Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

International Journal of Police Science & Management Volume 15 Number 2

A longitudinal examination of officer


deaths from vehicle pursuits

Richard R. Johnson
Criminal Justice Program, University of Toledo, 2801 W. Bancroft St., MS 119, Toledo,
OH 43606, USA. Tel: +1 419 530 4639; Fax: +1 419 530 2153;
email: Richard.Johnson4@utoledo.edu
Submitted 16 June 2013, revision submitted 1 October 2013, accepted
1 October 2013
Keywords: pursuit, police, law enforcement, officer safety, vehicle safety,
public policy

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


Richard R. Johnson, PhD, is an associate pro- majority of vehicle pursuits are initiated for
fessor of criminal justice at the University of minor criminal or traffic offences (Alpert,
Toledo in Ohio, and he is a former law enforce- 1997), 40 per cent of pursuits result in
ment officer. He holds a doctorate in criminal property damage from a motor vehicle
justice from University of Cincinnati. His research crash (Alpert, Dunham, & Stroshine, 2006),
interests include issues of police management with damage costs estimated to be over
and police–citizen interactions. $1,000,000 annually (Alpert et al., 2006).
Almost half of pursuit crashes also result in
physical injuries to the occupants of the
ABSTRACT
fleeing vehicle, police officers or
Little research has examined the changing
uninvolved third parties (Alpert et al.,
dangerousness of police high-speed pursuits over
2006). Approximately 1 in every 100 pur-
time, especially with regard to the safety of the
suits results in a fatality (Hill, 2002), with
officer. Using data from all pursuit-related officer
more than one third of these fatalities being
deaths in the USA between 1960 and 2011,
uninvolved third parties or police officers
this study tested the hypothesis that the officer
(Rivara & Mack, 2004).
death rate has been decreasing. Least-squares
Fortunately, the dangers posed by police
lines were calculated through the data point of
vehicle pursuits appear to have been
each year to estimate trends in officer death rates
decreasing in recent years. Improvements in
over time. The findings revealed that the death
vehicle safety technology (Vanderbilt,
rate for officers directly involved in pursuits has
2009), emergency medicine (National
decreased steadily, as has the death rate for officers
manning roadblocks. The death rate for officers Research Council, 2007), more restrictive
deploying spike strips or travelling to assist in the police departmental policies governing pur-
pursuit, however, has been steadily increasing. suits (Alpert et al., 2006) and anti-pursuit
The findings suggest further development of pur- technologies such as spike strips to deflate
suit policies to address these issues. the suspect’s vehicle (Hill, 2002) have com-
bined to reduce the broad risk to the public
posed by pursuits. One examination of
INTRODUCTION pursuit-related fatalities in the USA from International Journal of Police
Science and Management,
Police high-speed vehicle pursuits create 1994 to 2002 revealed that overall fatalities Vol. 15 No. 2, 2013, pp. 77–94.
DOI: 10.1350/ijps.2013.15.2.304
many hazards for the public. While the vast from pursuits decreased by almost 20 per

Page 77
Officer deaths from vehicle pursuits

cent during that nine-year period (Rivara & Vehicle crash survivability
Mack, 2004). Although some have studied Over the last several decades, vehicle manu-
the overall dangerousness of vehicle pursuits facturers have made great strides in improv-
(Hill, 2002; Rivara & Mack, 2004), little ing the survivability of vehicle occupants
research has examined whether vehicle pur- involved in crashes (Vanderbilt, 2009).
suits have become increasingly safer for the These improvements have included seatbelts
police officers involved. with shoulder restraints, front- and side-
On the one hand, it could be argued that impact airbags, fire-resistant fuel tanks and
the same changes and improvements that vehicle bodies that absorb kinetic energy
have helped reduce the overall fatality rate and direct it away from the occupants
for vehicle pursuits would have had a similar (Vanderbilt, 2009). Seatbelts, which were
effect on officer fatalities. On the other first introduced in the 1950s and only used
hand, the National Highway Traffic Safety 11 per cent of the time in 1980, have seen
Administration (2011) reported that, increased usage with an estimated three-
between 1980 and 2008, the raw number of
quarters of drivers wearing their seatbelts
police officer deaths from pursuits

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


today (National Highway Traffic Safety
increased. Thus the very limited knowledge
Administration, 1998). While studies of
about officer pursuit lethality is equivocal.
seatbelt effectiveness in reducing deaths in
It would be beneficial, therefore, to spe-
crashes have ranged from 8 to 86 per cent
cifically examine the officer fatality rate
(Levitt & Porter, 2001; Robertson, 1976),
(rather than raw numbers) for pursuit-
depending on the characteristics of the
related deaths. This study sought to do this
occupant and the circumstances of the
by longitudinally examining the number of
pursuit-related officer deaths per 100,000 crash, adult drivers wearing seatbelts in a
officers employed in the USA from 1960 to crash are generally 55 per cent more likely
2011. It tested the research hypothesis that to survive the crash than drivers not wear-
the officer death rate from pursuits has been ing seatbelts (Halman, Chipman, Parkin, &
decreasing since 1960. It also tested this Wright, 2002; Levitt & Porter, 2001).
hypothesis using three different types of Drivers’ side airbags were first introduced
pursuit involvement: direct pursuit of the into vehicles in 1987 and were being
suspect, executing a tactic to block the installed in all new vehicles sold in the USA
fleeing suspect, and travelling to reach and by 1998 (Levitt & Porter, 2001). As with
assist in a pursuit. seatbelts, the effectiveness of airbags varies
widely with the characteristics of the
Changes influencing dangerousness vehicle occupant and the circumstances of
of pursuits the crash. On average, however, adult
Over the last half-century, substantial drivers of vehicles with airbags who were
changes have occurred within the areas of involved in head-on crashes were 35 per
vehicle accident survivability, legal decisions cent more likely to survive after controlling
governing pursuits, police pursuit policies for seatbelt use (Braver, Ferguson, Greene,
and technological advancements. Taken col- & Lund, 1997; Levitt & Porter, 2001;
lectively, these changes would suggest that Zador & Ciccone, 1993). Vehicle handling
police pursuits have gradually become less characteristics have also improved, increas-
frequent, and, when they do occur, less ing one’s chances of avoiding a crash in the
likely to result in death than in years past. first place, and innovations in road con-
Each of these areas of change is discussed in struction and traffic signals have continued
greater detail below. to improve road safety (Vanderbilt, 2009).

Page 78
Johnson

Advances in trauma medicine and emer- per 100 million vehicle miles travelled
gency medical responses have improved the between 1966 and 1994, Levitt and Porter
chances that crash victims will survive their (2001) found that the average driver fatality
injuries. Ambulance services first became rate declined steadily, decreasing by almost
commonplace in the USA during the 70 per cent across the 29-year period. More
1950s, but were little more than a station recent data on motor vehicle crash fatalities
wagon with a stretcher (National Research in the USA between 1994 and 2011 found
Council, 2007). Often run by funeral that the number of fatalities per 100,000
homes, petrol stations, towing services or licensed drivers had declined from 23.21 in
police departments, between the 1950s and 1994 to 15.28 in 2011, a decrease of 34 per
the mid-1970s, ambulance crews typically cent over 18 years (National Highway Traf-
had no formal medical training or, at best, fic Safety Administration, 2013). It could
only basic first-aid training (National reasonably be assumed from these facts
Research Council, 2007). Emergency med- about motor vehicle accidents in general,
ical technician (EMT) training first began that police pursuits have also become pro-
in the early 1970s, expanding rapidly until gressively safer for the officers engaged in

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


almost all ambulance crews had received them. It is likely that police officers have
such training by 1990. As ambulance crew benefited, over time, from vehicles with
training became more sophisticated, so did better handling performance and occupant
the life-saving equipment available in most protection during crashes, as well as the
ambulances (National Research Council, emergency medical care available after
2007). Response times to emergencies has crashes.
also improved, in the 1980s, the industry
standard was a response time of 15 minutes, Changes in pursuit law
whereas the current industry standard is Since the mid-1980s, a number of landmark
arrival on the scene of an emergency in less federal court case decisions have involved
than 8 minutes (Pell, Sirel, Marsden, Ford, police vehicle pursuits. In Galas v McKee
& Cobbe, 2001; Pons & Markovchick, (1986), the US Court of Appeals for the
2002). Sixth Circuit formally acknowledged the
Providing even faster response and trans- permissibility of high-speed vehicle pursuits
port across great distances, air ambulances by the police, stating that the act of pursu-
have become more common over the last ing a fleeing felon or misdemeanant by
few decades. The first hospital-based air vehicle, even with its inherent dangers, does
ambulance service in the USA began in not constitute a violation of the Fourth
Denver, Colorado in 1972, but rapidly Amendment to the US Constitution. This
expanded in use in both urban and rural ruling was reaffirmed by the US Supreme
areas (Sullivent, Faul, & Wald, 2011). It is Court in County of Sacramento v Lewis
estimated that more than 1,000 air ambu- (1998). Through these and other cases, the
lances operated within the USA by 2010, courts have protected the right of the police
and trauma victims transported by air to engage in high-speed vehicle pursuits,
ambulance are 39 per cent more likely to yet several cases have placed limits and con-
survive than those transported by land sequences on police actions during these
ambulance (Sullivent et al., 2011). pursuits.
All of these medical and safety improve- One of the earliest landmark cases to
ments have combined to result in many tens influence police pursuit tactics was Tennessee
of thousands of lives saved in motor vehicle v Garner (1985). Although Garner involved
accidents. Examining US traffic fatality rates using a firearm to stop an unarmed felon

Page 79
Officer deaths from vehicle pursuits

fleeing on foot, the ruling was also applic- pursuit posed a danger disproportionate to
able to vehicle pursuits. In Garner, the US the seriousness of the offence for which the
Supreme Court eliminated the ability of the suspect was being chased (Becknell, Mays,
police to use potentially lethal force simply & Giever, 1999; Pipes & Pape, 2001).
to apprehend a fleeing criminal suspect These court decisions, occurring
unless the officer or a third party was in between 1985 and 1989, changed the legal
danger of imminent serious bodily injury landscape of police pursuits and influenced
(Tennessee v Garner, 1985). This ruling the development of stricter police pursuit
essentially banned such police pursuit prac- policies. It could be suggested that these
tices as shooting at fleeing vehicles from a legal changes would have made vehicle pur-
roadblock or moving vehicle, or ramming suits safer for officers in many ways. First,
fleeing vehicles with the likelihood of injur- restricting the use of firearms in vehicle
ing the driver. pursuits, officers would be less likely to be
Jamieson v Shaw (1985), decided by the injured by a stray bullet fired at the fleeing
Fifth Circuit of the US Court of Appeals, suspect’s vehicle. Second, restricting the use
extended Garner to include other roadblock of ramming the suspect’s vehicle, officers

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


tactics that had a high likelihood of severely would be less likely to be injured in the
injuring or killing a fleeing suspect. In ramming. Third, restricting the use of road-
particular, this case suggested that placing a blocks would make officers less likely to be
patrol car across the road on a blind curve, rammed by the fleeing suspect vehicle or hit
and shining bright lights in the fleeing as the suspect tried to swerve around the
driver’s eyes to prevent him from seeing the roadblock. Fourth, expanding police civil
roadblock before crashing into it, consti- liability for injuries or property damage
tuted an unreasonable seizure. The US caused by the pursuit, officers would be less
Supreme Court affirmed this ruling in likely to engage in pursuits at all, and more
Bower v Inyo County (1989), ruling that likely to call off pursuits when the risks to
roadblock tactics that create a likelihood of public safety outweigh the importance of
death or injury to the fleeing suspect con- catching the suspect. These propositions
stituted an unreasonable use of force. In this suggest that the police officer fatality rate
case, officers had placed a semi-trailer truck from vehicle pursuits has declined, espe-
completely across the highway around a cially since 1985.
blind curve, with a patrol car’s headlights
aimed to blind the fleeing driver on Changes in pursuit policies
approach so that he would crash into the Court decisions regarding pursuits have
truck. These two cases basically eliminated prompted many law enforcement agencies
the use of most police roadblock tactics. to establish written policies and procedures
Court decisions also opened up the regarding pursuits (Becknell et al., 1999;
police to civil liability for any injuries to Pipes & Pape, 2001). In the early 1960s,
third parties sustained because of a high- very few law enforcement agencies had
speed pursuit. In the case of City of Canton, written policies governing the use of force
Ohio v Harris (1989), the US Supreme or pursuits (Alpert & Fridell, 1992). This
Court decided that civil liability exists began to change in the 1970s as civil litiga-
when the police exhibit ‘deliberate indif- tion against the police increased. Police
ference’ to the rights of persons with whom departments found themselves losing law-
they contact. Although this case did not suits because they had failed to properly
involve a police pursuit, it has been widely train and supervise their officers, amounting
applied to pursuit situations in which the to a deliberate indifference to the rights of

Page 80
Johnson

citizens (Alpert & Fridell, 1992). The num- the added benefits of reducing the number
ber of law enforcement agencies with for- of permissible pursuits, and calling for the
mal, written policies on when and how termination of pursuits when the risks to
pursuits could be conducted, expanded the public far outweighed the benefits of
rapidly in the 1980s until such policies were apprehending the suspect. In fact, Alpert
commonplace in the mid-1990s. In a 1996 (1997) found that almost half of the law
nationally representative sample of 436 enforcement agencies surveyed had updated
municipal and county law enforcement their pursuit policy within the last two
agencies, Alpert (1997) found that 91 per years, and most of these changes had made
cent of these agencies had written policies their policy more restrictive than before, a
governing pursuits. Another national survey finding supported by Becknell et al. (1999).
in 1998 of 420 law enforcement agencies In one such agency, the more restrictive
revealed that 97 per cent reported having a policy coincided with an 82 per cent reduc-
written pursuit policy (National Institute of tion in pursuits the year after the intro-
Justice, 1998a). duction of the new policy (Alpert, 1997).
Although some of these pursuit policies

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


gave officers great discretion over when to Changes in pursuit training
initiate a pursuit, and the tactics to be Although pursuit driving makes up only a
employed during the pursuit, even the small proportion of police academy training
limited structure provided by these policies (Alpert, 1997), the number of hours
was an improvement over having no writ- devoted to academy training has grown over
ten policy at all (Alpert & Fridell, 1992). At the last five decades (Haberfeld, 2002).
the very least, these permissive policies Matheny (1996) reported that in 1961,
required officers to abide by the federal some New Jersey law enforcement agencies
court decisions prohibiting firing at a flee- required no pre-service academy training,
ing suspect or using roadblocks likely to whereas others had academy training
cause serious harm. Some agencies, how- requirements that ranged from 10 to 40
ever, did develop more restrictive pursuit hours. McManus (1970) surveyed a national
policies that strictly limited when vehicle sample of 360 city police departments and
pursuits could be initiated. These restrictive found that all required some sort of basic
policies often prevented pursuits for only police academy training, and the shortest
traffic offences, and required officers to take academy training reported was 120 hours.
into consideration the environment in In 1993, all 50 states had established basic
which the pursuit was taking place, such as police officer training and standards guide-
the excessive speeds involved or traffic con- lines, and the average police academy train-
gestion (Alpert & Fridell, 1992). A few ing length was 640 hours (Edwards, 1993).
agencies even had policies that prohibited In a 1997 survey of 436 law enforcement
vehicle pursuits except in extreme situations agencies, 60 per cent of agencies acknow-
(such as trying to stop a kidnapping in ledged covering pursuit driving in their
progress or apprehend a known murder academy curriculum, with a mean of 14
suspect), and even then, only with a super- hours of pursuit training (Alpert, 1997).
visor’s approval. These agencies also reported spending an
It could be argued that the proliferation average of 3 hours annually on in-service
of written pursuit policies has improved training with regard to pursuits.
officer safety by avoiding dangerous road- Police pursuit training has also improved
blocks and eliminating gunfire. More in quality. One way in which this has
restrictive pursuit policies would provide occurred is through the use of computer

Page 81
Officer deaths from vehicle pursuits

simulation systems that replicate the experi- Since the late 1980s, a number of new
ence of a high-speed pursuit, but do not technologies have been developed to disable
pose dangers to the trainee or department the vehicles of fleeing suspects. Devices
vehicles (Alpert et al., 2006). These com- have been developed for police use that can
puter simulations also allow instructors to turn off the fleeing vehicle’s engine through
systematically debrief trainees about their electronic pulses, but these devices are still
actions during each step of the pursuit in their infancy today (Alpert et al., 2006).
(Alpert et al., 2006; Hill, 2002). Police The most developed, and widely used,
officers, therefore, have progressively police technology to disable a fleeing sus-
received more, and better, training in pur- pect’s vehicle has been the use of spike
suit driving since the early 1960s, leading to strips. These strips, placed in the fleeing
the assumption that this training has made vehicle’s path, slowly deflate the vehicle’s
vehicle pursuits safer for officers. tyres without causing the vehicle to lose
control at high speeds (Alpert et al., 2006;
Changes in pursuit technology National Institute of Justice, 1998a).
According to the US Patents Office, in the

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


Several changes in police vehicle pursuit
technology that might potentially have 1960s, no patents were filed for spike strip,
tyre deflation devices. In the 1970s, two
made vehicle pursuits safer for officers have
such patents were filed, and in the 1980s,
occurred over the last several decades. Some
seven patents were filed (US Patent and
law enforcement agencies have incorpor-
Trademark Office, 2012). In the 1990s,
ated helicopters into their pursuit tactics,
however, 19 patents were filed, and another
permitting fleeing suspects to be followed
24 were filed in the first decade of the
safely from the air without endangering
2000s. A national survey of 419 large police
other motorists on the ground. Helicopters
departments found that these spike strip
can also direct ground units to the suspect devices were the anti-pursuit technology
when he stops or exits the vehicle (Alpert, used most by these agencies, and were
1998). Unfortunately, helicopters are pro- strongly supported by officers in the field.
hibitively expensive for most police depart- These technological advances all suggest
ments. Police radio systems, however, have that vehicle pursuits may have become safer
improved dramatically for all police depart- for police officers.
ments since 1960, when many radios in
police cars could not transmit, but only Officer deaths in pursuits
receive calls (Sanow, 2002). In a steady Although the innovations and policy
evolution of technological improvements, changes described above could be assumed
police agencies gradually developed two- to decrease the fatality rate for police
way radios, car-to-car communications, officers involved in pursuits, the only study
centralized dispatch centres, computer- of this issue to date produced unexpected
aided dispatch and in-car computers sys- results. The National Highway Traffic Safety
tems with instant messaging and Administration (2011) examined all
global-positioning capabilities. These com- vehicle-related deaths of police officers
munication improvements allowed better between 1980 and 2008. The total number
coordination of pursuits between multiple of police officer deaths from all types of
patrol cars, and allowed supervisors to mon- crashes, including pursuit-related crashes,
itor and control the pursuit by radio and had actually increased since 1980. Unfortu-
computer (National Institute of Justice, nately, this report did not control for the
1998b). growing number of law enforcement

Page 82
Johnson

officers employed in the USA across that covered data on officer deaths from 1971
same period (Lilley & Boba, 2008; Worrall (the data reported in the 1972 report) to
& Zhao, 2003). The higher number of 2011, additional data sources had to be used
deaths could be the result of a steady or to reconstruct officer deaths from 1960 to
declining fatality rate, but with significantly 1970.
more officers patrolling the roads. It would For these years, the primary information
be beneficial, therefore, to specifically source was the Officer-Down Memorial
examine the officer fatality rate for pursuit- Page website operated by the Concern of
related deaths. Police Survivors (COPS) organisation
This study sought to do this by examin- (Concerns of Police Survivors, 2013). On
ing longitudinally the number of pursuit- this website, law enforcement agencies post
related officer deaths per 100,000 officers the names of their officers who died in the
employed in the USA between 1960 and line of duty. Most of these entries include a
2011. It tested the research hypothesis that description of the death event and a photo-
the officer death rate from pursuits has been graph of the officer. While the majority of
decreasing since 1960. It also tested this entries are from post-World War II officer

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


hypothesis using three different types of deaths, many agencies even have entries
pursuit involvement: direct pursuit of the about officers killed in the 19th century.
suspect; executing a tactic to block the The validity of this website was checked by
fleeing suspect; and travelling to reach and comparing the pursuit-related officer deaths
assist in a pursuit.
from the LEOKA reports and the website
for the years 1971–2011. Although the data
METHOD from the 19th and early 20th centuries are
probably less complete, the data from the
This study examined longitudinally officer
website and LOKEA reports matched per-
death rates related to high-speed, vehicle
fectly for the period of overlap.
pursuits to test the research proposition that
pursuit death rates have been decreasing To supplement the pre-1971 data and
between 1960 and 2011. The study exam- any cases in the LOKEA reports that were
ined total officer death rate related to pur- unclear, articles were collected from the
suits, death rate due to direct pursuit archives of the local newspapers to garner
involvement, death rate due to participation further information about each officer
in blocking a pursuit and death rate related death circumstance. These newspaper art-
to travelling to assist in a pursuit. icles identified a few additional officers who
Data for the study were collected from were listed as dying in vehicle crashes, but
various sources. Officer death data were as details were reviewed, it was determined
collected from the Law Enforcement Officers that the crash occurred as the officer was
Killed and Assaulted (LEOKA) reports pub- travelling to assist in a pursuit.
lished annually by the Federal Bureau of To calculate the rate of officer deaths, the
Investigation from 1972 to the present. number of officers potentially at risk must
These reports provide one- or two- be known. The annual number of sworn
paragraph narratives of the death circum- law enforcement officers employed in the
stances of each officer that died in the line USA has been estimated by the Federal
of duty that year, providing enough detail Bureau of Investigation, and provided in
to determine if the officer’s death resulted their Uniform Crime Reports since 1962.
from a pursuit, and the role the officer These annual estimates were used to calcu-
played in the pursuit. As these reports only late the number of officer deaths per

Page 83
Officer deaths from vehicle pursuits

100,000 officers in the USA, with years measure changes in trends before and after a
1960 and 1961 using the 1962 estimate. specific point in time when an intervention
was introduced (Box et al., 1994). As was
Sample revealed earlier in this paper, changes in
In the years 1960–2011, a total of 479 law medical care, vehicle safety, police policies,
enforcement officers lost their lives as a training and technology occurred gradually,
result of vehicle-pursuit-related circum- rather than being implemented at one point
stances. The largest number of total deaths in time. The only exception to this may
occurred in the decade of the 1960s (106 have been changes to the law through
deaths), followed by the 1970s (101 deaths). appellate court decisions from 1985 to
The decade with the lowest number of 1989.
deaths was the 1990s, with only 64 deaths. It was decided, therefore, to calculate
The vast majority of all deaths (73.7 per regression lines through the data points to
cent) occurred with officers in direct pur- determine the ‘line of best fit’ (Cohen &
suit of a fleeing suspect. Another 19.6 per Cohen, 1983) to reveal whether the slope
cent died while attempting some sort of was declining or inclining over time. A

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


blocking technique, such as manning a regression line would reveal if the trend was
roadblock or deploying spike strips into the linear, and if it was positive or negative. Flat
path of the fleeing suspect. The remaining line trends would fail to demonstrate statist-
6.7 per cent died from vehicle crashes while ical significance. Using year as the inde-
travelling to assist with a pursuit, either by pendent variable, and each death rate
attempting to catch up with the pursuit or category as dependent variables, plots were
trying to get ahead of the pursuit to estab- constructed, regression models were calcu-
lish a roadblock or deploy spike strips. Table lated, and regression lines were applied to
1 displays the death rates per 100,000 the plots. This technique was repeated for
officers for each year in the sample, and by the total death rate, direct pursuit death
each type of death circumstance. rate, blocking activity death rate and travel-
ling to assist death rate. It was believed that
Procedure autocorrelation posed no threat, as none of
The objective of this study was to deter- each year’s deaths was dependent on pre-
mine trends in pursuit-related officer death vious year’s deaths. Substantial year-to-year
rates over a 52-year period in order to test variation was observed, and the death inci-
the research hypothesis that the death rate dents were widely dispersed across the
has been steadily decreasing. Although most nation each year.
longitudinal studies use the autoregressive
integrated moving average (ARIMA) stat-
istical technique, this was not appropriate RESULTS
for the current study. First, ARIMA gener- The overall pursuit-related death rate was
ally requires a minimum of 100 data points examined first. A longitudinal plot of the
(Box, Jenkins, & Reinsel, 1994). Our data, data points was created and inspected for
being annual, had only 52 data points. Even observable trends (Figure 1). As Figure 1
if the data had been converted to monthly shows, there is substantial year-to-year vari-
or weekly periods to increase the number of ation, but visual examination of the plot
data points, accurate calculations would suggested a slightly negative slope, with the
have been difficult because of the significant death rate decreasing over the 52-year
number of weeks or months with no deaths period. To confirm this visual observation, a
reported. Second, ARIMA is best used to least-squares regression line was calculated,

Page 84
Johnson

Table 1: Annual officer death rates per 100,000 officers

Year Total Direct pursuit Blocking activity Travelling to assist

1960 1.06 0.93 0.13 0.00


1961 0.79 0.79 0.00 0.00
1962 1.59 1.46 0.13 0.00
1963 1.46 1.32 0.13 0.00
1964 1.19 1.06 0.13 0.00
1965 1.19 1.06 0.13 0.00
1966 1.99 1.59 0.26 0.13
1967 1.46 1.19 0.26 0.00
1968 1.99 1.99 0.00 0.00
1969 1.32 0.79 0.53 0.00
1970 1.72 1.72 0.00 0.00
1971 0.93 0.93 0.00 0.00
1972 1.59 1.59 0.00 0.00
1973 1.19 0.93 0.26 0.00
1974 1.72 1.32 0.13 0.26

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


1975 1.06 0.93 0.13 0.00
1976 1.85 1.59 0.26 0.00
1977 0.93 0.66 0.26 0.00
1978 1.32 0.93 0.40 0.00
1979 1.06 0.40 0.40 0.26
1980 1.46 0.79 0.66 0.00
1981 0.79 0.40 0.40 0.00
1982 1.06 0.79 0.13 0.13
1983 1.59 0.93 0.40 0.26
1984 1.06 1.06 0.00 0.00
1985 1.06 0.93 0.00 0.13
1986 0.53 0.13 0.40 0.00
1987 1.46 1.19 0.26 0.00
1988 0.79 0.53 0.26 0.00
1989 1.85 1.32 0.40 0.13
1990 0.66 0.53 0.00 0.13
1991 0.93 0.79 0.13 0.00
1992 0.53 0.53 0.00 0.00
1993 0.79 0.66 0.00 0.13
1994 1.19 0.66 0.26 0.26
1995 1.46 1.06 0.26 0.13
1996 0.93 0.40 0.40 0.13
1997 0.40 0.40 0.00 0.00
1998 0.53 0.40 0.13 0.00
1999 1.06 0.93 0.00 0.13
2000 1.06 0.79 0.26 0.00
2001 1.46 0.79 0.53 0.13
2002 1.06 0.66 0.13 0.26
2003 1.99 0.93 0.93 0.13
2004 1.06 0.93 0.00 0.13
2005 1.32 0.79 0.40 0.13
2006 0.53 0.40 0.13 0.00
2007 0.79 0.26 0.13 0.40
2008 0.79 0.66 0.13 0.00
2009 0.53 0.13 0.40 0.00
2010 0.79 0.79 0.00 0.00
2011 1.32 0.40 0.66 0.26

Page 85
Officer deaths from vehicle pursuits

using years as the independent variable and from pursuits decreased by an average of
annual death rate as the dependent variable. 0.572 deaths per 100,000 officers across the
The regression line calculated was statistic- span of the study. This is a decline of
ally significant at p = 0.004, confirming the approximately 66 per cent over 52 years.
presence of a linear relationship between The research hypothesis suggested a neg-
year and death rate. The line that was com- ative slope, and, thus, the analysis of the
puted is superimposed as a dotted line over total pursuit-related officer death rate ap-
the plot in Figure 1. It has a negative slope, peared to support the research hypothesis.
confirming that the overall death rate has Next, the death rate for those killed
generally decreased. The resulting equation while in direct pursuit was investigated.
for this line of best fit was Y = 22.410 + Officers in this category were following
(-0.011) X. behind the fleeing suspect vehicle or
As this equation reveals, each year rammed the suspect vehicle when they
resulted in a mean decrease in the death rate experienced a crash that took their lives.
of only 0.011 per 100,000 officers. Because Some collided with another patrol car, the
the total number of deaths per 100,000 suspect vehicle or an innocent third-party

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


ranged only from 0.40 to 1.99, this slope vehicle. Others lost control and either
was actually quite pronounced. This equa- rolled the vehicle or struck an object, such
tion suggested that the overall death rate as a bridge support or power pole. The same

Figure 1
Pursuit-related death
rate per 100,000 officers,
1960–2011. Dotted line
represents line of best fit

Intercept Coefficient (SE) p

Regression line 22.410 –0.011 (0.004) 0.004

Page 86
Johnson

analysis procedures were used with this sub- pursuit, each year resulted in an average
sample and the total sample. A longitudinal decrease of 0.016 deaths per 100,000
plot was produced (see Figure 2) and exam- officers. This finding strongly supported the
ined visually for evidence of any trends. research hypothesis that vehicle pursuits
Again, year-to-year variation was substan- have been getting progressively safer for
tial, but a general declining trend was officers over the last 52 years. This decline is
clearly visible. A least-squares regression line more pronounced when only those officers
was then calculated to confirm this visual killed while in direct pursuit are
observation and is represented in Figure 2 as examined.
a dotted line. The third analysis involved officers killed
The regression line for this calculation while attempting to assist in the pursuit
was statistically significant at p > 0.000, through a blocking tactic. These tactics
confirming the presence of a linear relation- included establishing a roadblock or
ship between year and death rate. The equa- attempting to lay out spike strips into the
tion for this line of best fit was Y = 32.111 path of the fleeing suspect vehicle. This
+ (–0.016) X. As this equation reveals, the analysis proved to be more complex than

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


slope was negative, and the degree of slope the first two. When the plot was produced
was steeper than the line for all deaths. In and examined, it appeared that the data
the case of officers killed while in direct points began as a positive slope, dropped

Figure 2
Death rate per 100,000
officers in direct pursuit,
1960–2011. Dotted line
represents line of best fit

Intercept Coefficient (SE) p

Regression line 32.111 –0.016 (0.003) <0.000

Page 87
Officer deaths from vehicle pursuits

into a trough, and then began a positive During this first 25-year period, the
slope again. To address this non-linear trend, death rate for officers involved in blocking
multiple regression lines were calculated tactics increased by an average of 0.007
(Cohen, 1988). Starting with the first five deaths per 100,000 officers each year. This
years of data, numerous regression lines was a total increase of approximately 0.175
were calculated, adding one more year at a deaths per 100,000 officers for this first
time in a stepwise manner. Each additional period, a 27.5 per cent increase in officer
year of data decreased the statistical sig- death rate to 1984. Beginning in 1985,
nificance level, and increased the regression these types of officer deaths appeared to
coefficient of the line until the year 1985 decrease sharply, so a second regression line
was reached, when the significance levels was calculated using 1985 as a starting
increased, and the regression coefficients of point. As with Line A, various regression
the lines began to decrease again. The first lines were calculated by adding one year at a
regression line, therefore, spanned the years time. The regression coefficients of each
1960–1984. This line of best fit (Line A) successive line continued to strengthen until
had a positive slope and an equation of the year 1996 was reached. Starting with

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


Y = –13.11 + (0.007) X. 1996, the addition of each successive year

Figure 3
Death rate per 100,000
officers attempting
blocking tactics,
1960–2011. Dotted line
represents line of best fit

Intercept Coefficient (SE) p

Regression line A –13.110 0.007 (0.005) 0.057


Regression line B 43.340 –0.022 (0.014) 0.126
Regression line C –25.339 0.018 (0.015) 0.160

Page 88
Johnson

moved the line further from statistical sig- strength of the regression coefficients of
nificance. As a result, a second line segment, each line calculated increased until the last
Line B, was used for the years 1985–1995, year of the data, 2011, so Line C spanned
when the officer death rate from blocking the 16 years from 1996 to 2011. Line C had
tactics decreased. a positive slope and an equation of Y =
The equation for Line B was Y = 43.340 –25.339 + (0.018) X, indicating an annual
+ (–0.022) X, indicating a negative slope increase in death rate of 0.018 deaths per
and an average annual decrease in death rate 100,000 officers. Lines A, B and C revealed
of 0.022 per 100,000 officers. Over the that, although the death rate from direct
11-year span of this decrease, the death rate pursuits has decreased steadily, the death
dropped by approximately 0.242 deaths per rate from blocking tactics increased steadily
100,000 officers, or approximately 44 per until 1984, dropped sharply from 1985 to
cent. Although this line did not achieve 1995, and then increased again after 1996.
statistical significance at p > 0.005, this This finding does not give support to the
could easily have easily been the result of research hypothesis that death rates for pur-
calculating the line with only 11 cases suits have decreased steadily since 1960.

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


(Cohen, 1988). The final analysis involved officers who
The final line segment, Line C, was died while travelling to assist in a vehicle
calculated in the same manner as the first pursuit. These officers died in crashes while
two lines, but started from 1996. The either trying to catch up with the pursuit or

Figure 4
Death rate per 100,000
officers travelling to
assist in a pursuit,
1960–2011. Dotted line
represents line of best fit

Intercept Coefficient (SE) p

Regression line –4.891 0.003 (0.001) 0.007

Page 89
Officer deaths from vehicle pursuits

trying to get ahead of the pursuit to attempt DISCUSSION


a blocking tactic. Examination of the plot of This study sought to test the hypothesis that
this death rate (see Figure 4), revealed a the changes in emergency medicine, vehicle
steady increase since 1960. In the first safety technology, law, policy, police train-
decade-and-a-half of data, these types of ing and pursuit technology had resulted in a
deaths were very rare, but have increased lower death rate for police officers involved
steadily in later decades. The least-squares in vehicle pursuits. The results found here
line calculated through the data points had were that, in general, this was the case
an equation of Y = –4.891 + (0.003) X. This between 1960 and 2011. The overall death
equation had a positive slope and suggested rate for pursuits declined during this
an annual increase of 0.003 deaths per 52-year period. The decrease in death rate
100,000 officers, meaning that the average was even more substantial for officers spe-
death rate due to travelling to assist in cifically in direct pursuit of the suspect
pursuits increased by approximately 0.156 vehicle. When most people (including
deaths per 100,000 officers during the policy-makers and appellate court justices)

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


52-year period of data. Between 1960 and envisioned involvement in a police pursuit,
2011, the death rate for these types of they probably thought only of an officer in
deaths increased by 15.5 per cent. This direct pursuit, following behind the fleeing
outcome does not support the research suspect. It is possible, therefore, that the
hypothesis that vehicle pursuits have majority of the changes in policies, tactics,
become progressively safer for officers as equipment and laws were focused on mak-
these types of deaths have been gradually ing this sort of pursuit involvement safer
increasing. and less frequent. If this was the case, these
changes and improvements were successful.
To review, examination of the overall
These same changes, however, may have
pursuit-related death rate for police officers
had unintended consequences for other
from 1960 to 2011 revealed a steady
types of pursuit involvement. The death rate
decline, supporting the research hypothesis.
for officers involved in blocking tactics or
Examination of the death rate for officers
travelling to assist in pursuits increased dur-
directly involved in a pursuit also demon-
ing the same 52-year period, and both are
strated a steady decline. This result provided still on the increase today. A review of the
strong support for the research hypothesis. changes discussed above, with a new view
Examination of officer death rate related to regarding these indirect deaths from pur-
attempting to employ a blocking tactic, suits, may suggest potential causes of these
such as a roadblock or spike strips, did not rising death rates. For example, the death
produce a linear trend. The death rate for rate pattern from 1960 to 1984 increased
these types of deaths steadily increased until steadily for officers involved in blocking
reaching a peak in 1984, then plummeted tactics, while at the same time it decreased
until 1996, and began to increase again. steadily for officers in direct pursuit. All of
This result did not support the research the officer blocking-related deaths prior to
hypothesis that death rate was steadily 1985 occurred to officers manning road-
declining. Finally, examination of officer blocks, suggesting that roadblocks either
death rate related to travelling to assist in the steadily became more common or more
pursuit showed a steady increase from 1960 dangerous from 1960 to 1984.
of about 16 per cent. This increase also Perhaps the former was more likely than
failed to support the research hypothesis. the latter, owing to improvements in police

Page 90
Johnson

radio technology. In the early 1960s, many harm to fleeing suspects, strictly limiting
radios in police cars could only receive and the use of roadblocks in police vehicle
not broadcast, while others could broadcast, pursuits instantaneously. Shortly after these
but only to a limited range (National Insti- court decisions, more law enforcement
tute of Justice, 1998b). In the early years of agencies developed pursuit policies or made
the study, therefore, officers in pursuit were existing polices more restrictive (Alpert,
unable to communicate with a dispatcher or 1997; Alpert & Fridell, 1992). It is likely
other patrol cars about the pursuit. As radio these legal decisions (and their resulting
technology improved, patrol car radios policy changes within police agencies) were
developed greater range and signal strength the cause of the sudden decrease in officer
(National Institute of Justice, 1998b). These deaths related to blocking tactics.
improvements gradually enabled officers to As these cases have not been overturned,
communicate to dispatchers about the pur- what has caused the more recent steady
suit, its location and its direction of travel. increase in officer deaths due to blocking
Assisting officers could then be directed to tactics? Since 1995, the death rate from
assist by establishing roadblocks in the path blocking tactics has been increasing. One

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


of the fleeing suspect. possible explanation is the proliferation of
Perhaps, as roadblocks occurred more spike strip devices. Although first patented
frequently, so did officer deaths, as fleeing in the 1970s, spike strip use in pursuits grew
suspects crashed into them or swerved dramatically in the 1990s until the majority
around them, hitting the officers nearby. As of police agencies used them by 1997
the fleeing felon rule still applied during (Alpert, 1997; National Institute of Justice,
this era, when officers could shoot at any 1998a). Correspondingly, almost all block-
fleeing felony suspects in an attempt to ing deaths since 1995 have involved an
prevent their escape (Tennessee v Garner, officer struck while attempting to deploy or
1985), some of these roadblocks also retrieve spike strip devices. Some officers
included gunfire that could have accident- were killed when they were struck as the
ally hit an officer directly or by ricochet. suspect vehicle swerved around the spike
The gradual improvement in communica- strips. Others were struck as they misjudged
tions technology may be the key to the the speed or direction of the suspect, step-
increasing number of officer blocking tactic ping into the vehicle’s path as they deployed
deaths from 1960 to 1984. the strips. Still others were killed by police
What, then, might have caused the death vehicles following the suspect that accident-
rate for blocking tactics to plummet after ally struck the officer as he or she attempted
1984? Radio technology allowing the to retrieve the spike strip from the road.
coordination of roadblocks to continue to Although spike strips have been credited
exist, so what other possible changes could with safely ending high-speed vehicle pur-
have occurred so suddenly in 1985? As was suits (Alpert, 1997; National Institute of
discussed earlier, a number of landmark Justice, 1998a), one unintended conse-
appellate cases involving pursuits were quence may be that officers are more likely
decided between 1985 and 1989. The case to be killed attempting to use these
of Tennessee v Garner (1985) restrained devices.
police authority to shoot at fleeing suspect Finally, what might explain the gradual
vehicles, or ram them, under most circum- increase in death rate for officers killed
stances. Jamieson v Shaw (1985) discon- while travelling to assist with pursuits?
tinued the use of roadblocks that might Although the death rate for officers in
potentially cause death or serious bodily direct pursuit declined, rate of death has

Page 91
Officer deaths from vehicle pursuits

increased in officers travelling to assist in officers engage in these activities, the risk
these same pursuits. Perchance the answer likely increases.
is, again, linked to improved radio com- The results of this study have implica-
munication. As radio technology has gradu- tions for policy and practice. First, it sug-
ally improved over the decades, so has the gested that the technological and policy
ability to coordinate support from other advancements implemented in past decades
police vehicles to either join in the pursuit have worked to reduce the risk of death
or establish a blocking tactic. In the early from pursuits. Restrictive policies that limit
1960s, the existing radio technology often the number of pursuits, safer vehicles, better
did not permit pursuing officers to alert a emergency medical technology, and the use
dispatcher about the pursuit. Gradually, of helicopters have steadily driven down the
officers could communicate to the dis- death rate for officers in direct pursuit.
patcher but, as a result of weak radio signals, Legal decisions severely limiting the use of
other patrol cars could not hear about the firearms and roadblocks also demonstrated a
pursuit unless they were nearby (National quick decline in death rate for officers man-
Institute of Justice, 1998b). The dispatcher ning roadblocks. These policy changes

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


had to relay information to the other achieved their intended goals.
officers, which delayed the time taken for Second, the findings suggest that new
the information to reach assisting officers. policy changes may be in order for officers
Correspondingly, between 1960 and 1973, assisting in pursuits. The recent increase in
only one recorded death involved an officer officer death rate for those deploying spike
killed while travelling to assist in a pursuit. strips warrants a closer look at how these
Throughout the 1970s and 1980s, police devices are used. These deaths need to be
car radio signal strength increased, as did the closely examined to determine what tactical
officer death rate for those assisting in pur- changes need to be made in how they are
suits this way. Finally, by the late 1990s and deployed. Safe and unsafe techniques for
2000s, in-car computers with GPS systems deploying these devices need to be identi-
had become commonplace, permitting fied, and written policies developed in
officers to track the location of the pursuit accordance with the safe techniques.
without the use of a radio (National Insti- Officers should also be properly trained in
tute of Justice, 1998b). These changes the safe use of these devices. Similarly,
correspond with the steady increase in departmental pursuit polices need to address
officer death rate for those travelling to the behaviour of officers not directly
assist in pursuits. It is likely that, because involved in the pursuit. Perhaps the number
more officers are aware of the ongoing of assisting vehicles should be strictly
pursuit, and its exact location and direction limited, with assisting units clearly identi-
of travel, more officers have been travelling fied and coordinated by a supervisor. Other
to assist in the pursuit, thus increasing the units should be prohibited from involve-
likelihood of a crash. Because vehicle pur- ment in the pursuit. Pursuit policies should
suits are usually at high speeds, assisting also place limits on the speed and disregard
vehicles may often have to travel at even for traffic regulations that these assisting
greater speeds to catch up to the pursuit or officers can exhibit while travelling to
get ahead of it to deploy spike strips. Travel- assist.
ling at high speeds, while simultaneously This study revealed that police pursuits
monitoring the location of the pursuit on have generally become safer over the last
the radio or via computer terminal, are half-century. It revealed that policies and
risky activities in themselves. As more innovations directed at improving vehicle

Page 92
Johnson

safety and improving pursuit safety have Edwards, T. D. (1993). State police basic
worked. It also revealed, however, that new training programs: an assessment of course
threats have developed as a by-product of content and instructional methodology.
the technological improvements in radio American Journal of Police, 12, 23–45.
and spike strip technology. Further research Gallas v. McKee, 801 F.2d 200 (1986).
is warranted in these two increasing danger Haberfeld, M. R. (2002). Critical issues in police
training. Upper Saddle River, NJ: Prentice
areas, to inform police training, policy and
Hall.
practice better.
Halman, S. I., Chipman, M., Parkin, P. C., &
Wright, J. G. (2002). Are seat belt restraints
as effective in school age children as in
REFERENCES
adults? A prospective crash study. British
Alpert, G. P. (1997). Police pursuit: Policies and Medical Journal, 324, 1123–1125.
training. Washington, DC: National Institute Hill, J. (2002). High-speed police pursuits:
of Justice. dangers, dynamics, and risk reduction. FBI
Alpert, G. P. (1998). Helicopters in pursuit Law Enforcement Bulletin, 71(7), 1–5.
operations. Washington, DC: National Jamieson v. Shaw, 776 F.2d 1048 (1985).

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


Institute of Justice. Levitt, S. D., & Porter, J. (2001). Sample
Alpert, G. P., Dunham, R. G., & Stroshine, selection in the estimation of air bag and
M. S. (2006). Policing: Continuity and change.
seat belt effectiveness. Review of Economics
Long Grove, IL: Waveland Press.
and Statistics, 83, 603–615.
Alpert, G. P., & Fridell, L. A. (1992). Police
Lilley, D., & Boba, R. (2008). A comparison
vehicles and firearms: Instruments of deadly force.
of outcomes associated with two key law
Prospect Heights, IL: Waveland Press.
enforcement grant programs. Criminal Justice
Becknell, C., Mays, G. L., & Giever, D. M.
Policy Review, 19, 438–465.
(1999). Policy restrictiveness and police
Matheny, T. L. (1996). Entry level police
pursuits. Policing: An International Journal of
training: Evaluating the standard.
Police Strategies and Management, 22, 93–110.
Bower v. Inyo County, 109 S.Ct. 1378 (1989). (Unpublished master’s thesis). Jersey City
Box, G. E. P., Jenkins, G. N., & State College, Jersey City, NJ.
Reinsel, G. C. (1994). Time series analysis: McManus, G. P. (1970). Police training and
Forecasting and control. Englewood Cliffs, NJ: performance study. Washington, DC: Law
Prentice Hall. Enforcement Assistance Administration.
Braver, E. R., Ferguson, S. A., Greene, M. A., National Highway Traffic Safety
& Lund, A. K. (1997). Reductions in deaths Administration. (1998). Traffic safety facts,
in frontal crashes among right front 1997. Washington, DC: National Highway
passengers in vehicles equipped with Traffic Safety Administration.
passenger airbags. Journal of the American National Highway Traffic Safety
Medical Association, 278, 1437–1439. Administration. (2011). Characteristics of law
City of Canton, Ohio v. Harris, 109 S.Ct. enforcement officers’ fatalities in motor vehicle
1197 (1989). crashes. Washington, DC: National Highway
Cohen, J. (1988). Statistical power analysis for the Traffic Safety Administration.
social sciences. Hillsdale, NJ: Erlbaum. National Highway Traffic Safety
Cohen, J., & Cohen, P. (1983). Applied multiple Administration. (2013). Fatality analysis
regression/correlation analysis for the behavioral reporting system. Washington, DC: National
sciences. Hillsdale, NJ: Erlbaum. Highway Traffic Safety Administration.
Concerns of Police Survivors. (2013). Officer- National Institute of Justice. (1998a). Pursuit
Down Memorial Page. Retrieved May 30, management task force. Washington, DC:
2003 from http://www.odmp.org National Institute of Justice.
County of Sacramento v. Lewis, S.Ct. No. National Institute of Justice. (1998b). The
96-1337 (1998). evolution and development of police technology.

Page 93
Officer deaths from vehicle pursuits

Washington, DC: National Institute of American Journal of Public Health, 66,


Justice. 859–864.
National Research Council. (2007). Emergency Sanow, E. J. (2002). Encyclopedia of American
medical services: Future of emergency care. police cars. Chicago, IL: Crestline.
Washington, DC: National Academies Press. Sullivent, E. E., Faul, M., & Wald, M. M.
Pell, J. P., Sirel, J. M., Marsden, A. K., Ford, (2011). Reduced mortality in injured adults
I., & Cobbe, S. M. (2001). Effect of transported by helicopter emergency
reducing ambulance response times on medical services. Prehospital Emergency Care,
deaths from out of hospital cardiac arrest: 15, 295–300.
cohort study. British Medical Journal, 323, Tennessee v. Garner, 471 U.S. 1 (1985).
1365–1368. U.S. Patent and Trademark Office. (2012). U.S.
Pipes, C., & Pape, D. (2001). Police pursuits
patent search. Retrieved May 30, 2013
and civil liability. FBI Law Enforcement
from http://www.uspto.gov/patents/
Bulletin, 70(7), 16–21.
process/search/
Pons, P. T., & Markovchick, V. J. (2002). Eight
minutes or less: does the ambulance Vanderbilt, T. (2009). Traffic. New York, NY:
response time guideline impact trauma Vintage.
Worrall, J., & Zhao, J. (2003). The role of

Downloaded from psm.sagepub.com at MOUNT ALLISON UNIV on June 29, 2015


patient outcome? Journal of Emergency
Medicine, 23, 43–48. COPS office in community policing.
Rivara, F. P., & Mack, C. D. (2004). Motor Policing: An International Journal of Police
vehicle crash deaths related to police Strategies and Management, 26, 64–87.
pursuits in the United States. Injury Zador, P. L., & Ciccone, M. A. (1993).
Prevention, 10, 93–95. Automobile driver fatalities in frontal
Robertson, L. S. (1976). Estimates of motor impacts: air bags compared with manual
vehicle seatbelt effectiveness and use: belts. American Journal of Public Health, 83,
implications for occupant crash protection. 661–665.

Page 94

You might also like