Coverage of motor vehicle crashes with injuries in U.S. newspapers, 1999–2002

Abstract

Problem

The aims of the study were to evaluate information on motor-vehicle crashes with injuries provided in newspaper reports and to assess the frequency of thematic and episodic reporting of motor-vehicle crashes.

Method

The present study used Fatal Analysis Reporting System (FARS) derived variables to code a nationally representative sample of U.S. newspaper reports of motor-vehicle crashes from 1999–2002. A total 473 newspaper reports of motor-vehicle crashes with injuries were included. Information on the crash event, people involved, and vehicles was extracted. The reports were coded for episodic and thematic news framing.

Results

A majority of newspaper reports used episodic framing. The majority of reports included information on the type of crash, but characteristics about people and vehicles were rarely reported.

Discussion

Lack of information in newspapers makes them an incomplete source from which to influence public perceptions and attitudes.

Impact on industry

This provides an opportunity for news print media to improve public health content.

Introduction

The role of newspapers in expanding injury surveillance (Rainey & Runyan, 1992; Fine, Jones, Wrigley, Richards, & Rousculp, 1998) and in presenting health risk information (Frost, Frank, & Maibach, 1997) has been assessed. In addition, the reporting of injury prevention measures in newspapers has been discussed (Voight, Lapidus, Zavoski, & Banco, 1998). One study conducted in two counties in Colorado reported the relationship between newspaper reported and hospital discharge related injury rates, indicating motor-vehicle injury reports in newspapers underrepresented the number of hospitalized injuries resulting from motor-vehicle crashes (Rosales, Smith, & Stallones, 2006). In a study conducted in four Midwestern cities in the United States, newspaper reports were compared with data from the National Highway Transportation Safety Administration Fatal Analysis Reporting System (FARS) to evaluate the crash risks, causes, and reported consequences in newspapers (Connor & Wesolowksi, 2004). The newspaper reports did not accurately reflect real risk because crashes were reported as drama with victims and villains and factors associated with reducing injurious consequences such as seatbelt use were under reported (Connor & Wesolowski).

Newspapers have the power to inform and educate the public (Voight et al., 1998). Prevention information and sufficient details that may be helpful for the public to evaluate their own risk often are not included in reports, which influences risk perceptions about public health issues (Slater, Long, & Ford, 2006). By providing information about injury risk and the benefit of taking preventive measures, newspapers could make use of their power to produce social change. Media coverage can assist in defining health problems as public policy problems rather than as individual problems (Long, Slater, & Lysengren, 2006). Therefore, it is important to understand the nature of media reports to understand that media content may influence public concern and public policy (Slater & Rasinski, 2005).

Media coverage can increase the perceived importance of issues as well as alter attitudes toward issues (Long et al., 2006). Risk perception about relative severity and prevalence are affected by news reports on crime, crashes, and injuries (Slater & Rasinski, 2005). Views regarding issues like terrorism and poverty were found to depend on the point of view presented in the media (Iyengar, 1987). Therefore, the importance of documenting the extent to which particular media frames are more or less prevailing in news coverage has been suggested (Iyengar, 1987). Iyengar (1996) focuses on two specific types of framing, episodic and thematic. Episodic news framing refers to presentation of issues in terms of specific instances as isolated events (Iyengar, 1996). Thematic framing represents the presentation of issues in a broader societal context by placing issues in an appropriate context (e.g., temporal, geographical; Iyengar, 1996).

The purposes of this study were to use FARS derived variables to code a nationally representative sample of newspaper reports in order to determine the extent to which information often used to design motor-vehicle injury prevention programs was included, and to assess the frequency of thematic and episodic reporting of motor-vehicle crashes.

Method

Detailed descriptions of the sampling frame have been published (Long, Salter, Boiarsky, Stapel, & Keefe, 2005; Slater et al., 2006). Briefly, U.S. newspapers were sampled approximately proportional to their circulation or reach. Using Designated Marketing Areas (DMA) allows the selection of a sample based on media market size and is one of the most widely used approaches to defining media markets (Long et al., 2005). A DMA is a U.S. geographic division used to measure television audience size by Nielsen Media Research and it is the most widely used method for defining media markets (Slater, Long, & Ford, 2006; Standard Rate and Data Service, 2000). There are 210 DMAs in the United States that cover 99.8% of the U.S. population (Nielsen Media Research, 2006). DMAs have proven to be useful in segmenting and sampling newspaper media (Kunkel & Canepa, 1994; Kunkel & Goette, 1997; Slater & Rasinski, 2005). The 210 DMAs were stratified into six categories, with each category consisting of approximately 1/6th of U.S. households so that their probability of inclusion was proportional to their circulation or reach (Slater et al., 2006).

The newspaper sampling used in this report occurred over a four-year period (1999–2002). For each sampled year, a constructed month was created by constructing a week from each season of the year. This provided a balanced sample in terms of day of the week and season of the year, which provides a good representation of news coverage during each year. On each sampling date, one DMA was randomly selected from each of the six categories. In each DMA, the largest daily newspaper for that day was selected (Long et al., 2005; Slater et al., 2006). Then a median split was selected from each DMA on each sampling date. As a result, on each sampling day, 18 local newspapers were selected (Long et al., 2005; Slater et al., 2006). The sampling approach provided a geographically representative sample when contrasted with the four U.S. Census regions (Long et al., 2005).

Selection of newspaper reports

From the original database, a subset of newspaper reports citing motor-vehicle crashes with injuries or deaths were included; those involving motor-vehicles crashes with no reported injuries, deaths or near misses were excluded. Only newspaper reports that were not a part of the opinion, editorial, obituary, or sidebar sections were included. Reports mentioning motor-vehicle related injury or mortality in which the motor-vehicle crash occurred greater than one year before the story date were also excluded.

Coding of newspaper reports

The instrument used to extract information from the newspaper reports was developed using variables included in the FARS. The data set was structured to reflect the crash event (day of the week, date of event, holiday, time of day, speed limit, number of people involved in crash, number of fatalities involved in crash, weather conditions, road condition, type of road, city, county, and state), the driver and people involved (driver, age, sex, race, Hispanic ethnicity, seating position, taken to hospital, restraint used, airbag deployed, ejected, extricated, driver drunk, drug involvement, violations issued, license compliance), and the characteristics of vehicles involved (number of occupants in each vehicle, number of fatalities in each vehicle, travel speed of each vehicle, type of vehicle (make and model), extent of deformation, special use vehicle, roll-over, and owner registration information).

In addition to the FARS variables, each report was coded as representing episodic or thematic reporting (Ivengar, 1996). For this study, newspaper reports about motor-vehicle crashes were considered episodic unless they contained general statistics (e.g., one in 100,000 people will be involved in a motor-vehicle crash this year), descriptive information regarding the broader issue of motor-vehicle safety and crashes, or risk estimates about the relative importance of motor-vehicle crashes or injuries. General statements (e.g., young drivers are more likely to experience a motor-vehicle crash than older drivers) included in news articles qualified the articles as thematic.

A pilot study was conducted using 50 newspaper reports not included in the database to determine the usability of the data collection instrument. Inter-rater reliability was conducted using two independent coders. Due to initial low inter-rater reliability, stricter instructions were developed and the data collection instrument was modified. The reports were then re-coded by the two coders using the new guidelines and reliability was rerun. Variables were included in the final survey instrument only if inter rater reliability coding based on the Kappa scores were in a range of adequate agreement (0.41–0.60) or better (Landis & Koch, 1977).

Results

A total of 473 newspaper reports of motor-vehicle crashes were included in this study. The majority of the reports represented episodic reporting (459; 97%). Table 1 contains the distribution of type of crash reported. The majority of newspaper reports provided information on the type of crash, with the most frequently reported type of crash being single car events followed by two car collisions. Details about other characteristics related to the crashes were not often reported in the newspaper reports. Only 6 (1.3%) included information about the weather and only 11 (2.3%) included information about the condition of the road surface (e.g., wet, snow, or slush). The hour of the crash was reported in 43.3% of all newspaper reports. Whether the crash took place during a holiday was not reported for 98% of the crashes. The speed limit was reported in six (1.3%) of the reports. Drunk driving was reported in 11.2% of the crashes, no drunk driving was reported in 5.9% of the crashes, and suspected drunk driving was reported in 2.5% of the crashes; there was no mention of drunk driving in 80.3% of the crashes. Drug involvement was less likely to be included in the newspaper reports with 94.3% having no mention, 1.7% reported as yes, 3.6% reported as no, and 0.4% as suspected.

Table 1.

Type of crash and report frequency, 1999–2002

Type of Crash	Frequency	Percent
Single car collision	134	28.3
2 car collision	75	15.9
More 2 cars involved	33	7.0
Car/truck collision	19	4.0
Single truck	17	3.6
Head-on collision	15	3.2
Single Motorcycle	14	3.0
Car/bicycle collision	11	2.3
Car/motorcycle collision	11	2.3
Car/train collision	7	1.5
Car/van collision	4	0.8
Single van	4	0.8
Car/bus collision	4	0.8
Unknown	15	3.2
Other	110	23.3
Total	473	100.0

The newspaper reports contained information about 699 vehicles. Table 2 contains information about the type of vehicles involved and the extent of damage to the vehicles. For most vehicles there was no mention of a roll-over and whether the vehicle was being used for a special purpose. Estimated travel speed was not mentioned for 83% of vehicles, but 10% of vehicles were reported as not in motion when the crash took place. No fatalities were reported in 73% of the episodes.

Table 2.

Characteristics of motor vehicles involved in newspaper reported crashes, 1999–2002

Vehicle	Number (n = 699)	Percentage
Description
No mention	366	47.6
Mentioned	333	52.4
Extent of deformation
Totally deformed	10	1.4
Functional	8	1.1
Not functional	4	0.6
No mention	677	96.9
Involved in Roll-over
No	72	10.3
Yes	30	4.3
No mention	597	85.4
Special Use
No special use	198	28.3
Special use-Other	49	7.0
No mention	452	64.7

A total of 1,179 people were represented in the newspaper reports. Age and sex of those involved in the crashes are shown in Table 3. Age was not reported for 44% of those involved and sex was not reported for 28% of those involved. Race was not reported for any of the individuals and Hispanic ethnicity was not reported for 96.3% of individuals identified in the newspaper reports. Mexican ethnicity was reported for one individual and two individuals were identified as non-Hispanic. A Hispanic surname was reported for 3.4% (40) of individuals. In 25.6% of the reports there was only one person involved in the crash reported, two people were involved in 29% of the crashes, three people were involved in 15% of the crashes, four or more people were involved in 28% of the crashes. In 2.7% of the crashes there was no information provided on the number of people involved. There were no fatalities in 39.3% of the crashes reported; in 40.2% of the reports there was one fatality and in 20.5% of the reports there were two or more fatalities. Among the 1,179 people involved in motor-vehicle crashes, 44.4% were drivers. In terms of location, 26% were seated in the front, 1% were seated in the back, 1 (<1%) person rode on the outside of the car, 7 (<1%) were riding in a trailing unit, and for 60% there was no mention of where they were seated. A total of 13% of involved individuals were not in a motor vehicle.

Table 3.

Characteristics of individuals involved in newspaper reported motor vehicle crashes, 1999–2002

	Frequency	Percent
Age in years
1–4	24	2.0
5–9	23	1.9
10–14	48	4.0
15–19	112	9.5
20–29	129	10.9
30–39	79	6.7
40–49	83	7.0
50–59	60	5.0
60–69	64	5.4
70+	44	3.7
Not reported	513	44.0
Sex
Male	578	49.0
Female	270	23.0
Not reported	331	28.0
Role in car/crash
Driver	523	44.4
Passenger/occupant	475	40.3
Non-motorist	153	12.9
Unknown	8	0.7
No mention	20	1.7

The majority of males represented in the newspaper reports were identified as drivers (Table 4). A slightly higher number of females were reported as passengers or occupants than as drivers (Table 4). In addition, there was no mention of sex for a higher number of individuals described as passengers or occupants compared with drivers (Table 4).

Table 4.

Role in car/crash and sex in newspaper reported motor vehicle crashes, 1999–2002

Role of person	Sex
	Male	(%)	Female	(%)	No mention	(%)	Total	(%)
Driver	344	(29.2)	102	(8.7)	77	(6.5)	523	(44.4)
Passenger/occupant	178	(15.1)	113	(9.6)	184	(15.6)	475	(40.3)
Non-motorist	51	(4.3)	54	(4.6)	48	(4.1)	153	(12.9)
No mention	3	(0.2)	1	(0.1)	16	(1.4)	20	(1.7)
Unknown	2	(0.2)	0	(0.0)	6	(0.5)	8	(0.7)
Total	578	(49.1)	270	(22.9)	331	(28.0)	1179	(100.0)

Little information was provided about the legal driving status of those driving vehicles in the reported crashes. For 61% of individuals there was no mention of having a valid drivers license. Only six (<1%) people were reported to have a valid drivers license, 1.9% did not have a drivers license, 0.2% did not have a valid license for the type of vehicle being operated, and 36% were not driving. There was no mention of traffic violations in 88% of crashes reported.

Transport to the hospital was reported for 44.6% of all individuals, with 11.4% having been transported to a hospital and 33.2% not having been taken to a hospital. No comment was made regarding ejection from a vehicle among 90.6% of people involved in crashes and for 92.7% there was no information about extrication from the vehicle. Use of safety restraints was not mentioned for 77.7% of individuals involved in crashes. Among crashes where there was a report, 19.3% of individuals were not passengers in motor vehicles nor were they riding motorcycles or bicycles. For 86.4% of the individuals involved in crashes there was no mention of airbag deployment. Deployment of an airbag was mentioned for 0.7% of the individuals. Seatbelt use was reported for 2.5% of the individuals, motorcycle helmet use was reported for 0.2% of individuals, and use of a bicycle helmet was mentioned for one individual.

Discussion

While newspapers have been reported to be a potential source for injury surveillance data (Baullinger, Quan, Bennett, Cummings, & Williams, 2001; Rainey & Runyan, 1992), they may not accurately represent motor-vehicle crashes with injuries. In this study, newspaper reports contained limited information about factors that may contribute to motor-vehicle crash injuries and virtually no information about factors that predispose to crash risk, such as weather conditions and speed (Marmor & Marmor, 2006). Connor and Wesolowski (2004) reported that newspapers do not take into consideration the role many factors may play in crashes. Further, contrasting the conclusions based on newspaper reports may be misleading. For example, in the absence of any comment in the newspapers, readers might assume seatbelt use was present or absent based on their own perceptions about frequency of use. Based on FARS data, in more than 54–57.0% of fatal crashes over the period 1999–2002 seatbelts were not used (FARS, 2008). Nambisan and Vasudevan (2007) reported that in 2005 about 75% of unrestrained individuals involved in motor-vehicle crashes were ejected and killed. Underreporting seatbelt use is a concern, because adhering to this safety behavior has the potential of preventing fatalities in motor-vehicle crashes.

Infrequent reporting of alcohol and drug involvement is a finding this study shares with previous research (Baullinger et al., 2001; Rainey & Runyan, 1992; Slater et al., 2006). This underreporting is a concern because it may lead the reader to assume that alcohol and drugs are not usually involved in motor-vehicle crashes. Not providing an accurate representation of the involvement of alcohol and other drugs in newspaper reports may lead the public to minimize the risks and dangers of driving under the influence. Slater et al. (2006) reported that drunk driving has received substantial public health attention for decades. This has led to testing blood alcohol concentrations in serious fatal crashes; information that is readily available to reporters, however, was not included in these newspaper reports.

Information about race was rarely mentioned. Rainey and Runyan (1992) also found that newspapers did not provide adequate information about race. Rosales et al. (2006) used Spanish surname as a surrogate for Hispanic ethnicity because ethnicity and race were not reported in newspaper reports used in that study. Age and sex were reported for slightly more than half of individuals represented in these newspaper reports. Providing the age of those involved in fatal crashes can be important for regions in which motor-vehicle crashes are prevalent for specific age groups. For example, using FARS data, Frisch (2007) found that Texas had the highest number of crashes among individuals under the age of 15 compared to other states. The absence of age and driving speeds further compounds the potential for misperceptions about risky behavior contributing to motor-vehicle crashes. For example, driving at speeds that are too fast for road conditions has been found to be a contributing factor of motor-vehicle crashes among teenagers (Marmor & Marmor, 2006; McKnight & McKnight, 2003). Incorporating this type of information into newspaper reports can increase public awareness of age related motor-vehicle risk and alter public willingness to address high risk age groups in local communities. Public policies related to driving regulations may be influenced to reduce at-risk driving behaviors through appropriate use of newspaper reporting.

Implications for injury prevention

Research on newspaper coverage of motor-vehicle crashes has used local or regional samples in contrast to this study that was based on a nationally representative sample of U.S. newspapers. Newspaper reports were coded using variables derived from FARS. However, motor-vehicle crash information reported in newspaper reports appeared to lack sufficient detail based on these FARS variables. For example, FARS data have included race and Hispanic ethnicity since 1999 (Briggs et al., 2005). By the year 2000, national reporting reached 90% for race and 88% for Hispanic ethnicity (Briggs et al.). Adding this information into newspaper reports may provide readers with a more balanced view of their risk of motor-vehicle crashes and injuries, while avoiding some of the biased perceptions that have been reported by others (Heller, 1992; Mize & Leedham, 2000).

The media focuses on emphasizing distinctiveness and public health focuses on detecting trends and in identifying risk factors (Connor & Wesolowski, 2004). Injury related events are more likely to be covered when they seem to be out of the ordinary, rare, or dramatic (Connor & Wesolowki, 2004; Frost et al., 1997). This may lead to presenting such stories as isolated events (episodic), not as a public health concern. This type of reporting may provide the public with an inaccurate perception by overestimating infrequent causes of mortality and underestimating frequent causes. Episodic framing has been reported to cultivate individualistic as opposed to societal attributions of responsibility (Iyengar, 1996). Presenting complex issues as anecdotal cases (episodic framing) can be counterproductive (Iyengar, 1996). Encountering news reports that describe particular instances of national issues leads viewers to focus on specific individuals or groups depicted in the news instead of on historical, social, political, or other structural factors (Iyengar, 1996). Public health professionals should work with editorial boards of local newspapers to inform the editors and reporters about the importance of focusing on thematic reporting when addressing motor-vehicle crash information. Further, public health professionals should interact with undergraduate and graduate training programs in journalism and technical communications to highlight the role of the media in the development of health risk communication.

Impact on industry

Newspapers represent an important source of public information; they are, however, an incomplete source (Voight et al., 1998; Baullinger et al., 2001). To increase the accuracy of information provided to the public through media sources, there is a need for increased communication between public health professionals and reporters.

The results of this study raise concerns about the contents of motor-vehicle crash information provided in newspapers and suggest that newspapers do not provide information to allow public perception to be in accord with the importance of motor-vehicle crash injuries and health promoting actions to reduce risk of injury. More balanced and detailed information in newspapers would provide an opportunity for news print media to improve public health programs and public perception about the impact of motor-vehicle crashes on safety for all.

Biographies

Dr. Rosales completed her PhD in Applied Social Psychology at Colorado State University. Her research interests are in health psychology working primarily with Latino populations. She is a post doctoral fellow at the Center of Community Alliance for Research and Education (CCARE) Division of Population Sciences City of Hope National Medical Center, Los Angeles California.

Dr. Stallones is a Professor of Epidemiology and is the Director of the Colorado Injury Control Research Center. Her research interests include injury prevention and control among rural and underserved populations.