on police databases, not every police department tracked the same
information. This left many variables with only partial data. Some variables
this analysis examined had almost half the data missing because police
departments did not collect information on that variable for use of force
incidents. For example, call type was missing over 60% of its data, the reason
it was excluded in part of the regression analysis. Incomplete data proved
problematic when running multivariate regression analyses and some variables
had to be omitted. This proved especially challenging for CED use. The study
that gathered the data examined thirteen police departments. Only seven of
those police departments utilized CED in their use of force practices. That
alone cut the data almost in half for any analysis run on CED deployment.
There remain certain issues here for
further attentions. For instance, when running the regression about the First
Officer Years of Experience relates to CED used in incident, the statistical
significance of independent binary regression seems to contradict with all the
independent variables combined table. Though figures would be inherently
different within two different tables, their significance are poles apart, such
as b-coefficient, which refers to the coefficient or slope of the relationship
between IV (X) and DV (Y). The
independent table shows as -0.041, this negative value means that if you are in
first officer’s years of experience, you would be less likely to have the CED
device being used. For each unit decrease in the Officer’s years of Experience,
the CED used decreases by about 0.041 on average. However, in the combined
variates statistic table, b-coefficient shows as 0.027, which is a positive
value that means the first officer’s years of experience would be more likely
to have the CED device being used. Further, in the independent table, odds
ratio or Exp (B) is Close to 1 but below 1 means that being a case occurred in
such years of experience decreased the likelihood of being a Yes in CED used.
However, on the combined table, it shows to the contrary that E>1means being
a case occurred at the first officer's years of experience, would increase the
likelihood of being YES in DV. So, odds of being a YES on CED variable are
increased by 1.120 if you were in this IV. In addition, in independent table,
since b is close to 0, E is close to 1, if the sig. value is too high and the
Lower Bound < 1 and Upper Bound ＞
1, then we cannot reject the null. However, sig. value 0.033 ＜ Alpha 0.05, and Upper Bound is not ＞1, which is also ambiguous and
contradictory. It’s surprised that on combined coefficient table, it still
cannot prove statistically significant because its sig. value is too high that
is over the alpha, and both the lower and upper bound do not exceed 1. (though
b>0, E>1), we still cannot reject the null hypothesis of no difference
between the population mean. So, although we seem to get the same result of failing
to reject the null, the statistical issues remain and make the research
precision doubtful. Nonetheless, the same independent variable relates to OC
Spray DV was ran out without these contradictory issues. Another similar issues
are the interpretation of dummy variables in binary regression coefficient
table, there are many wrong numbers (all equal to 0.0 in Exp [B] & Lower
Bound) that don’t make any sense, including the First Suspect Age (B, S.E.,
Wald are all equal to 0.88) and the Maximum Suspect’s Behavior—Passive, the
S.E. is 18570.392, which is very
unusual.
Possible reasons may be related to be
considered, such as many data is masked, the research may not fully manage the
data well; the understanding of dummy variables as independent variables might be
varied between researchers, e.g. in call type group, I would run a regression
using the variable: Robbery, Burglary and Homicide together into the violent
crime category instead of excluding the Homicide into other group. If the
percentage of this kind of variables breaking down might be varied, then it
could be coded differently. Also, usually, we don’t assess the significance of
the constant or interpret it. In some
cases, constant does not make much sense because it is not possible to have an
X value---IV of zero in reality.
However, here in this case, it may have certain rare case for the first
officer absolutely has no years of experience, which is also ambiguous and hard
to interpret. Lastly, researchers should be cautious that even if their findings
are a statistical significance, it doesn’t mean these findings are practical
significance.
Much other research review prior to this
report had been limited to narrative literature review and theoretical
analysis, instead of empirical data analysis and review. A research article by
Colin Bolger in the American Journal of Criminal Justice attempted to
find results like the ones in this report to identify the most influential
predictors of police use of force decisions. Bolger encountered similar
limitations in research review, unable to use a substantial amount of surely
important research because of its lack in objectivity. His findings suggest
that the primary focus for future theories on use of force decisions should
regard the variables of the encounter instead of the theoretical perspective of
twenty-first century policing[16].
The authors of this report support Bolger’s conclusion. This type of research
and the analyses/conclusions henceforth can strengthen administrative policies
regarding use of force by focusing on variables in which force is used and
having evaluated evidence in prior research showing stricter policies reducing unnecessary
force.
References
Austin, G.J. (2019). 103-02 pg.46
Non-Lethal Use of Force. Rye Brook Police Department; New York State
Accreditation Standards. Retrieved from, Rye
Brook Police Department Manual (storage.googleapis.com)
Ba, B.; Grogger, J. (2018). The
Introduction of Tasers and Police Use of Force: Evidence from the Chicago
Police Department. National Bureau of Economic Research. DOI: 10.3386/w24202
Bolger,
C.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(1). Pp 466-492. DOI: 10.1007/s12103-014-9278-y
Brooks, S. (2019, October 03). CED (Taser). Retrieved April 27, 2021, from CED (Taser) | Norfolk Constabulary
Bulman, P. (2021) Police use of force: The
impact of less-lethal weapons and tactics. Retrieved April 27, 2021, from Police
Use of Force: The Impact of Less-Lethal Weapons and Tactics | National
Institute of Justice (ojp.gov)
Fowers, A.; Steckelberg, A.; Berkowitz, B.
(2020, June 05). A guide to the less-lethal weapons that law enforcement uses
against protesters. The Washington Post. Retrieved April 30, 2021, from The
less-lethal weapons law enforcement is using against protesters - The
Washington Post
Fryer Jr., R.G. (2018). An Empirical
Analysis of Racial Differences in Police Use of Force. National Bureau of
Economic Research. Retrieved from MergedFile
(nber.org) / Journal of Political Economy, 127(3).
Greaves, I.; Hunt, P. (2010). Less Lethal
Weapons. Responding to Terrorism. (First Edition). London: Churchill
Livingstone.
Hilmas, C.J. (2015). Riot Control Agents. Handbook
of Toxicology of Chemical Warfare Agents (Second Edition). Waltham:
Academic Press.
Law.Cornell.edu Admin. (2021). U.S.
Federal Regulation: Use of Deadly Force. Cornell Law School. Retrieved
April 29, 2021, from 10
CFR § 1047.7 - Use of deadly force. | CFR | US Law | LII / Legal Information
Institute (cornell.edu)
Levenson, E. (2021, April 06). Police
use-of-force instructor Says Derek CHAUVIN'S kneeling is not a Trained
restraint. CNN. Retrieved April 22, 2021, from George
Floyd: Police use-of-force instructor says Derek Chauvin's kneeling is not a
trained restraint - CNN
Matiash, C., & Rothman, L. (2016, March
03). Rodney King beating at 25: What happened in Los Angeles. Time Magazine.
Retrieved April 22, 2021, from Rodney King
Beating at 25: What Happened in Los Angeles | Time
Paoline III, E.A.; Gau, J.M.; Terrill, W.
(2016). Race and the Police Use of Force Encounter in the United States. The
British Journal of Criminology, 58(1). Pp 54-74. DOI: 10.1093/bjc/azw089
Simko-Bednarski, E. (2020, June 25).
Bipolar man with a SWORD dies after being tased by NYPD. CNN. Retrieved
April 30, 2021, from Bipolar
man with a sword dies after being tased by NYPD - CNN
Stephenson, C. (2015). Smart Policing
Initiative: Less Lethal Technologies for Law Enforcement. Department of Justice
(BJA). PowerPoint
Presentation (strategiesforpolicinginnovation.com)
Taylor, B., Koper, C. (2013, October 29) Evaluation of Less-Lethal
Technologies on Police Use of-Force Outcomes in 13 Sites in the United States,
1992-2007. U.S Department of Justice, Office of Justice Programs. Doi: 10.3886/ICPSR27561.v1 https://www.icpsr.umich.edu/web/ICPSR/studies/27561
Terrill, W.; Paoline III, E.A. (2016).
Police Use of Less Lethal Force: Does Administrative Policy Matter? Justice
Quarterly. DOI: 10.1080/07418825.2016.1147593
[1] Matiash, C., & Rothman, L. (2016, March 03).
Rodney King beating at 25: What happened in Los Angeles. Time Magazine.
Retrieved April 22, 2021, from Rodney King Beating at 25: What Happened in Los
Angeles | Time
[2] Levenson, E. (2021, April 06). Police use-of-force
instructor Says Derek CHAUVIN'S kneeling is not a Trained restraint. CNN.
Retrieved April 22, 2021, from George Floyd: Police use-of-force instructor says
Derek Chauvin's kneeling is not a trained restraint - CNN
[3] Fowers, A.; Steckelberg, A.;
Berkowitz, B. (2020, June 05). A guide to the less-lethal weapons that law
enforcement uses against protesters. The Washington Post. Retrieved
April 30, 2021, from The less-lethal weapons law
enforcement is using against protesters - The Washington Post
[4] Simko-Bednarski, E. (2020, June
25). Bipolar man with a SWORD dies after being tased by NYPD. CNN.
Retrieved April 30, 2021, from Bipolar man with a sword dies after
being tased by NYPD - CNN
[5] Bulman, P. (2021) Police use of
force: The impact of less-lethal weapons and tactics. Retrieved April 27, 2021,
from Police Use of Force: The Impact of
Less-Lethal Weapons and Tactics | National Institute of Justice (ojp.gov)
[6] Law.Cornell.edu Admin. (2021). U.S.
Federal Regulation: Use of Deadly Force. Cornell Law School. Retrieved
April 29, 2021, from 10 CFR § 1047.7 - Use of deadly force. | CFR | US Law
| LII / Legal Information Institute (cornell.edu)
[7] Austin, G.J. (2019). 103-02
pg.46 Non-Lethal Use of Force. Rye Brook Police Department; New York State
Accreditation Standards. Retrieved from, Rye Brook Police Department Manual
(storage.googleapis.com)
[8] Stephenson, C. (2015). Smart
Policing Initiative: Less Lethal Technologies for Law Enforcement. Department
of Justice (BJA). PowerPoint Presentation
(strategiesforpolicinginnovation.com)
[9] Brooks, S.
(2019, October 03). CED (Taser). Retrieved April 27, 2021, from CED
(Taser) | Norfolk Constabulary
[10] Greaves, I.; Hunt, P. (2010). Less
Lethal Weapons. Responding to Terrorism. (First Edition). London:
Churchill Livingstone.
[11] Hilmas, C.J. (2015). Riot Control
Agents. Handbook of Toxicology of Chemical Warfare Agents (Second
Edition). Waltham: Academic Press.
[12] Ba, B.; Grogger, J. (2018). The
Introduction of Tasers and Police Use of Force: Evidence from the Chicago
Police Department. National Bureau of Economic Research. DOI: 10.3386/w24202
[13] Terrill, W.; Paoline III, E.A.
(2016). Police Use of Less Lethal Force: Does Administrative Policy Matter? Justice
Quarterly. DOI: 10.1080/07418825.2016.1147593
[14] Fryer Jr., R.G. (2018). An
Empirical Analysis of Racial Differences in Police Use of Force. National
Bureau of Economic Research. Retrieved from MergedFile (nber.org) / Journal of Political Economy,
127(3)
[15] Paoline III, E.A.; Gau, J.M.;
Terrill, W. (2016). Race and the Police Use of Force Encounter in the United
States. The British Journal of Criminology, 58(1). Pp 54-74. DOI:
10.1093/bjc/azw089
[16] Bolger, C.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(1). Pp
466-492. DOI: 10.1007/s12103-014-9278-y
Comments
Post a Comment