Factoring Fatigue into Police Deadly Force Encounters: Decision-Making and Reaction Times

Abstract

Significant evidence exists demonstrating the negative impact of fatigue on human cognitive performance in such areas as decision making, reaction times, and memory. Law enforcement studies have shown that officers suffer from high levels of fatigue from lack of sleep, unusual shift schedules, and exorbitant hours awake; however, little empirical evidence exists directly relating the effects of fatigue to individual officer performance in police specific tasks, particularly performance in deadly force situations. The current study (N=53) examined effects of fatigue, including total time awake (TTA), shift work, hours slept, and subjective sleep quality, on officers’ decision making and reaction times when presented with simulated shoot/don’t shoot and ambiguous target paradigms. The author of this study hypothesized that fatigue would negatively impact officers’ accuracy of decision making and reaction times. The hypothesis was confirmed, in that many of the fatigue measures correlated significantly with decreases in decision making in the deadly force simulations and with increased reaction time. Specifically, poor sleep quality, greater TTA more days worked, and working night or swing shifts all decreased the accuracy of officers’ decision making, especially when officers were presented with no-shoot and ambiguous scenarios. Greater TTA, more days worked, and working swing shifts also increased officers’ reaction times during these deadly force simulations. Finally, the effects of fatigue also increased throughout each work day, with officers’ reaction times increasing consistently from their pre-shift assessment to their post-shift assessment. These findings have significant implications for police performance in deadly force encounters, training, and scheduling
Keywords: Sleep deprivation, sleep debt, fatigue, police, use of force, shift work, deadly force

Full text

Law Enforcement Executive Forum • 2015 • 15(1)44
A simple click of the mouse button while
searching through hundreds of YouTube
videos can provide single dimensional eyewit-
ness views of United States law enforcement
ofcers in a multitude of unusual and deadly
situations. The most intriguing ones revolve
around ofcers’ decision making during rap-
idly evolving, dynamic, and highly stressful
incidents. Unfortunately, some of these inci-
dents, about 0.2%, result in an ofcer’s most
powerful and devastating decision, the deci-
sion to use deadly force (Adams et al., 1999).
Also unfortunate is how society, through the
media, can sometimes misjudge these encoun-
ters based on limited information (Adams et al.,
1999; Johnson, 2007; Sharp & Hess, 2008).
Understanding police use of force in today’s
culture of unrelenting media access and per-
sonal video devices may require a paradigm
shift in how society looks at those who protect
and serve. As police are often held to scien-
tically deduced ideal human performance
standards that may be unattainable in real-life
encounters, they may be perceived as racist,
overly aggressive, or even murderous when
they fail to meet these standards (Johnson,
2007; Lewinski & Honig, 2008; Sharp & Hess,
2008). The idea of fallibility through human
performance error is rarely considered or
accepted when an ofcer uses lethal force.
It could be argued that social judgment
should and legal judgment must be derived
Factoring Fatigue Into Police Deadly Force
Encounters: Decision-Making and
Reaction Times
David M. Blake, MS, Blake Consulting and Training Group
Edward Cumella, PhD, Professor of Graduate Psychology, Kaplan University
Abstract
Signicant evidence exists demonstrating the negative impact of fatigue on human cognitive performance
in such areas as decision making, reaction times, and memory. Law enforcement studies have shown that
ofcers suffer from high levels of fatigue from lack of sleep, unusual shift schedules, and exorbitant hours
awake; however, little empirical evidence exists directly relating the effects of fatigue to individual ofcer
performance in police specic tasks, particularly performance in deadly force situations. The current study
(N = 53) examined effects of fatigue, including total time awake (TTA), shift work, hours slept, and sub-
jective sleep quality, on ofcers’ decision-making and reaction times when presented with simulated shoot/
don’t shoot and ambiguous target paradigms. The authors of this study hypothesized that fatigue would
negatively impact ofcers’ decision-making and reaction time accuracy. The hypothesis was conrmed in
that many of the fatigue measures correlated signicantly with decreases in decision making in the deadly
force simulations and with increased reaction time. Specically, poor sleep quality, greater TTA, more days
worked, and working night or swing shifts all decreased the accuracy of ofcers’ decision making, especially
when ofcers were presented with no-shoot and ambiguous scenarios. Greater TTA, more days worked, and
working swing shifts also increased ofcers’ reaction times during these deadly force simulations. Finally,
the effects of fatigue also increased throughout each work day, with ofcers’ reaction times increasing
consistently from their pre-shift assessment to their post-shift assessment. These ndings have signicant
implications for police performance in deadly force encounters, training, and scheduling.

Law Enforcement Executive Forum • 2015 • 15(1)45
from solid empirical evidence. This evidence
would do well to provide a full accounting of
all the human factors involved and the con-
text within which a specic situation exists,
along with close attention to the applicable
written legal codes.
To begin this accounting, a basic understand-
ing of the aspects involved in a use-of-force
incident and how they pertain to the perspec-
tive of the ofcer must be compiled. This jour-
ney may begin in the legal realm. Miller (n.d.),
Branch Chief within the Legal Division of the
Federal Law Enforcement Training Center
(FLETC), states that assessment of a police
ofcer’s use of force is generally based upon
the 1989 Supreme Court decision, Graham v.
Connor (1989). In brief, the decision provides
that an ofcer’s use of force in any Fourth
Amendment seizure (i.e., an arrest or deten-
tion) should be judged based on an objective
reasonableness standard. The Court went on
to dene the objective reasonableness stan-
dard as follows: “Would another similarly
experienced ofcer in a similar situation, uti-
lize a similar amount of force while under
split second timing restrictions and operating
in “tense, uncertain, and rapidly evolving cir-
cumstances?” (p. 1).
The Supreme Court in Graham provided
the legal backdrop for a fair and balanced
approach by which police use of force can
be judged. However, to truly understand
the event as it was experienced by the of-
cer involved, human factors relating to sub-
jective perspective must be introduced. Yet,
signicant uncertainty exists concerning the
human factors that ofcers may have experi-
enced, are able to testify about, or even know
to exist within the situation. The aspects of
rapidly evolving violent encounters from the
perspective of an involved police ofcer are
unique and have not been previously studied
with the detail necessary to fully understand
them. Cognitive psychology and the study of
human factors in use-of-force encounters are
just beginning to close the gap of understand-
ing (Honig & Lewinski, 2009).
Recent research provided insight into how
ofcers on the same scene might provide
differing accounts of an incident, how of-
cers vary in threat perception, and even how
they might justiably shoot a suspect in the
back (Blair et al., 2011; Lewinski & Honig,
2008; Lewinski & Hudson, 2003; Lewinski &
Redmann, 2009). Take for instance the case of
Randall Carr, who was shot and killed after
a deadly altercation with police ofcers. The
location of the fatal wound was questioned
because its placement in the buttocks meant
Carr was no longer a threat when shot by
ofcers. Ultimately, Dr. William Lewinski of
the Force Science Institute was able to ade-
quately explain the human dynamics behind
the incident, and the ofcers were exonerated
by a jury (Force Science Institute, 2005). His
breakthrough research into the dynamics of
police-involved shootings demonstrated the
uidity of a gunght and provided an of-
cer’s “stop shooting” reaction times, which
could result in a suspect justiably being
shot in the back (Lewinski, 2000). Without
Lewinski’s research into the human factors of
use-of-force encounters, the ofcers may have
been held accountable for what appeared to
be an unjustiable shooting when, in fact,
they were simply operating within the con-
nes of human performance.
One human factor that has been ignored for
too long in policing is fatigue. Shift work,
court appearances, special assignments,
and the long hours ofcers usually work
have been suggested as contributing factors
to human error (Vila, Kenney, Morrison,
& Reuland, 2000). The concept of fatigue
is closely linked to sleep deprivation, and
these labels are often used synonymously
(Samkoff & Jacques, 1991; Sundelin et al.,
2013). A depth of research exists in the area
of sleep deprivation (SD) as it applies to sev-
eral occupations, including policing, but its
application to specic police tasks has been
rather limited. Many of the other occupations
impact public safety and have been mandated
to have rest periods (Senjo, 2011). Policing,
which appears to have some of the poorest

Law Enforcement Executive Forum • 2015 • 15(1)46
working conditions in regard to sleep (Senjo,
2011; Vila et al., 2000), does not benet from
such mandates. In fact, little regard is given to
the many sleep disturbances ofcers experi-
ence. SD is often deemed just “part of the job”
(Vila et al., 2000).
Bonnet and Arand (1995) discussed in great
detail what constitutes adequate sleep in their
literature review. The rst prominent point
they make pertains to today’s society being
chronically sleep-deprived (sleep loss > 1 hour
nightly). They support an 8.5 hour standard
as being optimal and demonstrate that nightly
sleep lengths of 7.2 to 7.4 hours are decient.
The authors stated that chronic sleep depriva-
tion of less than 6.5 hours is potentially disas-
trous in regard to human performance.
A closer look should be taken at some nuances
of sleep deprivation because it may have dif-
ferent meanings depending on the type of
study involved. Following the denition sup-
ported by lead researchers (Dinges, Rogers, &
Baynard, n.d.; Durmer & Dinges, 2005; Lim &
Dinges, 2010), SD is simply a restriction of a
subject’s sleep to less than their usual amount
within any 24-hour period. SD can be as minor
as restricting a subject to 7 hours of sleep
nightly, which is the starting point for decien-
cies in cognitive performance (Dinges et al.,
n.d.).
The U.S. Department of Health and Human
Services (DHHS) (2012) discusses the need
to combat SD through maintaining regularly
scheduled sleep habits of 7 to 8 hours daily
for most adults. Studies show that SD leads to
problems in many areas of human function-
ing, the most notable being decits in decision
making, problem solving, attention, reaction
time (RT), and emotional control (Durmer
& Dinges, 2005; Rajaratnam et al., 2011; Vila
et al., 2000). Sadly, proof of the DHHS’s
pronouncements has been provided by the
National Highway Transportation Safety
Administration’s (NHTSA) (1996) report of
a yearly average of 56,000 trafc collisions
resulting in 1,550 fatalities occurring due to
driver fatigue. The NHTSA provides the main
characteristics of fatigued drivers as having
increased reaction times, attention decits,
and a decreased ability to process information.
Durmer and Dinges (2005) performed an
extensive meta-analysis of the consequences
of SD. Their review began by discussing the
many vehicle-related accidents which occur
as a result of fatigue. The research suggests an
analogy between fatigued driver performance
and that of alcohol impaired drivers. Studies
have shown that drivers who are awake for 17
to 19 hours operate a motor vehicle with sim-
ilar psychomotor skills to those with blood
alcohol content (BAC) between 0.05 and 0.1%,
with 0.08% being the typical legal denition
of driving while intoxicated (NHTSA, 2006).
Bryan Vila, Director of the Simulated
Hazardous Operational Task Laboratory at
the Washington State University Sleep and
Performance Research Center, and his col-
leagues (2000) have conducted studies on
police SD, which showed that 53% of U.S.
police ofcers receive less than the mean
amount of sleep needed per night. Study
results showed that 18% of ofcers experi-
enced fatigue and a lack of motivation, and
another 16% stated they had trouble simply
staying awake on the job. Performance issues
related to this study were noted in the areas of
reduced patience, diminished decision-mak-
ing capacity, decreased alertness, and slower
response times.
Neylan et al. (2002) conducted a study com-
paring subjective sleep quality in police of-
cers, examining the effects of trauma expo-
sure (critical incidents) and non-police routine
organizational stressors. The ndings showed
that although ofcers suffer from trau-
ma-related nightmares, the most signicant
aspect affecting sleep quality was based in
the routine stressors experienced within the
non-trauma-related work environment.
Senjo (2011) researched 15 Western state law
enforcement agencies in the U.S. The study

Law Enforcement Executive Forum • 2015 • 15(1)47
provided self-reported sleep needs of 7 to 9
hours a night by 70% of the responding of-
cers. However, two-thirds of the 70% reported
actual completed sleep ranging between 3 and
6 hours. Issues such as shift work, overtime,
secondary employment, and others were
listed as reasons for ofcers having insuf-
cient sleep.
Rajaratnam et al. (2011) conducted a critical
study of 4,957 police ofcers from across the
U.S. and Canada. The research involved both
online surveys and onsite interviews. Results
showed that 40% of those tested suffered from
at least one sleep disorder. Of those suffering
from a sleep disorder, 6.5% suffered from
moderate to severe insomnia, and 5.4% tested
positive for shift-work disorder. Results also
showed that those who suffered from sleep
disorders more often reported having made
administrative errors, falling asleep while
driving, and committing safety violations due
to fatigue.
Beyond the previously discussed fatigue
issues is the concept of chronic partial sleep
loss, often called sleep debt, which has also
been shown to affect alertness and perfor-
mance (Barger, Lockley, Rajaratnam, &
Landrigan, 2009). Sleep debt is often dis-
cussed in terms of its cumulative effect. Using
a simple example, cumulative sleep debt is
the total amount of time, typically measured
in hours, over a specied period in which the
required sleep was not achieved. Van Dongen,
Maislin, Mullington, and Dinges (2003) con-
ducted a study restricting the sleep of 48 par-
ticipants to either 6 hours or 4 hours over 14
days. Tests such as the Psychomotor Vigilance
Test (PVT; Dinges & Basner, 2011) and the
Stanford Sleepiness Scale (SSS; Hoddes,
Zarcone, Smythe, Phillips, & Dement, 1973)
were administered. The PVT measures alert-
ness through required sustained attention
while requiring quick reactions to random
stimuli; it has been deemed highly reliable
with test results comparable to real-world
behaviors (Dorrian, Rogers, & Dinges, n.d.).
The SSS is a subjective test with demonstrated
validity that can reliably determine levels of
sleepiness in an individual. The results of
the SSS have been found to correspond sig-
nicantly with performance on tasks related
to SD (Hoddes et al., 1973). Within this sleep
study, signicant differences existed in both
the 6- and 4-hour sleep groups in comparison
to the 8-hour group, indicating decits for
the 4- and 6-hour groups. Of importance for
this study was the result showing that sleep
restriction to 4 hours over 14 days resulted in
working memory and alertness levels equiva-
lent to those in persons who had not slept for
two days. Those in the 6-hour group showed
decits comparable to one day without sleep.
Thus, the empirical evidence from this study,
combined with the BAC comparisons, sug-
gest that persons who suffer from chronic
cumulative sleep debt could be functionally
equivalent to highly intoxicated individuals.
Another study concerning cumulative sleep
debt shows that even minor restrictions, such
as one hour per night, can cause performance
deciency (Belenky et al., 2003). Belenky et al.
(2003) conducted a study in which 66 partic-
ipants were sleep deprived at levels from 3
to 7 hours over 7 days. The study utilized the
PVT and SSS to measure sleepiness and per-
formance four times per day. Although the
7-hour group did not report having increased
sleepiness (SSS), they did show signicant
decreases in RT performance within PVT
results.
Couyoumdjian et al. (2009) discussed real-
world decision making by considering the
circumstantial uniqueness in which decisions
often occur. Some circumstances, such as
those of policing, require innovative think-
ing, distraction avoidance, ignoring irrele-
vant stimuli, and following unfolding events,
all of which are negatively impacted by SD.
These executive-level functions were assessed
through a task switching stimulus test. The
results indicated that one night of total SD
negatively impacted the participants’ ability
to shift between two different cognitive tasks.
This information is signicant in that ofcers

Law Enforcement Executive Forum • 2015 • 15(1)48
are often required to switch between tasks,
especially in use-of-force situations.
From the available literature, there is little doubt
that police ofcers work within a SD occupa-
tion and are ultimately exposed to SD at levels
which have adverse effects on human perfor-
mance (Alhola & Polo-Kantola, 2007; Antal,
1975; Barger et al., 2009; Couyoumdjian et al.,
2009; Durmer & Dinges, 2005; Edwards &
Waterhouse, 2009; Lewinski & Honig, 2008;
NHTSA, 1996; O’Brien et al., 2012; Rajaratnam
et al., 2011; Senjo, 2011; Vila et al., 2000).
Nevertheless, the specicity of research into
fatigue and police performance with the use
of rearms (i.e., deadly force) is lacking. A
few studies are suggestive, however.
Edwards and Waterhouse (2009) conducted
an experiment showing the effects of SD on
the ability to throw darts. This study provided
intriguing results because of its simplicity and
the demonstrated effects of relatively little
SD. Sixty participants were deprived of four
hours sleep on just one night and then asked to
throw darts at a dart board. Deciencies were
noted in accuracy and reliability. These de-
ciencies increased as the subjects were tested
over a span of several hours after awakening.
Antal (1975) conducted a study of circadian
rhythm disruption and its effects upon com-
petitive shooters. Although his data are rudi-
mentary and provide no specic hours of SD
or levels of fatigue, he shows a correlation
between the interruption of natural sleep
cycles and accuracy with rearms. His study
reported that shooters with SD suffered from
an inability to concentrate, complaints of
fatigue, and a lack of vitality. Another study
involving SD of 22 hours on a range of shoot-
ing skills was completed with a group of 20
military subjects. The study revolved around
the effects of caffeine and performance, but
it provided solid SD data in several areas.
This military study concluded that SD causes
decits in RT to engagement and accuracy of
shot placement (Tikuisis, Keefe, McLellan, &
Kamimori, 2004).
The effect of SD on human beings operating in
various environments appears clear. A stack
of empirical evidence shows lack of sleep
causes poor attention, errors in judgment and
decision making, and a slowing of reaction
times. There also exists a rather convincing
amount of evidence, indicating that SD exists
at signicant levels within law enforcement.
The last dot to connect is between the stated
effects of SD and police use-of-force incidents.
Bill Lewinski of the Force Science Institute
has collected scientic studies and discus-
sion papers suggesting such a link. Based on
reviews, Lewinski and Honig (2008) summa-
rize many of the human cognitive dynamics
of police use-of-force encounters and point to
attention, perception, decision making, pat-
tern recognition, and action/reaction time as
having much to do with ofcers’ successfully
overcoming violent encounters and making
correct decisions. Importantly, these very
cognitive functions are negatively affected by
SD (Alhola & Polo-Kantola, 2007; Barger et al.,
2009; Couyoumdjian et al., 2009; Van Dongen
et al., 2003).
Thus, a conuence of research in the areas of
SD, SD in policing, and use of force provides a
glimpse of the potential deadly consequences
created by a combination of SD-driven fac-
tors. SD is prevalent in policing, yet the very
cognitive functions that are so necessary for
attending to and ultimately making the cor-
rect decision in use-of-force environments are
decreased by SD. The need to more carefully
examine the association between use-of-force
decision making and SD is therefore surely
necessary. Similar research specic to other
professions has revealed serious deciencies,
prompting laws and regulations governing
these elds (NHTSA, 1996).
The literature review provided empirical evi-
dence to support the following hypothesis,
which is addressed in the current research
investigation. To begin, police ofcers’ job
demands likely create an environment of SD
through several means, most prominently the
disruption of the circadian rhythm resulting

Law Enforcement Executive Forum • 2015 • 15(1)49
from shift work, a continuously accumulating
sleep decit, and excessive total hours awake.
Overall cognitive abilities in the areas of infor-
mation processing, decision making, reaction
time, and attention are negatively affected
by SD. Thus, it is hypothesized that these
effects of SD will have a negative impact on
ofcers’ decision-making capabilities during
shoot/don’t shoot scenarios. Specically,
reaction time has been found to be negatively
impacted by SD and is expected to be affected
in this study through slower reaction times to
shoot/don’t shoot scenarios among ofcers
experiencing SD, with ofcers’ ability to react
quickly to perceived threats and to correctly
identify a shoot target being decreased by SD.
Methods
Participants
Participants were police ofcers from several
national police departments (N = 53): 50 men
and 3 women, ages 25 to 54, M = 40 (SD = 7.8)
years. Due to the specialized nature of this
study, participants were all experienced police
ofcers having completed basic police and
recurring inservice training concerning police
use of deadly force. Participating ofcers
were sampled from all shifts: Day Shift, n
= 17; Swing Shift, n = 21; and Night Shift,
n = 15. Following other studies on SD, partici-
pants were aware of the purpose of the study
as it has been determined such studies are
valid under these conditions (e.g., Edwards
& Waterhouse, 2009; Tikuisis et al., 2004;
Williamson & Feyer, 2000).
An additional and separate sample of police
ofcers from across the country (N = 277) com-
pleted a 10-question online Fatigue Survey.
The participants were gathered from elec-
tronic posts in police-specic online groups
such as the California Association of Force
Instructors, Law Enforcement Professionals,
and the Ofcer Involved Use of Force Group,
which are all hosted by LinkedIn (www.
linkedin.com), a professional networking
website. To protect the anonymity of these
ofcers and encourage their honesty about
a potentially difcult subject—the effects of
fatigue on their own police work—absolutely
no descriptive information about the partici-
pants was collected.
Procedure
An online electronic platform was created
based on previous studies and validated
measures. The online platform was named
the Thesis Computer Program (TCP) and has
built-in parameters to compensate for the lack
of an in-laboratory testing environment. These
parameters ensured participants logged in as
required, completed all prescribed tests, and
completed those tests within specied limits.
The TCP’s design favored validity over user
friendliness to provide the best outside of a
laboratory results.
Prior to engaging in the study, participants
were provided with an overview of the pur-
pose of the study and introduced to its meth-
odology. An online introduction to the TCP
followed in which in-depth instructions and
screen shots were provided. All information
provided prior to testing remained available
to participants throughout the duration of the
study. Additionally, participants had contin-
ued access to the facilitator to answer ques-
tions or resolve issues.
Participants were provided a URL which
allowed them access to the TCP online. Upon
entering the site, participants were required
to create an account using a typical password
and username security combination. The reg-
istration process included a request for cer-
tain non-identifying personal information
such as age, gender, shifts worked, and days
in the work week. Although all areas listed
were self-explanatory, the wide range of shift
denitions across law enforcement required
independent denitions of day, swing, and
night shifts. Day Shift was dened as most
duty hours during daylight. Swing Shift was
dened as half duty hours in daylight and

Law Enforcement Executive Forum • 2015 • 15(1)50
half during darkness. Night Shift was dened
as most duty hours during darkness.
During the registration process, the consent
form was displayed on the page, and partici-
pants were unable to register without selecting
a “consent/register” button, allowing them
to move forward. Participants received a val-
idation e-mail to the address they provided
and were required to activate their accounts
through a link provided to that e-mail account.
Activating the account allowed participants
to have access to the task sections of the TCP.
Once registered, participants were asked to
log into the TCP at the beginning of their duty
week. They were required to log-in as close
to the beginning and end of each individual
work shift as possible. Strict adherence to the
research design was required, and partici-
pants were aware of parameters invalidating
any improper actions or inputs by the user.
Participants understood that a failure to com-
plete all tests on all log-ins would result in
nullifying that day’s data.
The TCP itself adhered to a very strict set of
guidelines which allowed users little leeway
to operate outside its design. Participants
were guided step by step through the online
testing platform by displayed instructions
as well as automated movement to the next
task after the former task was completed.
The TCP did not allow for log-ins outside
of certain parameters, such as a mandated 8
hours between shifts or a requirement to log
in for post-shift tasks within 24 hours of the
pre-shift log-in. Participants were unable to
log-in for post-shift task completion without
having rst signed in for pre-shift completion.
Sleep Diary
The rst task required for each log-in to the
TCP was the completion of a sleep diary.
Participants completed a sleep diary for the
three days prior and all four days of the test-
ing cycle. The sleep diary required the partic-
ipants to enter the time they awoke each day,
the total hours of sleep prior to waking that
day, and their opinion about the quality of
their sleep.
Due to the subjective nature of asking par-
ticipants whether or not they had a good or
bad night’s sleep, the TCP dened each cat-
egory. A good sleep cycle was dened as an
“uninterrupted sleep cycle while awakening
well rested.” A bad sleep cycle was dened as
an “interrupted sleep cycle while awakening
poorly rested.” These denitions appeared on
each log-in to ensure consistency and valid-
ity. The TCP sleep diary task provided drop-
down menus or restricted data entry points
(e.g., HH:MM) for each required response,
ensuring only the correct type of answer was
provided.
The sleep diary information was requested for
several reasons. The rst is its ability to pro-
vide a static picture of changes in sleep pat-
terns between duty days and non-duty days.
Additionally, time awake, hours slept, and
sleep quality are all key points of correlation
to the performance tasks within the study,
and they allow a determination of whether or
not these factors have any effect on reaction
times or decision making (Dinges et al., 1997;
Lim & Dinges, 2010).
Epworth Sleepiness Scale (ESS)
Participants were asked to complete the ESS
daily during both pre- and post-shift log-ins
to the TCP. Johns (2000) studied various sleep-
iness scales and demonstrated the ESS as the
most valid and reliable test available for mea-
suring the appropriate amount of sleep. This
self-administered questionnaire (ESS) pro-
vides empirical evidence of whether or not
test subjects are fatigued. The instructions for
the ESS are very specic, yet simple, requiring
participants to subjectively rate their poten-
tial for “dozing” under a series of eight con-
ditions. These standardized instructions were
provided in two places within the TCP as well
as reprinted on the TCP’s ESS data input page.
Additionally, input selections were limited by

Law Enforcement Executive Forum • 2015 • 15(1)51
drop-down menus to the standardized ESS
responses. The drop-down menus were an
additional method of ensuring validity in the
responses provided.
Psychomotor Vigilance Task (PVT)
The PVT has been mentioned often as a prev-
alent and simple testing measure to deter-
mine the effects of sleep loss upon reaction
speed and lapses (Alhola & Polo-Kantola,
2007; Dinges & Basner, 2012). Gartenberg
and Parasuraman (2010) conducted a study
testing the validity of a shortened “reac-
tion test” using the iPhone/iPad platform,
with the application titled Mind Metrics. The
study provided evidence of validity in using
this 3-minute form of the PVT. Additionally,
other shorter duration PVT tests (i.e., 3 to 5
minutes) have demonstrated validity (Dinges
& Basner, 2012). The TCP included a version
of the 3-minute PVT due to the amount of
required testing sessions and the total time
required per log-in session.
Participants were required to complete the
PVT daily during both the pre- and post-shift
log-ins. Specic instructions were provided
to participants to ensure strict adherence to
the PVT methods. These instructions were
provided within the computer program and
had to be viewed before each test began. To
ensure validity, participants were required
to use their dominant hand middle nger
to perform the test. The hand was required
to be static, positioned just below the key-
board, helping to standardize the distance
the middle nger would be from the data
input device, which was the space bar. This
method enhanced both within-subject and
between-subject validity. In addition to the
physical restrictions asked of the participants,
the PVT had restrictions on acceptable reac-
tion time results. The PVT does not accept RT
scores faster than 100 ms or slower than 1,500
ms to further ensure validity of captured
data. The reaction time parameters are similar
to other studies measuring RT under similar
circumstances (Adam, Bays, & Husain, 2011;
Baumann & DeSteno, 2010; Correll et al.,
2007; Dinges & Basner, 2011; Gartenberg &
Parasuraman, 2010; Lewinski & Hudson,
2003).
Shoot/Don’t Shoot Situations (SDS)
Correll et al. (2007) tested police use of force
decision making on several occasions through
the use of SDS computer analysis. The stud-
ies used a computer-based simulation dis-
playing photographs of armed or unarmed
subjects in various settings. The photographs
remained on the screen for a short period
of time, between 500 and 850 ms, and were
intended to elicit SDS decisions from the sub-
jects. Points were added and subtracted based
upon the decisions made by the subjects.
Using similar methodology, a SDS process
was incorporated into the TCP. The process,
similar to Correll et al. (2007), records data
from SDS displays and participant inputs
regarding reaction time and decision making
in response to the stimulus photo. Due to the
different nature of measurements within this
project, a minor change in the SDS platform
from Correll et al. (2007) was required as fol-
lows. Police ofcers spend a sizable portion
of their day involved in low stress tasks, but
when necessary, they are required to switch to
aroused status in reaction to threatening stim-
uli (e.g., from report writing to a radio call of
an in-progress crime). Likewise, police shoot-
ing situations are often unexpected and occur
in conuence with any number of other low
to highly arousing daily duties. To replicate
a realistic switch between arousal states, or at
least provide for a realistic cognitive distrac-
tion, the TCP displayed a simple math equa-
tion between SDS stimuli. The math ques-
tions required the participants to respond by
striking the spacebar for correct answers. The
math problem remained on the screen for 2 to
10 seconds prior to the display of each new
SDS stimulus. This intervening math event
was not present in Correll et al. (2007).

Law Enforcement Executive Forum • 2015 • 15(1)52
Per log-in, participants in the current study
viewed 12 of 62 randomized and encoded
SDS photographs upon a computer screen:
six shoot, three no-shoot, and three ambiguous
scenarios. SDS decisions were made through
standard keyboard input used for gaming:
A = shoot, L = don’t shoot. Prior to inclusion,
the photographs in this study were reviewed
by a panel of tenured police use-of-force
instructors. Photographs which received any-
thing other than full agreement by the panel
were deemed ambiguous.
Thus, all SDS pho-
tographs used have 100% inter-rater reliability
by tenured police use-of-force instructors, repre-
senting denitive SDS situations or ambiguous
situations.
Participants were required to complete the
TCP daily during both pre- and post-shift
log ins. To avoid practice effects that could
degrade the validity of the testing process,
two procedural actions were put in place. The
rst was a randomization of the SDS photo-
graphs as to where each appeared during
each session. The randomization of the SDS
scenarios should ensure a lack of familiarity
with each stimulus. The second method of
avoiding practice effects is the sheer number
of scenario photographs, which were greater
than 60. A random viewing of 12 of 62 pho-
tographs over just four days should ensure a
lack of familiarity with each photograph as
no photograph was likely to appear several
times for each ofcer across the testing days.
Specic instructions were provided to partic-
ipants to ensure strict adherence to the TCP’s
parameters. These instructions were provided
within the computer program and had to be
viewed before each test began. Participants
were required to place both hands below the
keyboard in a specic manner while having
their “point” ngers hovering above the A
and L keys. In addition to the physical restric-
tions asked of the participants, the TCP con-
tained restrictions on acceptable reaction time
results. The reaction time parameters selected
were similar to those used in other studies
measuring RT under similar circumstances
(Adam et al., 2011; Baumann & DeSteno, 2010;
Correll et al., 2007; Dinges & Basner, 2011;
Gartenberg & Parasuraman, 2010; Lewinski &
Hudson, 2003).
Fatigue Survey
A survey created on the SurveyMonkey web-
site contained 10 questions. Table 1 lists the
questions, all based on self-report, related to
sleep, performance, and agency oversight.
The answers were limited to “Yes” or “No.”
The purpose was to assess a separate sample
of police ofcers, untainted by their experi-
ences completing the TCP, and obtain their
personal experiences with and views of the
effects of fatigue on their police performance.
Data Handling and Statistical Treatment
Participants were directed to the TCP web-
site to complete their assessments. The TCP
options were set properly to ensure none of
the participants’ names, police agency names,
and IP addresses were collected. All results
were presented in aggregate form to further
protect subjects’ identities and condential-
ity of information. Data were only accessible
through the online TCP system using a strong
password known only to the researcher.
Once the data collection was completed, data
were downloaded into Microsoft Excel and
then SPSS, stored only on the researcher and
advisor’s laptop computers, and deleted from
the online survey system. The SPSS database
used for data analysis was accessible only by
using a strong password known only to the
researcher and thesis advisor. Neither dataset
contained any coded identiers and, as such,
both are completely anonymous.
The thesis chair and the student researcher
had access to the downloaded SPSS data. The
data were stored on the two computers owned
by these individuals. The data resided in sep-
arate Windows folders on each computer,
segregated from unrelated les. The two com-
puters were locked by strong Windows pass-
words known only to the computer owners.

Law Enforcement Executive Forum • 2015 • 15(1)53
The data were retained on these computer sys-
tems for the duration of the research, and, fol-
lowing completion of the research, they were
retained on the researcher’s computer for a
minimum of ve years along with related les
in case questions arise about the analyses. The
dataset and related les will be transferred to
any future computer owned by the researcher
until the ve years have expired. Throughout
the study and subsequent ve years, the
researcher will implement a weekly backup
plan wherein the dataset and related les are
backed up using a secure online data backup
system. After the ve years, the researcher
will destroy the SPSS data le using then-cur-
rent Department of Defense data destruction
standards. An affordable technique, such as
encryption, will likely be chosen.
The various measures were scored according
to published norms. Then, the several inde-
pendent variables, which were measures of
fatigue, were correlated with the outcomes of
the SDS scenarios—scenario by scenario and
in the aggregate. Patterns of correlations were
detected by extracting signicant correlations
from the correlation matrix and presenting
such in tabular format. Because the direc-
tion of each correlation was predicted by the
hypotheses in the study, alpha levels were
one-tailed, set at p < 0.10 for signicance.
Results
In response to requests for participants printed
in the Police One Magazine and the Force Science
Institute Newsletter, 53 subjects completed the
study. It is not possible to know how many
subjects actually saw the research announce-
ment in these two venues; as such, a response
rate cannot be calculated for this study. To
protect subjects’ anonymity, minimal socio-
demographic data were collected; it appears
in Table 2. The mean subject age was 40; most
subjects were men. Table 2 also reveals that
subjects were fairly evenly distributed among
the three typical shifts worked in policing:
Days, n = 17; Swings, n = 21; Nights, n = 15.
As expected, participants slept more hours
on off-duty days than on-duty days: M = 6.8
hours vs. 6.4 hours. Participants were awake
between 15 and 17 hours at the completion of
each duty day (see Table 2 for details).
Table 1. Fatigue Survey Results
Measure % Yes
1. I believe shift work interferes with my ability to achieve a reasonably good
night of rest.
73.5
2. I have different sleep habits when I am not working as opposed to during
my work cycle.
82.2
3. I sleep much better on my days off as opposed to during my work cycle. 67.9
4. I believe lack of sleep has been the cause of a mistake or error I have made
while working.
68.5
5. I believe I perform better with more sleep. 92.7
6. I require about 8 hours of sleep to perform my best. 55.7
7. I believe I can perform adequately when required regardless of how many
hours I am awake.
41.4
8. I believe police departments should formally explore the impact of sleep
deprivation on ofcer performance.
94.5
9. I don’t want to explore aspects of sleep deprivation in police work because I
am concerned about a change in schedule or limitations on overtime.
12.0
10. I believe the law enforcement career eld (in general) does not adequately
concern itself with safety issues concerning sleep deprivation.
91.6

Law Enforcement Executive Forum • 2015 • 15(1)54
Instrument Validity
Correlations for Day 1 were computed to deter-
mine instrument validity. A sizable number
of signicant correlations occurred in the
predicted direction (see Table 3). Those cor-
relations were moderate for SQ and ESS,
moderate to strong for PVT and SDS RTs, and
strong for ESS and SDS RTs. Aggregate means
for the ESS and PVT over the course of the
study also showed movement in the predicted
direction. Table 3 demonstrates that subjec-
tive reporting of fatigue increased from pre-
shift to post-shift: ESS pre-shift, M = 6 (4.83),
and post-shift, M = 11 (6.27). Likewise, RT
increased from pre-shift to post-shift: PVT
pre-shift, M = 414 (63) ms, and post-shift,
M = 461 (87) ms. Table 3 includes data show-
ing daily increases in PVT RT on all but one
(Day 4 post-shift) for both pre- and post-shift.
These results suggest good predictive validity
for the TCP instrument.
Decision Making (DM)
Table 4 displays the coefcients of all signif-
icant DM (i.e., shoot, don’t shoot, or ambig-
uous) correlations and the number of signif-
icant correlations occurring in the predicted
direction for each independent variable and
the several DM outcome variables. Subjective
reports of sleep quality (SQ) yielded 20 sig-
nicant correlations in the direction of pre-
diction over the course of four days. Days 1
and 4 provided the strongest correlations
(> 0.41), with Days 2 and 3 providing mod-
erate correlations (0.26 to 0.40). Total time
awake (TTA) yielded 14 signicant correla-
tions in the direction of prediction on Days 3
and 4, with Day 3 providing the strongest sig-
nicant correlations (e.g., 0.727).
Table 5 displays the type of signicant DM
correlations (i.e., shoot, don’t shoot, or ambig-
uous). TTA and SQ produced six signicant
results in the shoot scenarios. TTA and SQ
produced 19 signicant no shoot results. TTA
and SQ produced 12 signicant results among
the ambiguous scenarios.
Reaction Times (RT)
Table 6 displays the coefcients of all signicant
RT correlations and the number of signicant
correlations occurring in the direction of
prediction. TTA in relation to RT yielded nine
signicant correlations in the direction of
Table 2. Background Characteristics of Participants (N = 53)
Measure Response
Shift 17 Day shift
21 Swing shift
15 Night shift
Age
Mean 40 (7.83)
Gender
Male 50
Female 3
Mean sleep hours per night
Off-duty 6.8 (1.76) hours
On-duty 6.4 (1.49) hours
Mean total time awake at posttest
Day 1 17 hours
Day 2 16 hours
Day 3 16 hours
Day 4 15 hours

Law Enforcement Executive Forum • 2015 • 15(1)55
Table 3. Correlations Suggesting Instrument Validity
Measure Day 1
SQ & ESS
Pre-shift 0.363
Post-shift 0.369
SQ & Shoot Response (Aggregate day 1)
Pre1NoShoot 0.311
Post1NoShoot 0.408
PVT & SDS RT Times
PrePVT/PreRT6 0.338
PrePVT/PreRT12 0.436
PostPVT/PostRT2 0.397
PostPVT/PostRT3 0.361
ESS & SDS RT Times
PreESS/PreRT4 -0.318
PreESS/PreRT8 -0.282
PostESS/PostRT9 0.424
Mean Psychomotor Vigilance Task (PVT)
Pre-shift
Day 1 411 ms
Day 2 410 ms
Day 3 436 ms
Day 4 434 ms
Post-shift
Day 1 437 ms
Day 2 484 ms
Day 3 486 ms
Day 4 467 ms
Pre-shift (aggregate) 414 (63) ms
Post-shift (aggregate) 461 (87) ms
Mean Epworth Sleepiness Scale (ESS)
Pre-shift (aggregate) 6 (4.83)
Post-shift (aggregate) 11 (6.27)
Table 4. Correlations of Participants’ Reaction Times with Independent Variables
Measure Day 1 Day 2 Day 3 Day 4
Total time awake -0.039 0.449 0.009 0.643
# Signicant correlations in predicted direction 0 5 1 3
Shift -0.169 -- -0.034 0.602
0 1 2
Days worked 0.063 0.010 -- 0.557
1 1

Law Enforcement Executive Forum • 2015 • 15(1)56
prediction. Days 2 and 4 produced strong
correlations (> 0.41), but Day 3 produced weak
correlations (< 0.20). The work shift assigned
showed strong correlations on Day 4, with
three signicant correlations occurring in the
direction of prediction. The total days worked
also had one signicant correlation moving in
the predicted direction on Day 4. Both work
shift and days worked correlations were
strong (> 0.41).
Table 7 displays the mean RT for the SDS with
all RTs moving in the direction of prediction.
SDS RT increased from pre- to post-shift on
Day 1: for SDS pre-shift, M = 719 (29) ms,
and post-shift, M = 767 (30) ms. SDS RT
increased from pre- to post-shift on Day 2:
for SDS pre-shift, M = 726 (21) ms, and post-
shift, M = 738 (33) ms. SDS RT increased
from pre- to post-shift on Day 3: for SDS
pre-shift, M = 705 (28) ms, and post-shift,
M = 746 (37) ms. SDS RT increased from
pre- to post-shift on Day 4: for SDS pre-shift,
M = 683 (20) ms, and post-shift, M = 729 (38) ms.
Aggregate mean RT for the SDS from Day 1 to
Day 4 increased between pre- and post-shift:
Table 5. Participants’ Signicant Decision-Making Types
Measure Signicant Correlations
TTA & Shoot response 3
SQ & Shoot response 3
Total 6
TTA & No shoot response 6
SQ & No shoot response 13
Total 19
TTA & Ambiguous response 8
SQ & Ambiguous response 4
Total 12
Table 6. Correlations of Participants’ Decision Making with Independent Variables
Measure Day 1 Day 2 Day 3 Day 4
Sleep quality/Mean correlation 0.440 0.383 0.276 0.697
# Signicant correlations in predicted direction 5 3 4 8
Total time awake/Mean correlation -0.846 -0.067 0.727 0.342
# Signicant correlations in predicted direction 1 2 7 7
Table 7. Aggregate Mean SDS Reaction Times
Measure Day 1 Day 2 Day 3 Day 4
Pre-shift SDS RT 719 ms 726 ms 705 ms 683 ms
SD 29.04 21.15 27.50 30.47
Post-shift SDS RT 767 ms 738 ms 746 ms 729 ms
SD 29.68 32.79 37.48 37.86
Total combined SDS RT Mean
Pre-shift 709 ms
SD 19.93
Post-shift 745 ms
SD 16.22

Law Enforcement Executive Forum • 2015 • 15(1)57
scenarios. Greater TTA, more days worked,
and working night or swing shifts also
increased ofcers’ reaction times during these
deadly force simulations. Finally, the effects of
fatigue also increased throughout each work
day, with ofcers’ reaction times increasing
consistently from their pre-shift assessment
to their post-shift assessment.
The body of scientic literature regarding
standard sleep requirements, sleep depriva-
tion, and cumulative sleep debt, along with
the effects of these factors on performance, is
large and continues to grow. Time and again,
the primary nding within the literature
was the statistically signicant relationship
between sleep deprivation and performance
in that sleep deciency leads to performance
deciency. The law enforcement eld is
aware of the decits from sleep deprivation,
but never before to the knowledge of this
researcher has a sleep-related study been so
directed to law enforcement’s most crucial
element—the application of deadly force.
The starting point of the discussion revolves
around the amount of sleep deprivation
experienced by the participants in this study.
Participants were not requested to change
sleep patterns or restrict sleep as is often the
case in sleep-related studies. The current
study simply looked at the police ofcers’ real
life data and compared their reported experi-
ences and assessment results to ndings from
the existing literature. Therein lay empiri-
cal evidence for a general determination of
fatigue’s impact on, and even expectation of,
poorer performance during crisis situations,
which could potentially involve deadly force.
The total hours slept data included days off-
duty as well as days on-duty to determine what,
if any, change occurred. The results showed
that participants had a negative mean change
of 20 minutes between off-duty and on-duty
sleep. In this light, it is important to note that
the literature review provided scientic evi-
dence that even minor sleep loss can cause
deciencies in performance (Belenky et al.,
for aggregate pre-shift, M = 709 (20) ms, and
post-shift, M = 745 (16) ms.
Fatigue Survey
Table 1 presents the results of the fatigue
survey. Most of the respondents (74%)
believed that shift work interferes with their
ability to achieve a good night’s rest. Most
said they had different sleep habits on and off
duty (82%), with 68% stating they slept better
on their days off. The vast majority said they
perform better with more sleep (93%), and 69%
of respondents pointed to lack of sleep as a
causal factor in one or more mistakes or errors
which they had made while working. About
half of the respondents said they require 8
hours of sleep for optimal performance, with
a minority, 41%, believing they can perform
adequately regardless of how many hours
they are awake. Almost all believed that the
law enforcement career eld does not ade-
quately concern itself with safety issues aris-
ing from sleep deprivation (92%). Likewise,
95% of respondents stated that police depart-
ments need to formally explore the impact
of sleep deprivation on ofcer performance.
A mere 12% of respondents did not want to
explore sleep deprivation research within law
enforcement due to concerns about changes
in scheduling or limitations on overtime.
Discussion
The authors of this study hypothesized that
SD would negatively impact ofcers’ accu-
racy of decision making during SDS scenar-
ios as well as their reaction times in such sce-
narios. This hypothesis appears to be amply
conrmed by the results of the present study
because many of the measures of fatigue cor-
related strongly with decreases in decision
making in the deadly force simulations and
with increases in reaction time. Specically,
poor sleep quality, greater TTA, more days
worked, and working night or swing shifts
all decreased the accuracy of ofcers’ deci-
sion making, especially when ofcers were
presented with no-shoot and ambiguous

Law Enforcement Executive Forum • 2015 • 15(1)58
The well-validated PVT was also utilized,
providing RT results for all days of the study
at the beginning and end of each duty day.
The results indicate that RT increased in the
direction of prediction over the duration of
the study: pre-shift PVT, M = 414 ms, but
post-shift PVT, M = 461 ms, an 11% increase
in RT. This is further empirical evidence of
increased fatigue and a coinciding perfor-
mance decrease measured by RT. Coinciding
with the ESS and providing validation to the
present method, this PVT increase from Day 1
pre-shift to Day 3 post-shift was the most sig-
nicant change in the direction of prediction.
Deciencies remained on Day 4 for both val-
idated tests but leveled off as expected, with
no increased sleep restriction, similar to what
others have documented (Banks & Dinges,
2007).
Deadly Force and Reaction Times
Based on the results from the PVT, it is clear
that RT was affected by increasing levels of
fatigue. However, a corollary question is
whether or not those RT deciencies trans-
lated to the SDS task. Many signicant cor-
relations emerged in the predicted direc-
tion regarding TTA and SDS RT scores. The
shift and the number of days worked also
negatively affected SDS RT scores. The cor-
relations between SDS RT and TTA, days
worked, and shift worked strongly suggest
that fatigue directly increases an ofcer’s
reaction time to deadly force decisions, at
least in the simulated environment of the
present study.
Deadly Force and Decision Making
The data clearly show that subjective SQ
and TTA had a great impact on the ofcers’
ability to decide correctly between the three
SDS possibilities. It should be noted that the
ambiguous responses were coded so that
only a total lack of action impacted the par-
ticipants’ results negatively. Although not
directly related to the hypothesis of this
study, it is important to point out that deci-
sion making on the SDS in fatigue-related
2003; Bonnet & Arand, 1995). That appears to
have been the case with the participants in the
present study.
Those same scientic articles point to both
sleep deprivation and cumulative sleep debt
as indicators of fatigue and performance de-
ciencies. The current participants not only
received less sleep than recommended, but
they also experienced a rather severe cumu-
lative sleep debt over seven days: 1.5 hours
daily x 7 days = 10.5 hours cumulative sleep
debt. Here again, the literature review noted
the negative effects of cumulative sleep
debt on performance (Barger et al., 2009;
Couyoumdjian et al., 2009; Hoddes et al.,
1973; Van Dongen et al., 2003), which was evi-
dent in the present study in that fatigue mea-
sures correlated with poor performance and
increased reaction times more during the later
days of the ofcers’ work weeks.
The TTA of the participants must also be
looked at to determine whether or not fatigue
is likely to be present. Study participants
reported TTA on the cusp of the hours scien-
tically shown to provide performance de-
ciencies, equivalent to a 0.05% blood alcohol
content (NHTSA, 2006; Senjo & Heward,
2007), with TTA, M = 16 hours. Based on these
results, it is empirically evident that the cur-
rent subjects did suffer some level of fatigue.
The present study used two additional scien-
tic measures to assess fatigue and provide
additional validation of the evidence of fatigue
discussed thus far. The rst method is the
well-validated ESS on which our participants
self-reported post-shift mean results indi-
cating excessive daytime sleepiness: M = 11
(Rajaratnam et al., 2011). This reported level
of daytime sleepiness concurs with the partic-
ipants’ reported TTA, hours slept, and sleep
debt. It should be noted that the most signif-
icant change in ESS scores occurred between
Day 1 pre-shift and Day 3 post-shift: Day 1
pre-shift ESS, M = 6.5, but Day 3 post-shift
ESS, M = 12.75.

Law Enforcement Executive Forum • 2015 • 15(1)59
performance changes were associated over-
whelmingly with the no shoot and ambigu-
ous targets. Likely, the reason for this is due
to the no shoot and ambiguous situations
requiring more cognitive processing power
than clear shoot situations, using a rule-based
decision-making model (Harrison & Horne,
2000; Maddox et al., 2009).
This is the rst time a use-of-force decision-
making sleep study has been conducted in
this manner. Specically, participants were
not asked to sleep less or stay awake longer
as is often the practice in sleep-related stud-
ies. Rather, participants simply worked the
shifts and hours required by their respective
agencies. The data supported the hypothesis
by showing that fatigue does appear to affect
both deadly force reaction times and decision
making.
One different outcome of the present study
is that it did not suggest the extreme effects
of fatigue that have been reported in much
larger studies. For example, Rajaratnam et al.
(2011) conducted a large study of police
ofcers (N = 4,957) in which 40% screened
positive for a sleep disorder and about 18%
later reported making serious administrative
errors. Senjo and Heward (2007) found of-
cers were working signicantly longer hours
(66 to 75 hours weekly) and receiving much
less sleep (3 to 6 hours per night) than was
found in the present study. Vila et al. (2000)
conducted a very large study involving sev-
eral law enforcement agencies and found that
59% of ofcers did not sleep an average of 7
or more hours per night, while 16% self-re-
ported trouble staying awake while driv-
ing. In light of the effects of fatigue on the
deadly force decisions discussed in the pres-
ent study, if these more extreme sleep decits
are occurring in some police agencies, these
greater amounts would raise a serious con-
cern about police decision making in deadly
force situations.
Fatigue Survey
The survey provided troubling results demon-
strating that a very large portion of police of-
cers believe that the law enforcement industry
needs to study the impact of sleep depriva-
tion on ofcer performance. Respondents also
reported that the law enforcement industry
is not sufciently concerned with the impact
of fatigue on police performance and errors.
These results support the literature review
about the negative effects of shift work,
changes in sleep patterns, and the relation-
ship between fatigue, errors, and general job
performance, suggesting that most ofcers are
aware of these issues, contend with them reg-
ularly, and would like to see solutions to pre-
vent deteriorated job performance and errors.
The survey also supports the TCP results in
areas such as changes in sleep patterns on and
off duty, shift work and fatigue, SDS errors
and slowed RT related to fatigue, and slowed
PVT RT based on fatigue, suggesting that
police in general have an awareness of these
factors, are doing their best to compensate for
them, and are requesting assistance in reme-
dying the causes of fatigue.
Validity
Validity concerns in this study were always
within the researcher’s purview. The SDS,
although a new measurement device, was
based on a similar computer-based shoot/
don’t shoot program used in other studies
(Correll et al., 2007). The ESS/PVT results
complemented those of the SDS, providing
solid evidence of concurrent validity for the
SDS. In addition, both subjective reporting
of fatigue and reaction times increased from
pre- to post-shift testing, suggesting that the
SDS has predictive validity as well.
Limitations
This study entails several limitations. The
rst limitation is the method of assessment
delivery. The TCP was administered online
and outside of tightly controlled laboratory

Law Enforcement Executive Forum • 2015 • 15(1)60
conditions, allowing for the possibility of
minor differences in how tasks were com-
pleted. To compensate, the TCP included pho-
tographs and clear instructions, but these do
not compete with the controls available within
a laboratory environment. Additionally, due
to technical issues with the TCP, we lost an
overwhelming amount of data. A total of 215
participants logged into the program, but
only a subset (n = 53) was actually able to
complete the assessment. Most of the techni-
cal issues revolved around compatibility and
could have been remedied within a labora-
tory environment. These issues have left this
investigation with a relatively small sample
size. To protect the anonymity of police of-
cers who volunteered to be in this study, min-
imal sociodemographic data were collected.
Yet, there may be relationships between
some sociodemographic variables and perfor-
mance. For instance, most participants within
this study were men. In addition, because a
portion of the subjects were unknown to the
researcher, although unlikely, it is possible
that not all subjects were in fact police of-
cers; some subjects’ professional credentials
were not possible to verify.
Follow Up
This study applied the ndings from previ-
ous investigations to the never before tested
area of ofcer fatigue and decision-making/
reaction time during deadly force encoun-
ters. The study found, not surprisingly, that
even minor amounts of sleep deprivation,
decreased sleep quality, and shift work all
have a negative effect upon ofcers’ speed
and ability to make appropriate decisions
in deadly force situations. What may stand
out to law enforcement administrators and
policymakers are the relatively low levels of
sleep deprivation among the subjects in this
study, which nevertheless were sufcient to
cause performance decits. Several national
studies with larger sample sizes have sug-
gested that ofcers are typically much more
sleep deprived than the present subjects. As
such, the probable impact of fatigue on the
outcomes of deadly force encounters may
be a serious concern in the law enforcement
community.
Clearly, much larger samples are needed
to provide a more detailed investigation of
ofcers’ sleep habits on and off duty over a
longer period of time and the effects of fatigue
on performance. Tracking for the nature of
the sleep disturbances (e.g., court appear-
ances, overtime, and other work assignments)
should be included. A single day’s sleep dis-
turbance could greatly increase total hours
awake and cumulative sleep debt, which
both evidenced powerful effects on decision
making and reaction times within the pres-
ent SDS task. A between-subjects compari-
son comparing sleep deprived and non-sleep
deprived ofcers’ performance could also be
very productive.
Former Police Chief and current sleep
researcher Bryan Vila has been studying
police ofcer fatigue for decades. He has
established the necessity for changes in the
police community regarding sleep. Vila and
Kenney (2002) provided a list of what some of
those changes should be: (1) Police executives
should be concerned with the total number
of employee work hours; (2) Police executive
should provide employees a voice in their
shift and work hours; (3) Police executives
should assess levels of employee fatigue; and
(4) Police executives should provide employ-
ees with sleep- and fatigue-related training to
ensure good habits.
The results of the present study suggest that
law enforcement executives, risk managers,
and their legal representatives may need to
come to terms with the necessity for change
within law enforcement to reduce the adverse
effects of fatigue, particularly on the out-
come of deadly force encounters. Ignoring the
risks of excessive overtime, randomized shift
schedules, and unforgiving court appearance
schedules would appear to be unwise in light
of the data. Empirical evidence published
prior to the present study has already shown

Law Enforcement Executive Forum • 2015 • 15(1)61
the negative effects of sleep deprivation on
performance and resulted in legislation and
policy changes for some industries involved
in ensuring public safety such as truck driv-
ers, commercial airline pilots, medical resi-
dents, and air trafc controllers (Arora, 2010;
Halsey, 2012; Lockridge, 2014; Trejos, 2014). It
may be time for law enforcement to address
this long-standing issue.
The current study demonstrates agreement
with previous sleep deprivation studies in
regard to performance, and it builds on these
previous investigations by suggesting that
sleep deprivation adversely impacts law
enforcement ofcers’ most difcult decision
at the moment ofcers are faced with deadly
force encounters. Just as no law enforcement
executive would place an intoxicated ofcer
on the street, they may come to understand
the dangers of placing a fatigued ofcer in the
line of duty. Tired cops make inferior deci-
sions and react more slowly, placing them-
selves and the public they serve at unneces-
sary risk.
References
Adams, K., Alpert, G. P., Dunham, R. G., Garner,
J. H., Greenfeld, L. A., Henriquez, M. A.,
. . . Smith, S. K. (1999). Use of force by police:
Overview of national and local data. Retrieved
from https://www.ncjrs.gov/pdfles1/nij/
176330-1.
pdf.
Adam, R., Bays, P. M., & Husain, M. (2012).
Rapid decision-making under risk. Cogni-
tive Neuroscience, 3(1), 52-61.
Alhola, P., & Polo-Kantola, P. (2007). Sleep
deprivation: Impact on cognitive perfor-
mance. Neuropsychiatric Disease and Treat-
ment, 3(5), 553-567.
Antal, L. C. (1975). The effects of the changes of
the circadian body rhythm on the sport
shooter. Retrieved from www.ncbi.nlm.nih.
gov/pmc/articles/PMC1859287.
Arora, V. (2010). Patient handoffs limit residents
work hour cap gains. Retrieved from www.
kevinmd.com/blog/2010/03/patient-handoffs-
limit-residents-work-hour-
cap-gains.html.
Banks, S., & Dinges, D. F. (2007). Behavioral
and physiological consequences of sleep
restriction. Journal of Clinical Sleep Medicine,
3(5), 519-552.
Barger, L. K., Lockley, S. W., Rajaratnam, S. M.,
& Landrigan, C. P. (2009). Neurobehavioral,
health, and safety consequences associated
with shift work in safety-sensitive profes-
sions. Current Neurology and Neuroscience
Reports, 9(2), 155-164.
Baumann, J., & DeSteno, D. (2010). Emotion
guided threat detection: Expecting guns
where there are none. Journal of Personality
and Social Psychology, 99(4), 595-610.
Belenky, G., Wesensten, N. J., Thorne, D. R.,
Thomas, M. L., Sing, H. C., Redmond, D. P.,
. . . Balkin, T. J. (2003).
Patterns of perfor-
mance degradation and restoration during
sleep restriction and subsequent recovery:
A sleep dose-response study. Journal of Sleep
Research, 12(1), 1-12.
Blair, J. P., Pollock, J., Montague, D., Nichols,
T., Curnutt, J., & Burns, D. (2011). Reason-
ableness and reaction time. Police Quarterly,
14(4), 323-343.
Bonnet, M. H., & Arand, D. L. (1995). We are
chronically sleep deprived. Retrieved from
www.soundersleep.com/uploads/Bonnet-
Arand.pdf.
Correll, J., Wittenbrink, B., Park, B., Judd, C. M.,
Sadler, M., & Keesee, T. (2007). Across the
thin blue line: Police ofcers and racial bias
in the decision to shoot. Journal of Personality
and Social Psychology, 92(6), 1006-1023.
Couyoumdjian, A., Sdoia, S., Tempesta, D.,
Curcio, G., Rastenllini, E., De Gennaro, L.,
& Ferrara, M. (2009). The effects of sleep and

Law Enforcement Executive Forum • 2015 • 15(1)62
sleep deprivation on task-switching perfor-
mance. Journal of Sleep Research, 19(1), 64-70.
Dinges, D. F., & Basner, M. (2011). Maximiz-
ing sensitivity of the Psychomotor Vigi-
lance Test (PVT) to sleep loss. SLEEP, 34(5),
581-591.
Dinges, D. F., & Basner, M. (2012). An adap-
tive-duration version of the PVT accurately
tracks changes in psychomotor vigilance
induced by sleep restriction. SLEEP, 35(2),
193-202.
Dinges, D. F., Pack, F., Williams, K., Gillen, K.
A., Powell, J. W., Ott, G. E., . . . Pack, A. I.
(1997). Cumulative sleepiness, mood distur-
bance, and psychomotor vigilance perfor-
mance decrements during a week of sleep
restricted to 4-5 hours per night. American
Sleep Disorders Association and Sleep Research
Society, 20(4), 267-277.
Dinges, D. F., Rogers, N. L., & Baynard, M. D.
(n.d.). Chronic sleep deprivation. Retrieved from
www.med.upenn.edu/uep/user_documents/
dfd22.pdf.
Dorrian, J., Rogers, N. L., & Dinges, D. F. (n.d.).
Psychomotor vigilance performance: Neurocogni-
tive assay sensitive to sleep loss. Retrieved from
www.med.upenn.edu/uep/user_documents/
Dorrianetal.PVTchapterinKushida2005.pdf.
Durmer, J. S., & Dinges, D. F. (2005). Neuro-
cognitive consequences of sleep depriva-
tion. Seminars in Neurology, 25(1), 117-129.
Edinger, J. D., & Fins, A. I. (1995). The distri-
bution and clinical signicance of sleep time
misperceptions among insomniacs. SLEEP,
18(4), 232-239.
Edwards, B. J., & Waterhouse, J. (2009). Effects
of one night of partial sleep deprivation
upon diurnal rhythms of accuracy and con-
sistency in throwing darts. Chronobiology
International, 26(4), 756-768. http://dx.doi.
org/10.1080/07420520902929037.
Force Science Institute. (2005, January). Force
Science News #10: Was suspect’s shooting a
police execution? Detailed ndings from Force
Science Research Center help federal jury
decide.
Retrieved from www.forcescience.org/
fsnews/
10.html.
Gartenberg, D., & Parasuraman, R. (2010).
Understanding brain arousal and sleep quality
using a neuroergonomic smart phone application.
Retrieved from www.crcnetbase.com/doi/
abs/10.1201/EBK1439835012-c20.
Graham v. Connor, 490 U.S. 386 (1989).
Retrieved from http://supreme.justia.com/
cases/federal/us/490/386/case.html#399.
Halsey, A. (2012, June 15). Air trafc controllers
violating no-doze schedule. The Washington
Post. Retrieved from www.washingtonpost.
com/local/trafcandcommuting/air-trafc-
controllers-arent-keeping-to-no-doze-schedule/
2012/06/15/gJQAJD5FfV_story.html.
Harrison, Y., & Horne, J. A. (2000). The impact
of sleep deprivation on decision making: A
review. Journal of Experimental Psychology:
Applied, 6(3), 236-249. Retrieved from http://
postcog.ucd.ie/files/Hrrison%20and%20
horne.pdf.
Hoddes, E., Zarcone, V., Smythe, H., Phillips,
R., & Dement, R. (1973). Quantications of
sleepiness: A new approach. Psychophysiol-
ogy, 10(4), 431-436.
Honig, A., & Lewinski, W. J. (2008). A survey
of the research on human factors related
to lethal force encounters: Implications for
law enforcement training, tactics, and testi-
mony. Law Enforcement Executive Forum, 8(4),
129-152. Retrieved from www.forcescience.
org/articles/Honig129-152.pdf.
Johns, M. W. (2000). Sensitivity and specicity
of the Multiple Sleep Latency Test (MSLT),
the maintenance of wakefulness test and
the Epworth Sleepiness Scale: Failure of the

Law Enforcement Executive Forum • 2015 • 15(1)63
MSLT as a gold standard. Journal of Sleep
Research, 9(1), 5-11.
Johnson, J. L. (2007). Use of force and the
Hollywood factor. Retrieved from www.aele.
org/law/2007-04MLJ501.pdf.
Lewinksi, W. J. (2000). Why is the suspect shot
in the back? Finally, hard data on how fast the
suspect can be in 11 different shooting scenar-
ios. Retrieved from www.forcescience.org/
articles/shotback.pdf.
Lewinski, W. J., & Honig, A. (2008). A survey
of the research on human factors related to lethal
force encounters: Implications for law enforcement
training, tactics, and testimony. Retrieved from
www.forcescience.org/encounter.html.
Lewinski, W. J., & Hudson, B. (2003). Time to
start shooting? Time to stop shooting? The
Tempe Study. The Police Marksman, 28(5),
26-29. Retrieved from www.forcescience.org/
articles/
tempestudy.pdf.
Lewinski, W. J., & Redmann, C. (2009). New
developments in understanding the behav-
ioral science factors in the “stop shooting”
response. Law Enforcement Executive Forum,
9(4), 35-54.
Lim, J., & Dinges, D. F. (2010). A meta-analy-
sis of the impact of short-term sleep depri-
vation on cognitive variables. Psychological
Bulletin, 136(3), 375-389.
Lockridge, D. (2014). Ferro discusses split sleep
study, CSA, hours of service, electronic logs
at MATS eet forum. Retrieved from www.
truckinginfo.com/channel/drivers/news/
story/2014/04/ferro-discusses-split-sleep-
study-csa-hours-of-service-electronic-logs-
at-mats-eet-forum.aspx.
Maddox, W. T., Glass, B. D., Wolosin, S. M.,
Bowen, C., Mathews, M. D., & Schnyer, D. M.
(2009). The effects of sleep deprivation on
information-integration categorization per-
formance.
SLEEP, 32, 1439-1448. Retrieved from
www.cti-home.com/wp-content/uploads/
2014/01/Effect-of-Sleep-Deprivation-On-
Performance.pdf.
Miller, T. (n.d.). Graham v. Connor. Retrieved
from https://www.etc.gov/sites/default/
files/imported_files/
training/programs/
legal-division/podcasts/fletc-legal-
division-
use-of-force/podcast-transcripts/
PartIGrahamvConnor.pdf.
National Highway Transportation Safety Ad-
ministration (NHTSA). (1996). Drowsy driving
and automobile crashes. Retrieved from www.
nhtsa.gov/people/injury/drowsy_driving1/
drowsy.
html.
NHTSA. (2006). DWI detection & standard-
ized eld sobriety testing: Instructor manual.
Retrieved from www.aschemansmith.com/
Websites/aschemansmith/Images/NHTSA/
SupportDocs_DRE_Forms_Manuals_dwi_
Instructor%20Manual%20-%20February%
202006.pdf.
Neylan, T. C., Metzler, T. J., Best, S. R., Weiss,
D. S., Fagan, J. A., Liberman, A., . . . Marmar,
C. R. (2002). Critical incident exposure and
sleep quality in police ofcers. Psychosomatic
Medicine, 6 4 (2), 345-352. Retrieved from www.
ncbi.nlm.nih.gov/pubmed/11914452.
O’Brien, M. J., O’Toole, R. V., Zadnik-Newell,
M., Lydecker, A. D., Nascone, J., Sciadini,
M., . . . Eglseder, W. A. (2012). Does sleep
deprivation impair orthopeadic surgeons’
cognitive and psychomotor performance?
The Journal of Bone and Joint Surgery, 94,
1975-1981.
Rajaratnam, S. M., Barger, L. K., Lockley,
S. W., Shea, S. A., Wang, W., Landrigan, C.
P., . . . Czeisler, C. A. (2011). Sleep disorders,
health, and safety in police ofcers. Journal
of the American Medical Association (JAMA),
306(23), 2567-2578.
Samkoff, J. S., & Jacques, C. H. M. (1991).
A review of studies concerning effects

Law Enforcement Executive Forum • 2015 • 15(1)64
of sleep deprivation and fatigue on resi-
dents’ performance. Academic Medicine, 66,
687-693. Retrieved from http://journals.
lww.com/academicmedicine/Abstract/1991/
11000/A_review_of_studies_concerning_
effects_of_sleep.13.aspx.
Senjo, S. R. (2011). Dangerous fatigue con-
ditions: A study of police work and law
enforcement administration. Police Practice
and Research, 12(3), 235-252.
Senjo, S. R., & Heward, M. E. (2007). Sleep and job
performance in law enforcement: Measuring differ-
ences between highway patrol, sheriff, and munic-
ipal police ofcers. Retrieved from https://
kucampus.kaplan.edu/documentstore/
docs09/pdf/picj/vol2/issue2/Sleep_and_
Job_Performance_in_Law_Enforcement.pdf.
Sharp, M. J., & Hess, A. B. (2008). To shoot or
not to shoot: Response and interpretation of
response to armed assailants. Retrieved from
www.forcescience.org/articles/
study%20
shoot-or-not.pdf.
Sundelin, T., Lekander, M., Kecklund, G.,
Van Someren, E. J. W., Olsson, A., & Axelsson,
J. (2013). Cues of fatigue: Effects of sleep depri-
vation on facial appearance. SLEEP, 36, 1355-
1360. Retrieved from www.journalsleep.org/
ViewAbstract.aspx?pid=29095.
Tikuisis, P., Keefe, A. A., McLellan, T. M.,
& Kamimori, G. (2004). Caffeine restores
engagement speed but not shooting preci-
sion following 22 h of active wakefulness.
Aviation, Space, and Environmental Medicine,
75(9), 771-776.
Trejos, N. (2014, January 3). New pilot fatigue
rules go into effect this weekend. USA
Today. Retrieved from www.usatoday.com/
story/todayinthesky/2014/01/03/pilot-
fatigue-mandatory-rest-new-faa-rules/
4304417.
U.S. Department of Health and Human Ser-
vices (DHHS) (2005). Your guide to healthy
sleep. Retrieved from www.nhlbi.nih.gov/
les/docs/public/sleep/healthy_sleep.pdf.
Van Dongen, H. P. A., Maislin, G., Mullington,
J. M., & Dinges, D. F. (2003). The cumulative
cost of additional wakefulness: Dose-response
effects on neurobehavioral functions and
sleep physiology from sleep restriction and
total sleep deprivation. SLEEP, 26(2), 117-126.
Retrieved from www.med.upenn.edu/uep/
user_documents/dfd16.pdf.
Vila, B., & Kenney, D. J. (2002). Tired cops:
The prevalence and potential consequences
of police fatigue. NIJ Journal, 248, 16-21.
Retrieved from https://www.ncjrs.gov/
pdfles1/jr000248d.pdf.
Vila, B., Kenney, D. J., Morrison, G. B.,
& Reuland, M. (2000). Evaluating the effects
of fatigue on police patrol ofcers: Final report.
(Unpublished report). Retrieved from www.
ncjrs.gov/pdfles1/nij/grants/184188.pdf.
Williamson, A. M., & Feyer, A. M. (2000). Mod-
erate sleep deprivation produces impair-
ments in cognitive and motor performance
equivalent to legally prescribed levels of alco-
hol intoxication. Occupational & Environmental
Medicine, 57, 649-655. Retrieved from http://
oem.bmj.com/content/57/10/649.full.
David M. Blake is a retired law enforcement
ofcer and avid student of law enforce-
ment use of force. He is a Force Science
Certied Analyst with instructor certica-
tions in Defensive Tactics, Firearms, Force
Options Simulator, and Reality Based
Training. He currently teaches Human
Factors and Force Encounters Analysis for
the California Training Institute. He is also
an adjunct professor of Criminal Justice, a
police academy instructor, and an inservice
use-of-force instructor. He owns the Blake
Consulting and Training Group.
Dr. Edward Cumella is a professor of Grad-
uate Psychology at Kaplan University. He

Law Enforcement Executive Forum • 2015 • 15(1)65
received his Bachelor of Arts at Harvard,
and his Master of Arts/PhD in Psychology
at University of North Carolina Chapel
Hill. He has worked in mental health for 29
years. Previously, he was Executive Direc-
tor at America’s largest eating disorder facil-
ity and in private practice. Dr. Cumella has
published 50 peer-reviewed articles and has
been interviewed on TV, radio, and news-
papers (e.g., ABC, FOX, New York Times).
He is the editor of a book on eating
disorders.
Contact Information
David M. Blake
dblake66@gmail.com
Dr. Edward Cumella
ECumella@kaplan.edu