TRIP-100073; No of Pages 7
Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

Contents lists available at ScienceDirect

Transportation Research Interdisciplinary Perspectives
journal homepage: https://www.journals.elsevier.com/transportation-researchinterdisciplinary-perspectives

Does news coverage of traffic crashes affect perceived blame and preferred
solutions? Evidence from an experiment
⁎

Tara Goddard a, , Kelcie Ralph b, Calvin G. Thigpen b, Evan Iacobucci b
a
b

Department of Landscape Architecture and Urban Planning, Texas A&M University, 3134 TAMU, College Station, TX 77843, USA
Edward J. Bloustein School of Planning and Public Policy, Rutgers, The State University of New Jersey, 33 Livingston Avenue, New Brunswick, NJ 08901, USA

A R T I C L E

I N F O

Article history:
Received 9 August 2019
Received in revised form 21 October 2019
Accepted 23 October 2019
Available online xxxx

A B S T R A C T

Traffic crashes are one of the leading causes of preventable death in the United States. Nearly 20% of these roadway fatalities are people who were hit and killed while walking or bicycling. Pedestrian deaths increased by 46% between 2009
and 2016. Despite these troubling statistics, there has not been a sustained and widespread public call to action to improve
road safety. Researchers and advocates are increasingly focused on how traffic crashes are reported in the media, and
whether existing editorial patterns contribute to victim-blaming and distract from systems-level solutions. However, no
previous study has examined whether editorial patterns in traffic crash coverage actually influence perceptions. This
study conducted an experiment in which 999 subjects were randomly assigned to read one of three versions of a news article describing a traffic crash involving a pedestrian. After reading the description, subjects were asked to apportion
blame, identify an appropriate punishment for the driver, and assess various approaches for improving road safety. In
comparing the three groups, even relatively subtle differences in editorial patterns significantly affected readers' interpretation of both what happened and what to do about it on nearly every measure. Shifting from pedestrian-focused to driverfocused language reduced victim-blaming and increased perceived blame for the driver. A thematic frame significantly
increased support for infrastructure improvements. This study provides strong evidence that efforts to change public perceptions of road safety should include a focus on improving editorial patterns in traffic crash reporting.

1. Introduction
Traffic crashes are one of the leading causes of preventable death in the
United States. Nearly a fifth (18%) of these roadway fatalities is people who
were hit and killed while walking or bicycling (National Highway Traffic
Safety Administration, 2016). Worse still, pedestrian deaths increased by
46% between 2009 and 2016 (Hu and Cicchino, 2018). Despite these troubling figures, there has not been a sustained and widespread public call to
action to improve road safety.
One plausible reason for the lack of widespread public concern may be
how these traffic crashes are reported in the media. The field of media studies has consistently demonstrated that news coverage meaningfully shapes
public perceptions (McCombs and Shaw, 1972; Iyengar, 1994). Seemingly
trivial editorial differences have been shown to influence how readers
make sense of a story (Iyengar, 1994, 1996). Critically, these changes also
influence how readers would like policymakers to respond, in turn influencing which policies are implemented (Boykoff and Boykoff, 2007; Hinchcliff
et al., 2010; Jacoby, 2000).
⁎ Corresponding author.
E-mail addresses: goddard@tamu.edu, (T. Goddard), kelcie.ralph@ejb.rutgers.edu, (K. Ralph),
thigpen.calvin.g@gmail.com, (C.G. Thigpen), evan.iacobucci@rutgers.edu. (E. Iacobucci).

http://dx.doi.org/10.1016/j.trip.2019.100073
2590-1982/©2019 The Authors. Published by Elsevier Ltd. This is an open access
article
under
the
CC
BY-NC-ND
license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Researchers and advocates are increasingly focused on whether these
editorial patterns may help explain the lack of concern about traffic deaths.
Thus far, scholars have cataloged the editorial patterns in pedestrianinvolved crash coverage over a one-year period in a Canadian city
(Magusin, 2017), in 200 local news articles of pedestrian and bicyclist
crashes from throughout the United States (Ralph et al., 2019), and
media coverage of bicyclist fatalities over a ten-year period in Florida
(Scheffels et al., 2019). These three studies collectively demonstrate widespread imbalanced editorial and linguistic patterns, which media and psychology studies identify as contributing to victim-blaming and distracting
from systematic issues and solutions.
However, no previous study has examined whether editorial patterns in
traffic crash coverage actually influence perceptions. This paper fills that gap.
Specifically, it asks three research questions:
1) Do editorial patterns affect how readers apportion blame for a crash?
2) Do editorial patterns affect what punishment readers think is appropriate, and/or how severe that punishment should be?
3) Do editorial patterns influence readers' support for various solutions for
improving road safety?
To answer these questions, an experiment was conducted. Subjects were
randomly assigned to read one of three versions of a news article describing

 T. Goddard et al.

Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

Table 1
Editorial patterns in crash coverage.
Status quo

Better practice

Effect of status quo

Accident: “Pedestrian killed in accident on Main Street.”
Non-agentive: “A pedestrian was hit and killed.” (no agent)
Focus on pedestrian: “A pedestrian was hit and killed by a car.”
Object-based language: “A car jumped the curb.”
Counterfactual statements: “The pedestrian darted into the street.”
Episodic framing: Treats the crash as an isolated incident.

Crash: “Pedestrian killed in crash on Main Street.”
Agentive: “A pedestrian was hit and killed by a car.”
Focus on vehicle: “A car hit and killed a pedestrian.”
Person-based language: “A driver drove over the curb.”
Not included
Thematic framing: “This is the tenth fatal collision this year.”

Obscures preventable nature of crashes
Obscures role of a human actor
Increases blame for the focus of the sentence
Obscures role of a human actor
Increases perceived blame for the victim
Prevents readers from connecting the dots between
incidents and thus shifts attention to individual-level
rather than systematic solutions

agent obfuscate the perpetrator's role in the action, decreasing apparent
blameworthiness (Coates and Wade, 2004). The use of non-agentive language is so powerful that it can even affect how people recall an event
that they actually observed takes place (Fausey et al., 2010).

a traffic crash involving a pedestrian. After reading the description, subjects
were asked to apportion blame, identify an appropriate punishment for the
driver, and assess various approaches for improving road safety. In comparing the outcomes across the three groups, we find that editorial patterns significantly affected readers' interpretation of both what happened and what
to do about it on nearly every measure.

2.1.3. Focus
The concept of focus describes the way that a sentence communicates
who (or what) is the center of attention. “Sarah opened the door” focuses
on Sarah; it about what Sarah is doing. “The door was opened by Sarah” focuses on the door; it is a sentence about something that happened to the
door.2 In crash coverage, journalists overwhelmingly tend to focus on the
victim. Ralph et al. found that victims were the focus of sentences 73% of
the time (Ralph et al., 2019). By contrast, vehicles and drivers were the
focus just 13% and 11% of the time. Focus is important because readers
are more likely to attribute blame to the focus of a sentence (Niemi and
Young, 2016).

2. Background
Media studies is a field of scholarly inquiry that focuses on how media
coverage influences our collective understanding of issues (Marris and
Thornham, 2000). This body of work consistently finds that the media exerts influence over public perception (Cobb and Primo, 2004; Frost et al.,
1997; Harcup, 2001; Jeanis and Powers, 2017). For instance, “agenda-setting” refers to the media's ability to influence which issues are discussed
and which are not (McCombs and Shaw, 1972). Novel or shocking events
and those that involve ideal victims—children, the elderly, or people otherwise likely to be unquestioningly accepted as victims—tend to receive
much more coverage than events that seem regular and mundane
(Christie, 1986; Jeanis and Powers, 2017). This disproportionate coverage
on “newsworthy” events heightens public interest in those events and
dampens interest in others, thereby subtly shifting public opinion and altering the demands on policymakers (Harcup, 2001). For instance, airline
crashes receive much more media attention than road crashes, even though
the latter cause vastly more injuries and deaths (Cobb and Primo, 2004).

2.1.4. Object-based and person-based language
When authors ascribe agency to inanimate objects, they are using
object-based language. In crash coverage, object-based language feels fairly
natural (e.g. “A car hit a person”), yet in other settings it sounds strange indeed (“A hammer hit a person”). By keeping the actor hidden, object-based
language obscures blame. When rewritten with person-based language,
these sentences clearly identify the responsible party (“A man hit a person
with a car/hammer”).
Object-based language is widespread in traffic crash coverage. Ralph
et al. found that articles referred to a vehicle 81% of the time (when an
agent was identified at all) and referred to a driver just 19% of the time
(Ralph et al., 2019). Moreover, writers use object-based language unevenly.
They tend to describe people in cars using object-based language (e.g. “car”
or “traffic”), but use human-based language when describing VRUs (e.g.
“bicyclist” or “pedestrian”) (Ralph et al., 2019). A study of driver's perception of other roadway users found this same imbalance in the language people used to describe roadway interactions (Walker, 2005). This practice
portrays agency in an uneven way and may implicitly assign greater responsibility to the VRU.

2.1. Editorial patterns of interest
Table 1 summarizes six editorial patterns of interest, each of which have
been shown to influence perceptions and are widely used in crash coverage.
2.1.1. “Accident”
The specific terms used to refer to a crash can influence how readers interpret what happened. For instance, the British Medical Journal banned
use of the term “accident” in 2003 because it conveys an undue sense of
faultlessness and inevitability (Davis and Pless, 2001). In 2016, the Associated Press Style Guide changed to encourage journalists to use “crash, collision, or other terms” instead of “accident” (Transportation Alternatives,
2016) Eliminating this term would be more in line with the nature of traffic
crashes, which are “predictable and preventable” (Davis and Pless, 2001).
Nevertheless, the use of “accident” remains widespread, with one study
finding the term in 47% of news articles about VRU crashes (Ralph et al.,
2019).

2.1.5. Counterfactual statements
Other kinds of statements imply that a victim could have avoided death
or injury if they had behaved differently. These “counterfactuals” refer to
an implied or perceived possible alternative sequence of events where a
change in circumstances prior to an event would make the outcome of
that event different (Goldinger et al., 2003). In crash coverage, journalists
sometimes include details about the actions or characteristics of the pedestrian—that they were crossing outside of a marked crosswalk, wearing dark
clothing, or darting into the roadway. These types of counterfactual statements occur in roughly one-third of articles about crashes (Ralph et al.,
2019). The widespread inclusion of counterfactuals is troubling because
they are known to shape how readers perceive blame. For instance,

2.1.2. Agentive and non-agentive language
The presence or absence of an agent informs how readers perceive
blame. “Allison spilled the milk” uses agentive language, while “The milk
spilled” describes the same event in a non-agentive way.1 When it comes
to crash coverage, non-agentive sentence are common. In a recent study,
one third of the sentences describing VRU crashes were non-agentive
(Ralph et al., 2019). This is troubling because sentences that omit an

2
Note that focus also differs from active/passive-voice. “Jocelyn's wallet was stolen” is an
example of passive voice. “Natalie stole Jocelyn's wallet” and “Jocelyn had her wallet stolen
by Natalie” both use active voice, but focus on different actors: in the first, the victim, and in
the second, the perpetrator.

1
This linguistic device is distinct from passive voice: “The milk spilled” uses active voice,
but the agent that did the spilling has been omitted.

2

 T. Goddard et al.

Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

United States at large on these three dimensions. Respondents were paid
$1.25 for their participation, which was set by estimating a 10-minute participation at Prolific.ac's rate of $7.50/h. 999 participants were recruited to
attain a statistical power threshold of 0.8 for the three-group, betweensubjects experimental design employed here.

including counterfactuals in a description of a rape increases reader's tendency to blame the victim (Branscombe et al., 1996). Another study showed
that participants serving as hypothetical jurors awarded less money to victims whose incidents were described with a counterfactual (Goldinger
et al., 2003). Importantly, evidence suggests that this kind of statements induce counterfactual thinking automatically (Goldinger et al., 2003), without needing the alternative sequence of events and outcome spelled out
explicitly.
Notably, the accuracy of the counterfactual statements is not at issue (although comments about the actions of the victim may be subject to
survivor's bias (Smith, 2014). Rather it is that details about the VRU tend
to get included, while others—like the speed of the driver—are omitted.
By asymmetrically drawing our attention, crash coverage subtly shifts
blame towards VRUs.

3.2. Experimental treatments
The survey began with a brief description of a traffic crash. Participants
were randomly assigned to see one of three versions of the text (Table 2).
Version #1 (the pedestrian-focused text) is typical of local new coverage
of crashes. It refers to the incident as an “accident”, uses non-agentive language, focuses on the victim, uses object-based language, and includes a
counterfactual about the victim. Version #2 (the driver-focused text) describes the same crash but has been rewritten slightly. It refers to the incident as a “crash”, uses agentive language, focuses on the driver, uses
person-based language, and does not include a counterfactual. Version #3
(the thematically-themed text) includes all of the previously mentioned
changes as well as contextualizing information and thematic framing.

2.1.6. Thematic v. episodic framing
Iyengar, a political science and communications expert, distinguishes
between two types of framing in news articles: episodic and thematic
(Iyengar, 1994; Scheufele, 1999). Episodic framing portrays events as isolated incidents, whereas thematic framing connects the dots for readers,
providing context and identifying other events that are part of a pattern.
Generally speaking, readers who encounter episodic framing tend to attribute blame to individuals and tend to support individual-level solutions
(Iyengar, 1994). By contrast, readers who encounter thematic framing are
more likely to call for broader, systematic changes. Thematic framing remains rare in coverage of car crashes, occurring in just 6% of news stories
in one study (Ralph et al., 2019) and in 29.6% of stories in another
(Scheffels et al., 2019).

3.3. Variables
After reading one of the randomly assigned news articles, participants
answered several questions grouped into three topics: apportioning
blame, selecting appropriate punishments, and identifying preferred solutions for improving safety.
Table 2
Article text.

2.2. Hypotheses
In sum, there are a number of editorial patterns that have been shown
across multiple fields to influence reader perceptions. Moreover, these patterns are widespread in crash overage. While editorial patterns may shape
public perceptions about crash coverage, it has yet to be confirmed. This
project aims to fill that gap.
This project tests the hypothesis that relatively minor changes in the description of a crash will measurably affect how readers perceive blame, select appropriate punishments, and identify preferred solutions. Specifically,
this experiment hypothesizes that:

Version #1
Pedestrian-focused

Version #2
Driver-focused

Pedestrian struck and
killed on east side

Driver hits, kills
pedestrian on east side

Version #3
Thematically-framed

Driver hits, kills
pedestrian on east side as
pedestrian deaths
continue to increase
city-wide
Police are investigating a
Police are investigating a
Police are investigating a
fatal crash on the east side fatal crash on the east side
fatal accident on the
Saturday evening. Shortly
Saturday evening. Shortly
east side Saturday
evening. Shortly before 7 before 7 p.m., police were before 7 p.m., police were
called to the 2500 block of called to the 2500 block of
p.m., police were called
North Main Street on a
North Main Street on a
to the 2500 block of
report of a driver striking report of driver striking a
North Main Street on a
a pedestrian. Police say a pedestrian. Police say a
report of a pedestrian
46-year-old man was
46-year-old man was
struck by a car. Police
attempting to cross Main
attempting to cross Main
say a 46-year-old man
Street between a bus stop
wearing dark clothing Street when the driver
struck him with his car.
was attempting to cross
and the Walgreens when
The injured pedestrian was the driver struck him
Main Street when he
was struck. The injured unresponsive at the scene
with his car. The injured
and reportedly died on the pedestrian was
pedestrian was
way to the hospital. Police unresponsive at the scene
unresponsive at the
say the driver was not
scene and reportedly
and reportedly died on the
impaired and remained at
died on the way to the
way to the hospital. Police
the scene.
hospital. Police say the
say the driver was not
driver was not impaired
impaired and remained at
and remained at the
the scene.
scene.
This is the eighth death
of a pedestrian in the city
this year, an increase of
20% from last year at this
time. Three pedestrians
have died on this stretch
of Main Street, which,
despite being a busy
shopping area, has high
traffic speeds and a lack
of streetlights.

1) rewriting the text to focus on the driver will reduce the pedestrian's perceived blame and increase the driver's;
2) shifting to driver-focused text will increase support for punishing the
driver; and
3) thematic framing will lead readers to increase their support for pedestrian infrastructure and decrease their support for educational campaigns that focus on the actions of the pedestrian.
3. Methods
To test these hypotheses, a survey instrument was developed with three
experimental conditions. Experiments are widely used to study the effects
of linguistic patterns, including counterfactuals (Branscombe et al., 1996;
Goldinger et al., 2003), agentive versus non-agentive language (Fausey
et al., 2010; Fausey and Boroditsky, 2010), and focus (Niemi and Young,
2016).
3.1. Sampling
Participants were recruited through Prolific.ac, an online data collection tool with a pool of over 35,000 adult US citizens (Palan and Schitter,
2018). Using this research-ready online participant recruitment tool allows
for rapid data collection of a large sample with a more diverse population
than a convenience sample. Prolific.ac stratifies its participants on age,
race, and gender. The sample used in this study is representative of the

Note: Words or phrases that differ between article types are denoted in bold for
demonstrations purposes. Thematic elements in article #3 are denoted with underline. In the actual experiment, all three articles had plain text.
3

 T. Goddard et al.

Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

Fig. 1. Shifting focus and frame significantly affects how people apportion blame in a hypothetical traffic crash.

test this, one-way ANOVA and chi-squared tests were run for age, sex,
race, and travel mode (Table 2). While the three groups were broadly similar, they did differ with respect to age and extent of driving. The statistical
models below account for these important differences between the experimental groups.
To test the hypotheses, a series of linear regression models were estimated for each outcome variable. In all cases, the independent variable of
interest was the experimental group (i.e. did the respondent read Version
#1, #2, or #3). To test for differences in the coefficients for the experimental group variables, a linear hypothesis tests was performed using the finitesample F statistic (with approximate F distribution).

3.3.1. Apportioning blame
Participants were asked to assign responsibility for the crash to “the
driver of the car”, “the person walking”, or “other”. Responses were required to sum to 100%. Respondents who indicated “other” were asked
to provide details via a text box.
3.3.2. Selecting appropriate punishments
Next, respondents were instructed to imagine that the driver described
in the article was on trial and to indicate the type of punishment that would
be appropriate: community service, a fine, jail time, vehicle impounded, or
no punishment. Respondents could select more than one punishment (unless they selected “no punishment”).
In addition, respondents were asked to indicate the severity of punishment a judge should administer (e.g. “If the judge sentenced the driver to
PAY A FINE, in your opinion how much should the fine be in dollars?”).
This question was repeated for hours of community service, dollar value
of a fine, and jail time in years.

4. Results
4.1. Q1: apportioning blame
Editorial patterns significantly influenced how readers apportioned
blame for the hypothetical crash (Fig. 1, Table 3). Shifting from the
pedestrian-focused text (Version #1) to the driver-focused text (Version
#2) caused readers to assign 30% more blame to the driver and 30% less
blame to the pedestrian.
Shifting to the thematically framed text also influenced how readers apportioned blame. Relative to readers of Versions #1 and #2, readers who
encountered the thematic frame were less likely to blame the pedestrian
and much more likely to blame “other” factors (about 150% more than Version #2 and over 250% more than Version #1).

3.3.3. Assessing preferred solutions
The survey included two metrics for assessing preferred solutions. First,
respondents were asked to indicate their level of support for three options
for improving safety on the hypothetical Main Street:
1) a Walk Safe! Campaign to train pedestrians to cross the street more carefully;
2) increase enforcement of traffic laws, especially speeding, distracted
driving, and reckless driving; and
3) improve safety by widening the sidewalks, adding a new crosswalk, and
lowering the speed limit.

4.2. Q2: selecting appropriate punishments

Respondents indicated their support using a five-point scale ranging
from “strongly oppose” to “strongly support”. Second, respondents were
asked to assess two proposed alternatives for a newly constructed roadway.
The options were modeled after similar work by Mouter et al. and were designed to force participants to trade off two competing outcomes: higher vehicle speeds and fewer traffic deaths (Mouter et al., 2018).

Editorial patterns also shaped readers' perceptions about punishments
(Table 4). Whereas nearly one third of those who read the pedestrianfocused text thought the drivers should not be punished, just 19% of
those who read the driver-focused text agreed. Specifically, readers of the
driver-focused text were more likely to support jail time, a vehicle impound, and a fine. Moreover, these readers supported more severe punishments for drivers (30% more hours of service, 30% higher fines, and double
the amount of jail time).
Shifting to a thematically-framed text also meaningfully altered opinions about punishment. Relative to those who read the pedestrian-focused
text, readers who encountered the thematically framed text were more
likely to approve punishment for drivers (jail time and vehicle impound)
and approved stricter punishments (24% more service, 22% higher fines,
and double the amount of jail time).
Finally, there was no difference in punishment between the driverfocused and thematically-framed texts.

3.3.4. Other personal information
In addition to the outcome variables of interest, participants also indicated how often they drive, ride in a car as a passenger, walk, or bicycle in
a typical week. Demographics reported by the recruitment company included
age group, gender, and the simplified race categories used in the US Census.
3.4. Statistical testing
Since experimental conditions were randomly assigned, the three
groups should be relatively similar in terms of personal characteristics. To
4

 T. Goddard et al.

Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

Table 3
Personal characteristics and travel behaviors of the experimental groups.

Age (mean)
Mean (SD)
Min - max
Gender: female
Race
White
Black
Asian
Mixed
Other
Typical mode
Drive
1–2 times
3–4 times
5–6 times
7+ times
Never
Passenger
1–2 times
3–4 times
5–6 times
7+ times
Never
Walk
1–2 times
3–4 times
5–6 times
7+ times
Never
Bike
1–2 times
3–4 times
5–6 times
7+ times
Never

#1 – pedestrian focused
(N = 334)

#2 – driver focused
(N = 332)

#3 – thematically framed
(N = 333)

Total
(N = 999)

45.4 (16.0)
18.0–79.0
157 (47.0%)

44.1 (15.5)
18.0–77.0
182 (54.8%)

42.5 (16.1)
18.0–76.0
173 (52.0%)

44.0 (15.9)
18.0–79.0
512 (51.3%)

267 (79.9%)
37 (11.1%)
21 (6.3%)
4 (1.2%)
5 (1.5%)

238 (71.7%)
43 (13.0%)
26 (7.8%)
15 (4.5%)
10 (3.0%)

248 (74.5%)
48 (14.4%)
19 (5.7%)
9 (2.7%)
9 (2.7%)

753 (75.4%)
128 (12.8%)
66 (6.6%)
28 (2.8%)
24 (2.4%)

39 (11.7%)
38 (11.4%)
73 (21.9%)
127 (38.0%)
57 (17.1%)

48 (14.5%)
55 (16.6%)
48 (14.5%)
127 (38.3%)
54 (16.3%)

37 (11.1%)
51 (15.3%)
48 (14.4%)
120 (36.0%)
77 (23.1%)

124 (12.4%)
144 (14.4%)
169 (16.9%)
374 (37.4%)
188 (18.8%)

150 (44.9%)
62 (18.6%)
24 (7.2%)
16 (4.8%)
82 (24.6%)

150 (45.2%)
44 (13.3%)
25 (7.5%)
24 (7.2%)
89 (26.8%)

145 (43.5%)
63 (18.9%)
32 (9.6%)
24 (7.2%)
69 (20.7%)

445 (44.5%)
169 (16.9%)
81 (8.1%)
64 (6.4%)
240 (24.0%)

116 (34.7%)
46 (13.8%)
22 (6.6%)
38 (11.4%)
112 (33.5%)

113 (34.0%)
38 (11.4%)
20 (6.0%)
44 (13.3%)
117 (35.2%)

122 (36.6%)
37 (11.1%)
26 (7.8%)
50 (15.0%)
98 (29.4%)

351 (35.1%)
121 (12.1%)
68 (6.8%)
132 (13.2%)
327 (32.7%)

32 (9.6%)
8 (2.4%)
6 (1.8%)
5 (1.5%)
283 (84.7%)

32 (9.6%)
8 (2.4%)
5 (1.5%)
5 (1.5%)
282 (84.9%)

24 (7.2%)
14 (4.2%)
3 (0.9%)
10 (3.0%)
282 (84.7%)

88 (8.8%)
30 (3.0%)
14 (1.4%)
20 (2.0%)
847 (84.8%)

p value
*

n.s.
n.s.

**

n.s.

n.s.

n.s.

Note: One-way ANOVA and chi-squared tests, at statistical significance levels: *** = p < 0.01, ** = p < 0.05, * = p < 0.1.

The other results were inconsistent across groups (Table 4). For instance,
the thematically-framed text increased support for enforcement and willingness to trade off slower speeds for fewer deaths, but only relative to the
pedestrian-focused text (i.e. there was no difference relative to the driverfocused text). Similarly, the thematically-framed text decreased support for
a pedestrian safety campaign, but only relative to the driver-focused text.
There was no difference in preferred approaches between the
pedestrian-focused and driver-focused texts (not shown).

4.3. Q3: assessing preferred solutions
The use of thematic framing significantly influenced readers' preferred
approaches for improving safety (Table 5). Shifting to a thematicallyframed text clearly increases support for pedestrian infrastructure (relative
to both the pedestrian- and driver-focused texts) (Fig. 2). This result is consistent with the work of Iyengar who found that thematic framing increases
demand for systematic public policy solutions (Iyengar, 1994, 1996).

Table 4
Influence of editorial patterns on assessment of blame and punishment.
#1
Ped-focused

#2
Driver-focused

#3
Thematic frame

Ped v. driver
1 v. 2

Ped v. thematic
1 v. 3

Driver v. thematic
2 v. 3

50.2 (27.2)
43.1 (26.7)
6.7 (18.7)
100

65.5 (25.0)
30.0 (23.6)
4.5 (13.4)
100

54.9 (27.6)
28.6 (24.7)
16.5 (22.1)
100

***
***
n.s.

*
***
***

***
n.s.
***

Type of punishment (% approve)
None
31.4
Community service
46.1
A fine
38.6
Jail time
23.4
Vehicle impound
13.2

18.7
47.6
46.7
37.3
22.9

22.2
46.2
44.4
37.2
19.8

***
n.s.
*
***
***

**
n.s.
n.s.
***
n.s.

n.s.
n.s.
n.s.
n.s.
n.s.

Severity of punishment (mean)
Hours of service
101.4
Amount of fine
$535.30
Years of jail time
1.03

132.2
$698.10
2.05

125.7
$653.30
2.08

***
***
***

***
***
***

n.s.
*
n.s.

Share of blame (mean (SD))
The driver
The person walking
Other

Note: Using a linear hypothesis test, with the finite-sample F statistic (with approximate F distribution), at statistical significance levels: *** = p < 0.01, ** = p < 0.05, * = p
< 0.10. Models control for age and extent of driving.
5

 T. Goddard et al.

Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

Table 5
Influence of editorial patterns on preferred approaches for improving safety.

Share who support
Wider sidewalks, marked crosswalk, lowered speed limit
A Walk Smart! Campaign to “train pedestrians to cross the street more carefully”
More enforcement for drivers, including speeding and distracted driving
Share willing to trade lower speeds for fewer traffic deaths

#1
Ped-focused

#2
Driver-focused

#3
Thematic frame

Ped v. thematic
1 v. 3

Driver v. thematic
2 v. 3

85.3
76.9
84.4
88.3

85.2
80.7
88.3
91.6

93.4
75.1
89.2
94.3

***
n.s.
*
***

***
*
n.s.
n.s.

Note: Using a linear hypothesis test, with the finite-sample F statistic (with approximate F distribution), at statistical significance levels: *** = p < 0.01, ** = p < 0.05, * = p
< 0.10. Models control for age and extent of driving.

5. Discussion

clear, focusing on the actor with the ability to cause harm may help shift
focus to potential solutions instead of victim-blaming. Journalists may be
understandably wary of assigning blame to drivers; however, an article
can focus on the driver without making any assignment of fault.
In some cases neither the pedestrian nor the driver were acting illegally
in the moments leading up to the crash. Drivers traveling within the posted
speed limit routinely crash and pedestrians crossing legally are routinely hit
by drivers. In these cases, it is often the system, not individuals, that is primarily responsible. In these cases, thematically-framed articles are the
best option, and writing in this manner also aligns with the Vision Zero traffic safety movement (Johansson, 2009). These articles do not unduly blame
the driver, and they provide context that serves as a call to action. In short,
if the goal is to call public attention to external circumstances that cause
road deaths, thematic framing helps to accomplish this goal.
In this study, readers encountered just a single article, which was
enough to measure altered perceptions. The magnitude of the differences
is remarkable given the simplicity of the differences in text. The cumulative
effect of shifting to driver-focused and thematically-framed crash coverage
could result in powerful cultural change related to traffic crashes.
Of course, changing editorial patterns is somewhat difficult. In addition
to plain old inertia, there are many structural pressures in the news industry. There is considerable pressure for overstretched journalists on tight
deadlines to simply parrot police reports. To the extent that it is feasible,
journalists should take a few steps to move towards best practices (listed
below in ascending order of the effort required to implement them).

To recap, minor changes in crash coverage can shift a reader's perceptions of what happened and what to do about it. The results confirm that
sentence level editorial patterns do indeed shape perceptions about
blame, punishment, and preferred solutions. Specifically, shifting from a
pedestrian-focused to a driver-focused text increased perceived blame for
the driver. In turn, shifting to a thematically-framed text slightly reduced
blame on the driver and increased blame on “other” factors.
The results about preferred policy solutions are consistent with the
work of Iyengar who found that thematic framing increases demand for systematic public policy solutions (Iyengar, 1994, 1996). The present study
finds that thematic framing increases support for pedestrian infrastructure,
enforcement, and lower speeds—all of which are systems-level solutions—
and reduces support for an educational campaign, an individual-level
solution.
At the moment, the default setting is for crash coverage to blame the pedestrian (Magusin, 2017; Ralph et al., 2019; Scheffels et al., 2019). In most
cases, it is premature to assign blame because an investigation is inconclusive, still underway, or was never even started. Despite the lack of evidence,
writers rarely break from the practice of blaming the victim. In fact, the default is so powerful that writers often use it even when the driver was either
charged with or convicted of a crime related to the crash (Ralph et al.,
2019). Moreover, writers rarely make corrections to shift the blame away
from the pedestrian when exculpatory evidence comes to light (Flax,
2018). And, as this study shows, how an article is written affects how
readers perceive crashes. Given all of this, as a general rule, it is a good
idea to shift away from pedestrian-focused texts.
Driver-focused texts represent an improvement over pedestrian-focused
reporting. Some drivers are clearly responsible for crashes, such as when
they drive distracted, at excessive speeds, or in a reckless manner. In
these cases, using driver-focused text instead of standard practice of focusing on the pedestrian, is associated with an increased assessment of blame
and greater willingness to justly punish drivers. Even when fault is not

o Avoid non-agentive and object-based language.
o Shift the focus away from the pedestrian and towards the driver (or if
necessary, the vehicle).
o Be conscious of the counterfactuals that you include. Specifically, if you
mention that the pedestrian was outside a crosswalk, check Google
Street View to quickly determine whether there are any crosswalks
available and note that in many jurisdictions it is legal to cross outside
marked crosswalks.

Fig. 2. Thematic frame significantly increases support for safety-related infrastructure improvements.
6

 T. Goddard et al.

Transportation Research Interdisciplinary Perspectives xxx (xxxx) xxx

publishing fees for this article have been covered by the Texas A&M University Open Access to Knowledge Fund (OAKFund), supported by the University Libraries and the Office of the Vice President for Research.

o Include data on the number of crashes, injuries or deaths, preferably locally. Time permitting, consider contacting a local transportation, urban
planning, or public health expert to provide further context.

Declaration of competing interest

5.1. Strengths and limitations

None.

The study design has a number of strengths. In particular, the experimental nature of design enables causal inference. The findings are not
merely associations, they are the result of subtle changes in the text. Moreover, respondents only encountered one version of the text so it is unlikely
that they could anticipate the purpose of the study and act strategically.
Despite these strengths, some caveats deserve mention. First, use of an
online sample is relatively new. Even though the panel was representative
of the U.S. public in terms of age, sex, and race, the panelists may differ systematically in some unknown way. Second, there was remarkably broad
support for improving safety (>75% support in all cases). In part this may
be due to the fact that the proposed changes were purely hypothetical
and do not ask respondents to envision changing their own behavior.
Thus, respondents may overestimate their willingness to sacrifice slower
travel speeds and/or more enforcement for improved safety. There may
also social desirability bias if respondents thought we would like for them
to approve of these options. On the other hand, the level of support observed here may only appear high relative to the fierce opposition on display at public meetings, which may themselves overstate public
opposition to safety improvements. The treatment in this study was a
brief encounter with a news article, so the observed effect may be shortlived for experiment participants. This possible limitation could be ameliorated in practice, though, if the majority of crash reporting used driverfocused and thematic reporting techniques.

References
Boykoff, M.T., Boykoff, J.M., 2007. Climate change and journalistic norms: a case-study of us
mass-media coverage. Geoforum 38 (6), 1190–1204.
Branscombe, N.R., Owen, S., Garstka, T.A., Coleman, J., 1996. Rape and accident counterfactuals: who might have done otherwise and would it have changed the outcome? J. Appl.
Soc. Psychol. 26 (12), 1042–1067.
Christie, N., 1986. The ideal victim. From Crime Policy to Victim Policy: Reorienting the Justice System. Palgrave Macmillan UK, London, pp. 17–30.
Coates, L., Wade, A., 2004. Telling it like it Isn't: obscuring perpetrator responsibility for violent crime. Discourse Soc. 15 (5), 499–526.
Cobb, R.W., Primo, D.M., 2004. The Plane Truth: Airline Crashes, the Media, and Transportation Policy. Brookings Institution Press.
Davis, R.M., Pless, B., 2001. Bmj bans “accidents”: accidents are not unpredictable. BMJ 322
(7298), 1320–1321.
Fausey, C.M., Boroditsky, L., 2010. Subtle linguistic cues influence perceived blame and financial liability. Psychon. Bull. Rev. 17 (5), 644–650.
Fausey, C.M., Long, B.L., Inamori, A., Boroditsky, L., 2010. Constructing agency: the role of
language. Front. Psychol. 1, 162.
Flax, P., 2018. Why Is it So Hard to Charge Motorists With Murder? Outside Online
Frost, K., Frank, E., Maibach, E., 1997. Relative risk in the news media: a quantification of
misrepresentation. Am. J. Public Health 87 (5), 842–845.
Goldinger, S.D., Kleider, H.M., Azuma, T., Beike, D.R., 2003. “Blaming the victim” under
memory load. Psychol. Sci. 14 (1), 81–85.
Harcup, T., 2001. D. O'Neill. What is news? Galtung and Ruge revisited. Journal. Stud. 2 (2),
261–280.
Hinchcliff, R., Chapman, S., Ivers, R.Q., Senserrick, T., Du, W., 2010. Media framing of graduated licensing policy debates. Accid. Anal. Prev. 42 (4), 1283–1287.
Hu, W., Cicchino, J.B., 2018. An Examination of the Increases in Pedestrian Motor Vehicle
Crash Fatalities During 2009–16. Arlington, VA. .
Iyengar, S., 1994. Is Anyone Responsible?: How Television Frames Political Issues. University
of Chicago Press.
Iyengar, S., 1996. Framing responsibility for political issues. The Annals of the American
Academy of Political and Social Science 546, 59–70.
Jacoby, W.G., 2000. Issue framing and public opinion on government spending. Am. J. Polit.
Sci. 44 (4), 750–767.
Jeanis, M.N., Powers, R.A., 2017. Newsworthiness of missing persons cases: an analysis of selection bias, disparities in coverage, and the narrative framework of news reports. Deviant Behav. 38 (6), 668–683.
Johansson, R., 2009. Vision zero – implementing a policy for traffic safety. Saf. Sci. 47 (6),
826–831.
Magusin, H., 2017. If you want to get away with murder, use your car: a discursive content
analysis of pedestrian traffic fatalities in news headlines. Earth Common Journal 7 (1),
28–61.
Marris, P., Thornham, S., 2000. Media Studies: A Reader. New York University Press, New
York.
McCombs, M.E., Shaw, D.L., 1972. The agenda-setting function of mass media. Public Opinion
Quarterly 36 (2), 176–187.
Mouter, N., van Cranenburgh, S., van Wee, B., 2018. The consumer-citizen duality: ten reasons why citizens prefer safety and drivers desire speed. Accid. Anal. Prev. 121, 53–63.
National Highway Traffic Safety Administration, 2017. Quick Facts 2016. Washington, DC. .
Niemi, L., Young, L., 2016. When and why we see victims as responsible: the impact of ideology on attitudes toward victims. Personal. Soc. Psychol. Bull. 42 (9), 1227–1242.
Palan, S., Schitter, C., 2018. Prolific.Ac—a subject pool for online experiments. J. Behav. Exp.
Financ. 17, 22–27.
Ralph, K., Iacobucci, E., Thigpen, C.G., Goddard, T., 2019. Editorial patterns in bicyclist and
pedestrian crash reporting. Transp. Res. Rec. 2673 (2), 663–671.
Scheffels, E., Bond, J., Monteagut, L.E., 2019. Framing the bicyclist: a qualitative study of
media discourse about fatal bicycle crashes. Transp. Res. Rec. 2673 (6), 628–637.
Scheufele, D.A., 1999. Framing as a theory of media effects. J. Commun. 49 (1), 103–122.
Smith, G., 2014. Standard Deviations: Flawed Assumptions, Tortured Data, and Other Ways to
Lie With Statistics. First edition. Gary Smith, First edition. ed. Overlook Duckworth.
Transportation Alternatives, 2016. AP stylebook to reporters: avoid the word “accident”. URL.
Transportation Alternatives https://www.transalt.org/news/releases/9545, Accessed
date: 19 October 2019.
Walker, I., 2005. Road Users' Perceptions of Other Road Users: Do Different Transport Modes
Invoke Qualitatively Different Concepts in Observers?

6. Conclusion
This study provides evidence that editorial patterns in traffic crash
reporting influence people's interpretation of what happened and what to
do about it. While the media studies and psychology literatures suggested
this would be the case, this study is the first to test the power of editorial
patterns with respect to traffic crash coverage. The results are particularly
concerning given that non-agentive language, a focus on vulnerable road
users, object-based language, and episodic framing are commonplace in
crash reporting (Magusin, 2017; Ralph et al., 2019; Scheffels et al., 2019).
Encouragingly, though, editorial practices are not immutable. It costs little
to use agentive language, appropriate focus, person-based language, and—
to a lesser extent—thematic framing. Additionally, improved editorial
choices do not take any special training or technical understanding of traffic
safety. Adopting simple improvements in crash reporting, however, offers a
potentially powerful tool in shifting cultural awareness of traffic crashes as
a preventable public health issue. Given the potential to save human lives
and prevent injury on a large scale, implementing more intentional editorial patterns may be nothing less than an ethical imperative.
Author contributions
The authors confirm contribution to the paper as follows: study conception, experimental design, and survey development: T. Goddard, E. Iacobucci,
K. Ralph, C. Thigpen; data collection: T. Goddard; analysis and interpretation
of results: T. Goddard, E. Iacobucci, K. Ralph, C. Thigpen; draft manuscript
preparation: T. Goddard, E. Iacobucci, K. Ralph, C. Thigpen. All authors
reviewed the results and approved the final version of the manuscript.
Funding source
This research did not receive any specific grant from funding agencies in
the public, commercial, or not-for-profit sectors. The open access

7