Accident Analysis and Prevention 73 (2014) 109–115
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Research-informed evidence and support for road safety legislation: Findings from a national survey Katherine Clegg Smith a,∗ , Beata Debinski a,b , Keshia Pollack a , Jon Vernick a , Stephen Bowman a,c , Alicia Samuels a , Andrea Gielen a a
Johns Hopkins Center for Injury Research and Policy, Johns Hopkins Bloomberg School of Public Health, United States Wake Forest School of Medicine, Department of Social Sciences and Health Policy, United States c University of Arkansas for Health Sciences, Fay W. Boozman College of Public Health, Department of Health Policy and Management, United States b
a r t i c l e
i n f o
Article history: Received 31 January 2014 Received in revised form 17 August 2014 Accepted 18 August 2014 Keywords: Injury Policy Legislation Research evidence Public opinion Survey
a b s t r a c t Public opinion is influential in the policymaking process, making it important to understand the factors that influence popular support or opposition to public health policies. Researchers and policymakers tend to agree that scientific evidence can inform decision-making, but this influence has not been explored sufficiently, especially in the area of injury prevention. This paper considers the potential for the communication of evidence-based research and public health data to influence opinion about legislation that could reduce road-related injury. We conducted a nationally-representative online survey to assess public attitudes toward four road-safety laws; ignition interlock, school zone red-light cameras, restrictions on infotainment systems, and children’s bicycle helmets. For each law, we assessed initial support and then provided a research-informed statistic on either the injury risk posed or the law’s efficacy reducing risk and re-examined the law’s support or opposition. The survey was completed by 2397 U.S. adults. Each law was initially supported by a majority of respondents, with greatest support for ignition interlock (74.4%) and children’s bicycle helmets (74.8%). Exposure to research-informed statements increased legislative support for 20–30% of respondents. Paired analyses demonstrate significant increases toward supportive opinions when comparing responses to the initial and research-informed statements. The study demonstrates considerable public support for evidence-based road-related laws. Overall support was augmented by exposure to research data. Injury prevention practitioners can capitalize on this support in efforts to build support for legislation that would prevent injury. Researchers should be encouraged to expand their efforts to share research results with both the public and policymakers. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Considerable progress has been made in reducing motor vehicle crash (MVC) deaths and injuries around the world. Still, motor vehicle crashes remain the leading cause of death for Americans ages 5–24 years, and, among all unintentional injury deaths, they are the leading cause for the entire population (Centers for Disease Control and Prevention, 2012; Levi et al., 2012). In 2011, there were approximately 32,000 road-related fatalities (including both occupants and non-occupants) and over 2.6 million vehicle occupants were treated in emergency departments for nonfatal injuries sustained in a vehicle crash (Centers for Disease Control and Prevention, 2012). Approximately one in three deaths on the road involve an alcohol impaired driver (Levi et al., 2012; Shults and Bergen, 2013), and
∗ Corresponding author. Tel.: +1 4105020025. E-mail address: [email protected] (K.C. Smith). http://dx.doi.org/10.1016/j.aap.2014.08.016 0001-4575/© 2014 Elsevier Ltd. All rights reserved.
current estimates are that about one in six fatal crashes involves a distracted driver (Levi et al., 2012). Globally, estimates from the Global Burden of Disease Report show the burden of road traffic injuries. Between 1990 and 2010, road traffic injuries went from the 12th to 10th rank (a 34% increase) among the leading causes of the global burden of disease (Murray et al., 2013). By 2030, estimates are that road traffic injuries will be the 5th leading causes of death in the world (World Health Organization, 2013). Policies, such as legislation and regulation, that prohibit risky behaviors and/or products can be effective in shifting social norms, changing behavior, and preventing injuries and fatalities. Policies often provide incentives for people to follow safe practices and in so doing, protect themselves and others (Levi et al., 2012). For example, effective policies to reduce motor vehicle crashes and resulting death and injury include primary enforcement of seat belt laws (Dinh-Zarr et al., 2001), installation of ignition interlocks for convicted drunk drivers (Willis, 2005), mandatory motorcycle (Liu et al., 2008), and bicycle helmet use (Macpherson and Spinks, 2008),
110
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
graduated driver licensing (Russell et al., 2011), installation of redlight cameras (Aeron-Thomas and Hess, 2005), and speed cameras (Institute of Medicine. Committee on Assuring the Health of the Public in the 21st Century, 2003; Wilson et al., 2010). Despite the evidence base regarding the efficacy of public policy in reducing road-related injury, a recent report indicates low state-level utilization of legislation (Levi et al., 2012) for preventing injury and death among road users. For example, as of fall 2012 (when our survey data were collected), only 21 states and the District of Columbia (DC) required children to wear bicycle helmets; 17 states and Washington D.C. had comprehensive, primary seat belt laws; and 16 states required mandatory ignition interlocks for all convicted drunk drivers, including first-time offenders (Levi et al., 2012). There are many factors that can influence road safety policies including (but not limited to) the framing of such policies in relation to road safety or mobility (Bates et al., 2010; Hirsch, 2003). Public opinion changes can also be important factors in influencing the legislative process (Page and Shapiro, 1983), and public opinion change often (but not always) precedes policy change. It is therefore critical to understand the factors that lead people to become more supportive of policies that have been shown to be effective in preventing road-related injury. Prior work has demonstrated the potential impact of substantive information on the magnitude of a problem on policy stakeholders’ consideration of potential solutions (Sabatier, 1991). Moreover, research findings can serve to define a problem and frame potential solutions (Gold, 2009). The question emerges, therefore: Can public opinion regarding effective road-related laws be impacted by providing research-informed statements about such laws? A recent review of the literature (Debinski et al., 2014) found 26 studies and reports addressing public opinion about legislation pertaining to motor vehicle and bicycle-related topics (e.g., allterrain vehicles, automated enforcement with red-light and speed cameras, distracted driving, drinking and driving, graduated driver licensing, bicycle and motorcycle helmets, and seat belts). Results suggested that public opinion toward injury prevention legislation is generally favorable, but few studies reported on factors that might serve to influence support or opposition for such initiatives. Understanding public perceptions and influences can inform consideration of the discourse about legislation so as to better protect the public from preventable injury – both on the road and more broadly. The present study seeks to address two key gaps in the current road safety literature. Our first aim is to document the magnitude of support and opposition for four laws among a nationally representative sample of U.S. adults. Second, we examine whether support and opposition to the selected laws is impacted by the provision of factual information on risk or efficacy of the given law in relation to injury risk. Addressing these questions will demonstrate the value of research evidence in shaping public opinion, and may facilitate public information efforts focused on road-related risks, responsible road use practices, and effective laws. We explore whether research-informed statements that are supportive of the selected laws are associated with an increased level of public support (and reduced opposition) for that law.
cameras in school zones as a structural/community-level initiative, and laws pertaining to in-motion infotainment system disablement, so as to represent emerging safety concerns. We chose laws that have at least some evidence supporting their effectiveness in reducing injury risk. The resulting survey included questions pertaining to the four specific laws: (1) mandatory ignition interlock installation for all driving under the influence (DUI) offenders even after their first offense; (2) a mandatory bicycle helmet law for all children under 16 years old; (3) allowing communities to install red light cameras in school zones to help improve pedestrian safety; and (4) requiring that in-vehicle infotainment systems ‘lock’ when the car is moving. There have been five prior studies and reports addressing public opinion regarding mandatory installation of ignition interlocks, and one study about bicycle helmets for riders under 16 specifically, but no data have as yet been reported about support for red light cameras in school zones or restrictions to infotainment system operation (Debinski et al., 2014). With regard to the influence of evidence on public opinion, Girasek (2013) found that the public strongly endorsed the concept that they “expect evidence-based traffic safety policies”. For each of the selected laws, we used the most reliable evidence available to provide a relevant research-informed risk or risk reduction statement. These included two statements that addressed the magnitude of the injury risk posed and two that addressed the efficacy of the law in reducing injury risk. Each statement presents injury risk and/or reduction differently. Statistics were obtained from the CDC Injury and Violence Prevention and Control site, the CDC Web-based Injury Statistics Query and Reporting System (WISQARS), and reviews published in the Cochrane Database of Systematic Reviews (see Table 1). 2.2. Data collection Our data are the result of a nationally representative online survey conducted over two days (October 29–30, 2012). We contracted with the polling firm, Harris Interactive Inc. to include a set of road-related questions about specific laws in a survey with U.S. adults (ages 18 and older) via its Harris Poll Quick Query omnibus product. An online format was chosen over a telephone survey to yield a greater number of responses in a more cost-efficient manner. The Harris Poll OnlineSM Panel members are recruited for surveys using a “double opt-in” process, which requires that each registrant confirms participation in a panel by clicking on a link within an email that is sent to their email address, after voluntarily registering to be a part of the Panel. The survey is designed to permit only one respondent per household. Harris Interactive generates a nationally representative sample using a two-stage approach, first balancing the sample that will receive a survey invitation, and then weighting collected data on at least eight demographic factors (Harris Interactive, 2009). In addition to our set of questions, we also collected demographic information about each respondent including residence by region of country, age, sex, marital status, highest completed education level, annual household income, household size, employment status, and whether or not there are any children under 18 in the home.
2. Methods 2.3. Survey items 2.1. Topics We selected four laws that pertain to road safety, two of which (alcohol ignition interlock and bicycle helmets), were selected from the Trust For America’s Health report of state-level injury indicators (Levi et al., 2012). We also selected laws that require red light
All respondents were asked to indicate support or opposition to each of the four sample laws. In the online survey, each law and the questions related to it were presented individually and in turn to each respondent. Each law was introduced through a brief descriptive text accompanied by an illustrative photo
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
111
Table 1 Questions posed and research-informed statement provided. Topic
Question posed
Evidence provided and source
Alcohol ignition interlock
Would you support or oppose a law in your state that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law would include people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car Would you support or oppose a law in your state that requires all children under 16 to wear helmets when riding a bicycle?
Research shows that when an alcohol ignition is on the car, the driver’s risk of being caught again for drunk driving decreases to 1/3 of what the risk would be without the interlock (Willis et al., 2004)
Bicycle helmets
Vehicle infotainment system
Would you support or oppose a law requiring that infotainment systems cannot work when the car is moving?
Red light camera
Would you support or oppose a law in your state that allows communities to install red light cameras in school zones to help improve pedestrian safety?
Research shows that bicycle helmets provide a 63% reduction in head injury and 86% reduction in brain injury in children younger than 15 (Thomas et al., 1994; Thompson et al., 1999) Research shows that in the U.S., 2 out of 10 serious crashes involve distracted driving, resulting in over 5000 deaths and 440,000 injuries per year (NHTSA, 2009; CDC, 2012) Every year there are about 43,000 injuries from motor vehicle crashes with pedestrians younger than 18 (CDC WISQARS, 2012)
1. An alcohol ignition interlock is a device that can be attached to a car to determine if the driver has been drinking alcohol. The alcohol ignition interlock works by having the driver breathe into the device. If a preset level of alcohol is detected, the car will not start.
Photo: Colourbox.com 2. Would you support or oppose a law in your state that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law would include people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car. Strongly oppose Somewhat oppose Neither support nor oppose Somewhat support Strongly support Research shows that when an alcohol ignition is on the car, the driver’s risk of being caught again for drunk driving decreases to 1/3 of what the risk would be without the interlock. Based on this information, would you support or oppose a law in your state that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law would include people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car. Strongly oppose Somewhat oppose Neither support nor oppose Somewhat support Strongly support 3. Does the state that you live in have a law that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law includes people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car. Yes No I don’t know Fig. 1. Ignition interlock survey questions.
112
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
(see Fig. 1 for an example of the question presentation). The poll questions and research-informed statements were assessed for face validity through cognitive interviewing with a convenience sample of fifteen participants. Cognitive interviewing is a qualitative method used to identify potential problems with survey questions where the survey is administered and then respondents are probed on various aspects of the questions (Willis, 2005). For our study, interviewees were shown the statements and accompanying photos and were asked to respond to the questions pertaining to each topic, and then asked to orally describe their understanding of the topic, including whether their understanding was supported by the text and/or picture; to interpret the law that was being addressed and explain how they made their decision; and to assess what they thought about the piece of evidence that was provided in the second question of each set. After completion of the cognitive interviews and consideration of issues raised by participants, modifications were made to the survey questions. Modifications included replacing pictures that were not beneficial for understanding the topic, and clarifying that each state would be responsible for specifying details of a given law. After being presented with the introduction and the photo, respondents were asked to indicate whether they would support or oppose each law. Possible responses were on a 5-point Likert scale, ranging from “strongly oppose” to “strongly support.” After responding to the initial question, respondents were prompted to the next screen where they were provided with a research-informed statement about the law (See Table 1). They were once again asked to indicate support or opposition for the same law, and were presented with the same five response options. It was not possible for respondents to go back and change their answers to the initial question after being presented with the research-informed statement. This research was reviewed and approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board.
question was coded as becoming more oppositional. To examine the proportion of participants who became more supportive or less oppositional, we divided the total number of people who changed their response in a positive direction by the total sample and subtracted the number of people who were strongly supportive in response to the initial question. Those who were initially strongly supportive were removed from the denominator because they were not eligible to become any more supportive in response to the evidence informed decision. To examine the proportion of participants who became more oppositional or less supportive, the same procedure was followed, and those who were strongly opposed in response to the first question were eliminated from the denominator. In addition to comparing the proportions of respondents by level of support and opposition after the initial and research-informed statements, we also conducted a paired analysis of the mean scores before and after exposure to the research-informed statement. 3. Results 3.1. Sample The Harris Omnibus Survey (Quick Query) was completed by 2397 respondents in the two-day window that it was available to panel members. Characteristics of the sample weighted to the U.S. population are presented in Table 2. The proportion of male and female respondents was roughly equal. The largest percentage (20%) of respondents was 45–54 years of age, with few older adults (>75 years) participating. More than half (56%) of respondents reported a family income of less than $75,000 and few (8%) reported an income of $150,000 or more. More than a quarter (29%) of respondents had at least a college degree (see Table 2). 3.2. Initial support and opposition
2.4. Statistical analysis Weighted data were analyzed using Stata 12.1 (StataCorp LP, College Station, TX). Harris Interactive assigned each respondent a weight for their contribution to the nationally representative sample (based on March 2011 Current Population Survey data). A survey design was declared for the dataset and the weight provided by Harris Interactive was designated as a probability, or sampling, weight. All Stata estimation commands subsequently utilized the ‘svy’ prefix, which accounts for the complex survey sampling. We report overall support and opposition for each law as well as the levels of strong support and strong opposition for each. We also report the proportion of ‘supportive’ movement and ‘oppositional’ movement for each law following exposure to the research-informed statement. We present data pertaining to each of the five response categories (“strongly oppose” to “strongly support”). Strong opposition was scored as 1, and strong opposition was scored as 5, such that a higher mean indicates higher support. A variable for “net support” and “net oppose” was created by condensing the “strongly” and “somewhat” responses for both support and opposition. A variable to describe movement in responses from the initial question to the research-informed second question was generated using the five original categories. Individuals who provided the same response for both questions were considered to have not moved or changed opinion. Any response to the second, research-informed question that was more supportive or less oppositional than the original response was considered a change in the positive direction. Similarly, any response to the second question that was less supportive or more oppositional than the response to the initial
Table 3 displays the proportion of respondents who initially supported and opposed each of the four laws in the study. Each of the four laws was supported by the majority of respondents, with the highest net support for ignition interlocks and bicycle helmet laws for children (74.4 and 74.8% respectively), followed by vehicle infotainment system laws (61.4%). There was least initial support for red light cameras laws (58.0%). The highest proportion of strong opposition was seen in response to red light camera laws (16.3%) and vehicle infotainment system laws (14.6%), which is more than double the proportions who strongly opposed ignition interlock laws (6.6%) and bicycle helmet laws (5.9%). 3.3. Research-informed support and opposition Table 3 also displays the proportion of respondents who supported and opposed each of the four laws after exposure to the research-informed statement. Each law saw a net increase in support following exposure. The greatest increase was observed for laws that restricted vehicle infotainment systems (30.7% of respondents became more supportive). In addition, exposure to a research-informed statement was associated with increased support for laws allowing communities to introduce red-light cameras among 22.2% of respondents and 20% increased support for laws mandating bicycle helmets for children (see Table 4). While opinion movement about each law was largely positive following exposure to research data, there were also some respondents who became less supportive after such exposure. Decreased support (increased opposition) was most apparent in relation to mandatory ignition interlock laws (7.6% of respondents who were not already strongly opposed became more oppositional), followed
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
113
Table 2 Sample characteristics (weighted to the national sample). Characteristic
U.S. populationa
Weighted sample n
%
n
%
1176 1221
49.05 50.95
Total pop.: 309,349,689 153,267,861 158,324,058
49.2 50.8
Age 18–24 years 25–34 years 35–44 years 45–54 years 55–64 years 65–74 years 75–84 years 85 years and over
310 430 393 486 377 340 60 1
12.93 17.95 16.38 20.29 15.71 14.20 2.49 0.04
Population over 18: 237,681,218 31,209,548 41,540,346 40,827,710 44,653,387 38,065,201 22,489,229 13,197,352 5,698,445
13.1 17.5 17.2 18.8 16.0 9.5 5.6 2.4
Income Less than $34,999 $35,000–$74,999 $75,000–$149,999 $150,000 or more Missing income data
642 706 646 181 222
26.78 29.44 26.95 7.56 9.2
40,973,894 36,631,417 27,367,632 10,018,782 NA
35.7 31.8 23.8 8.7 NA
1017 694 686
42.45 28.94 28.61
87742516 59,838,341 58,890,813
42.5 29.0 28.5
Sex Male Female
Education High school or less Some college OR associate’s degree College grad or more a
Data acquired from American Community Survey 2011 (1-year estimate).
Table 3 Initial and research-informed opinions regarding road safety policies. Ignition interlock
Initial response Net support Strongly support Somewhat support Neither support or oppose Net oppose Somewhat Oppose Strongly Oppose
Bicycle helmet
Vehicle infotainment
Red light camera
n
%
n
%
n
%
n
%
1784 1165 618 272 341 182 159
74.4 48.6 25.8 11.3 14.2 7.6 6.6
1792 1321 471 315 290 149 141
74.8 55.1 19.6 13.1 12.1 6.2 5.9
1473 1033 440 346 578 229 349
61.4 43.1 18.3 14.4 24.1 9.6 14.6
1391 843 549 374 632 241 391
58.0 35.2 22.9 15.6 26.4 10.0 16.3
76.4 49.1 27.3 12.0 11.6 6.3 5.4
1859 1386 473 300 238 121 118
77.5 57.8 19.7 12.5 9.9 5.0 4.9
1717 1204 512 323 357 177 180
71.6 50.2 21.4 13.5 14.9 7.4 7.5
1508 843 549 374 509 241 391
62.9 35.2 22.9 15.6 21.2 10.0 16.3
Response after exposure to research-informed statement Net Support 1832 Strongly support 1177 655 Somewhat support 287 Neither Support or Oppose 278 Net oppose 150 Somewhat oppose 128 Strongly oppose
by laws pertaining to infotainment systems (5.9%), laws permitting community installation of red-light cameras (4.2%), and laws mandating bike helmets for children (2.4%). Consistent with the findings in Table 3, the paired analyses demonstrate significant increases toward supportive opinions of the laws when comparing responses to the initial and research-informed statements (Table 5).
4. Discussion This study adds to the literature by outlining the magnitude of support and opposition for four evidence-based road safety laws among a nationally representative sample of U.S. adults. Although this research focuses on laws in the U.S., the findings are also
Table 4 Describing movement in support and opposition.
Those who were not already strongly supportive (of total sample) People who became more supportive/less oppositional (of not already strongly supportive) Those who were not already strongly opposed (of total sample) People who became more opposed/less supportive (of not already strongly opposed)
Ignition interlock
Bicycle helmet
Vehicle infotainment
Red light camera
n
%
n
%
n
%
n
%
1232
51.4
1076
44.9
1364
56.9
1554
64.8
246
20.0
215
20.0
418
30.7
345
22.2
2238
93.4
2256
94.1
2048
85.4
2006
83.7
170
7.6
53
2.4
122
5.9
84
4.2
114
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
Table 5 Comparison of mean scores between initial and research-informed opinions regarding road safety laws. Topic of laws
Initial mean
Research-informed mean
Difference in means
Test statistic, p valuea
Ignition interlock Bicycle helmet Vehicle infotainment Red light camera
4.022 4.119 3.658 3.505
4.086 4.205 3.994 3.678
0.064 0.086 0.336 0.173
F(1, 2396) = 10.62, p = 0.001 F(1, 2396) = 43.27, p < 0.001 F(1, 2396) = 94.26, p < 0.001 F(1, 2396) = 69.55, p < 0.001
a
Analysis consisted of a weighted paired t-test, using svy command in Stata 12.1.
relevant to other counties. Globally there is a push for the enactment and enforcement of laws that address the key risks factors for road traffic injuries (speed, alcohol, motorcycle helmets, seat belts, and child restraints). The WHO Global Status Report on Road Safety mentions a need for political will to advance these laws, and understanding the public’s support for them, may bolster the political will needed for legislative initiatives. (World Health Organization, 2013) In this research, we consider whether support and opposition to the selected legislative change after exposure to a researchinformed statement either about the risk posed by a particular behavior or the efficacy of the law to reduce injury risk. We found that a majority (58–75%) of those surveyed supported the four presented laws. Moreover, supporters tend to be strong supporters. There was relatively little ambivalence about these laws, as evidenced by only 11–16% of respondents providing neutral responses. While rates of strong opposition to the laws were low, more people were more strongly opposed to laws related to infotainment systems (15%) and red light cameras (16%) than to bicycle helmets (6%) and ignition interlock systems (7%). The mean scores for each of the laws were generally supportive both in response to the initial and research-informed statements (between 3.5 and 4.2 on a 5-point scale). The increases in support seen when comparing the two responses, while statistically significant, were modest in size. For each of the laws, there were some respondents who became less supportive or more oppositional following exposure to the research-informed statement. It may be that the data presented suggested a smaller risk or less efficacious law than respondents had been assuming. We saw the highest level of increase in opposition (7.6%) for the alcohol interlock law. The statement provided included comparative recidivism rates with and without the device, and it may be that the level of risk reduction was not as much as people would have otherwise assumed. It is also possible that a statement related to a change in behavior is less impactful than one about absolute risk. Alternatively, this change could be that asking a second time caused people to consider the question more carefully and they were actually more opposed originally or in general, regardless of the research evidence, or it could simply be simple response variance. The current analysis only allows us to speculate about why research-informed statements would reduce support among a small (but not inconsequential) subset of respondents. Nevertheless, our results provide some potentially useful guidance for injury prevention professionals interested in advancing specific motor vehicle safety legislation. Motor vehicle injury prevention is one of the CDC’s Winnable Battles, and the data collected in this study may inform efforts to promote and implement legislation that not only are supported by evidence, but also have an additional benefit of being supported by the public. This is the first study to our knowledge that examined the impact of providing research evidence on a respondent’s opinion about a specific law. We found substantial benefits in terms of increased public support when presented with this information; 20–30% of U.S. adults became more supportive/less oppositional. Policies are informed by factors beyond research results, including whether a countermeasure is framed in relation to safety or mobility (Bates et al., 2010). Regardless, these results provided actionable information that can be used by practitioners and researchers alike. Our
results suggest there may be great utility in including evidence on risk and risk reduction as part of a legislative advocacy campaign. The findings also should encourage researchers to share their study findings with the public as well as with policymakers as a means of increasing public support for effective legislative solutions. 4.1. Limitations Certain limitations should be considered with regard to this study. All surveys and polls (both those that employ probability sampling and those that do not), are subject to multiple sources of error. Much of this error is difficult (or impossible) to quantify or estimate; sources of error include sampling error, non-response error, error associated with question wording and response options, and post-survey weighting and adjustments. More specific to our study, we acknowledge the limitations associated with generalizing from online surveys that inevitably exclude portions of the population who are not online. According to recent research (Zickuhr, 2013), 15% of Americans do not use the internet, and not being online is associated with being older, having lower education and lower income. Moreover, due to limited resources there were survey design compromises. For instance, the structure of the Quick Query poll did not allow us to manipulate the ordering of the questions so that laws were asked about in varying order. In order to facilitate weighting these data to provide a nationally representative sample, we exposed all respondents to the question about the law, followed by the statement and a follow-up question about the law. Our design may have been more robust had we conducted a between groups design and compared the opinion of those exposed to the data with those who had not been exposed. We would argue, however, that there is value in considering how exposure to risk or legislative efficacy information might change the opinion of a respondent. We are also aware that we cannot isolate the effect of repeated measures on the same individual as all respondents saw all 4 laws; however, any repeated measures bias would likely have the same effect across all four laws so it is valid to compare the proportions who support/oppose the evidenceinformed statements across the policies. Finally, we tested a limited number of laws and research-informed statements; it may be that certain theory-based types of evidence (e.g., personal susceptibility) would be more or less impactful on changing public opinion in a pro-safety direction, and that this may vary by the specific law under consideration.1 5. Conclusions The likelihood of adoption of effective policies is enhanced when we understand current public opinion, as well as what influences opinion change. These data demonstrate considerable support on the part of the U.S. adult population for a set of four road-related laws. Moreover, we established that exposure to the kinds of data produced through injury research can help strengthen public support further, pointing to the potential benefit to be gained from educational campaigns that utilize risk and efficacy data. There is
1
Harris only allows one individual per household to be included in its Online Poll.
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
some work still to do to better understand why a small subset of people become more oppositional when presented with information about risks faced or the law’s impact. However, we should be greatly heartened by the fact that in relation to these laws, injury control starts from a position of strength in terms of public attitudes that can serve to keep people safer when they are on the road. These results bode well for injury prevention efforts that seek to utilize legislation to reduce risk and enhance safety. References Aeron-Thomas, A., Hess, S., 2005. Red-light cameras for the prevention of road traffic crashes. Cochrane Database Syst. Rev. 2, http://dx.doi.org/10.1002/ 14651858.CD003862.pub2, Art. No.: CD003862. Bates, L., Watson, B., King, M., 2010. Mobility and safety are conflicting goals for transport policy makers when making decisions about graduated driver licensing. Int. J. Health Promot. Educ. 48 (2), 46–51. Centers for Disease Control and Prevention, 2012. Injuries and Violence Are Leading Causes of Death: Key Data & Statistics, Retrieved, 2013 from http://www.cdc.gov/injury/overview/data.html Debinski, B., Smith, K.C., Gielen, A.C., 2014. Public opinion on motor vehicle related injury prevention policies: a systematic review of a decade of research. Traffic Inj. Prev. 15 (3), 243–251. Dinh-Zarr, T.B., Sleet, D.A., Shults, R.A., Zaza, S., Elder, R.W., Nichols, J.L., Sosin, D.M., 2001. Reviews of evidence regarding interventions to increase the use of safety belts. Am. J. Prev. Med. 21 (4), 48–65. Girasek, D.C., 2013. Gauging popular support for traffic safety in the United States. Accid. Anal. Prev. 50, 1112–1117. Gold, M., 2009. Pathways to the use of health services research in policy. Health Serv. Res. 44 (4), 1111–1136. Harris Interactive, 2009. ESOMAR 26 Questions and Answers, Retrieved, 2013 from http://www.harrisinteractive.com/vault/HI Corp ESOMAR26 QandA.pdf Hirsch, P., 2003. Adolescent driver risk taking and driver education: evidence of a mobility bias in public policymaking. J. Saf. Res. 34 (3), 289–298. Institute of Medicine (US). Committee on Assuring the Health of the Public in the 21st Century, 2003. The Future of the Public’s Health in the 21st Century. National Academies Press, Washington, DC. Levi, J., Segal, L., Kohn, D., 2012. The Facts Hurt: A State-by-State Injury Prevention Policy Report. Trust for America’s Health, Washington, DC.
115
Liu, B.C., Ivers, R., Norton, R., Boufous, S., Blows, S., Lo, S.K., 2008. Helmets for preventing injury in motorcycle riders. Cochrane Database Syst. Rev. 1, http://dx.doi.org/10.1002/14651858.CD004333.pub3, Art. No.: CD004333. Macpherson, A., Spinks, A., 2008. Cochrane review: bicycle helmet legislation for the uptake of helmet use and prevention of head injuries. Evid. Based Child Health Cochrane Rev. J. 3 (1), 16–32. Murray, C.J., Vos, T., Lozano, R., Naghavi, M., Flaxman, A.D., Michaud, C., Abdalla, S., 2013. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the global burden of disease study 2010. Lancet 380 (9859), 2197–2223. Page, B.I., Shapiro, R.Y., 1983. Effects of public opinion on policy. Am. Polit. Sci. Rev. 77 (1), 175–190. Russell, K.F., Vandermeer, B., Hartling, L., 2011. Graduated driver licensing for reducing motor vehicle crashes among young drivers. Cochrane Database Syst. Rev. 10, http://dx.doi.org/10.1002/14651858.CD003300.pub3, Art. No.: CD003300. Sabatier, P.A., 1991. Toward better theories of the policy process. Polit. Sci. Polit. 24 (2), 147–156. Shults, R.A., Bergen, G., 2013. Attitudes towards requiring ignition interlocks for all driving while intoxicated offenders: findings from the 2010 HealthStyles survey. Inj. Prev. 19 (1), 68–71. Thomas, S., Acton, C., Nixon, J., Battistutta, D., Pitt, W.R., Clark, R., 1994. Effectiveness of bicycle helmets in reducing head injury in children: case-control study. British Medical Journal 308, 173. Thompson, D.C., Rivara, F., 1999. Thompson R. Helmets for preventing head and facial injuries in bicyclists. Cochrane Database of Systematic Reviews 4, http://dx.doi.org/10.1002/14651858, Art. No.: CD001855. Willis, C., Lybrand, S., Bellamy, N., 2004. Alcohol ignition interlock programmes for reducing drink driving recidivism. Cochrane Database Syst Rev. 18 (4), CD004168. Willis, G.B., 2005. Cognitive Interviewing: A Tool for Improving Questionnaire Design. Sage, Thousand Oaks, CA. Wilson, C., Willis, C., Hendrikz, J.K., Le Brocque, R., Bellamy, N., 2010. Speed cameras for the prevention of road traffic injuries and deaths. Cochrane Database Syst. Rev. 11 (10), Article CD004607. World Health Organization, 2013. WHO Global Status Report on Road Safety 2013: Supporting a Decade of Action. World Health Organization, http://www.who.int/violence injury prevention/road safety status/2013/en/ (accessed 10.07.14). Zickuhr, K., 2013. Who’s Not Online and Why. Pew Internet & American Life Project, http://www.pewinternet.org/2000/09/21/whos-not-online/ (accessed 07.10.14).
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Research-informed evidence and support for road safety legislation: Findings from a national survey Katherine Clegg Smith a,∗ , Beata Debinski a,b , Keshia Pollack a , Jon Vernick a , Stephen Bowman a,c , Alicia Samuels a , Andrea Gielen a a
Johns Hopkins Center for Injury Research and Policy, Johns Hopkins Bloomberg School of Public Health, United States Wake Forest School of Medicine, Department of Social Sciences and Health Policy, United States c University of Arkansas for Health Sciences, Fay W. Boozman College of Public Health, Department of Health Policy and Management, United States b
a r t i c l e
i n f o
Article history: Received 31 January 2014 Received in revised form 17 August 2014 Accepted 18 August 2014 Keywords: Injury Policy Legislation Research evidence Public opinion Survey
a b s t r a c t Public opinion is influential in the policymaking process, making it important to understand the factors that influence popular support or opposition to public health policies. Researchers and policymakers tend to agree that scientific evidence can inform decision-making, but this influence has not been explored sufficiently, especially in the area of injury prevention. This paper considers the potential for the communication of evidence-based research and public health data to influence opinion about legislation that could reduce road-related injury. We conducted a nationally-representative online survey to assess public attitudes toward four road-safety laws; ignition interlock, school zone red-light cameras, restrictions on infotainment systems, and children’s bicycle helmets. For each law, we assessed initial support and then provided a research-informed statistic on either the injury risk posed or the law’s efficacy reducing risk and re-examined the law’s support or opposition. The survey was completed by 2397 U.S. adults. Each law was initially supported by a majority of respondents, with greatest support for ignition interlock (74.4%) and children’s bicycle helmets (74.8%). Exposure to research-informed statements increased legislative support for 20–30% of respondents. Paired analyses demonstrate significant increases toward supportive opinions when comparing responses to the initial and research-informed statements. The study demonstrates considerable public support for evidence-based road-related laws. Overall support was augmented by exposure to research data. Injury prevention practitioners can capitalize on this support in efforts to build support for legislation that would prevent injury. Researchers should be encouraged to expand their efforts to share research results with both the public and policymakers. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Considerable progress has been made in reducing motor vehicle crash (MVC) deaths and injuries around the world. Still, motor vehicle crashes remain the leading cause of death for Americans ages 5–24 years, and, among all unintentional injury deaths, they are the leading cause for the entire population (Centers for Disease Control and Prevention, 2012; Levi et al., 2012). In 2011, there were approximately 32,000 road-related fatalities (including both occupants and non-occupants) and over 2.6 million vehicle occupants were treated in emergency departments for nonfatal injuries sustained in a vehicle crash (Centers for Disease Control and Prevention, 2012). Approximately one in three deaths on the road involve an alcohol impaired driver (Levi et al., 2012; Shults and Bergen, 2013), and
∗ Corresponding author. Tel.: +1 4105020025. E-mail address: [email protected] (K.C. Smith). http://dx.doi.org/10.1016/j.aap.2014.08.016 0001-4575/© 2014 Elsevier Ltd. All rights reserved.
current estimates are that about one in six fatal crashes involves a distracted driver (Levi et al., 2012). Globally, estimates from the Global Burden of Disease Report show the burden of road traffic injuries. Between 1990 and 2010, road traffic injuries went from the 12th to 10th rank (a 34% increase) among the leading causes of the global burden of disease (Murray et al., 2013). By 2030, estimates are that road traffic injuries will be the 5th leading causes of death in the world (World Health Organization, 2013). Policies, such as legislation and regulation, that prohibit risky behaviors and/or products can be effective in shifting social norms, changing behavior, and preventing injuries and fatalities. Policies often provide incentives for people to follow safe practices and in so doing, protect themselves and others (Levi et al., 2012). For example, effective policies to reduce motor vehicle crashes and resulting death and injury include primary enforcement of seat belt laws (Dinh-Zarr et al., 2001), installation of ignition interlocks for convicted drunk drivers (Willis, 2005), mandatory motorcycle (Liu et al., 2008), and bicycle helmet use (Macpherson and Spinks, 2008),
110
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
graduated driver licensing (Russell et al., 2011), installation of redlight cameras (Aeron-Thomas and Hess, 2005), and speed cameras (Institute of Medicine. Committee on Assuring the Health of the Public in the 21st Century, 2003; Wilson et al., 2010). Despite the evidence base regarding the efficacy of public policy in reducing road-related injury, a recent report indicates low state-level utilization of legislation (Levi et al., 2012) for preventing injury and death among road users. For example, as of fall 2012 (when our survey data were collected), only 21 states and the District of Columbia (DC) required children to wear bicycle helmets; 17 states and Washington D.C. had comprehensive, primary seat belt laws; and 16 states required mandatory ignition interlocks for all convicted drunk drivers, including first-time offenders (Levi et al., 2012). There are many factors that can influence road safety policies including (but not limited to) the framing of such policies in relation to road safety or mobility (Bates et al., 2010; Hirsch, 2003). Public opinion changes can also be important factors in influencing the legislative process (Page and Shapiro, 1983), and public opinion change often (but not always) precedes policy change. It is therefore critical to understand the factors that lead people to become more supportive of policies that have been shown to be effective in preventing road-related injury. Prior work has demonstrated the potential impact of substantive information on the magnitude of a problem on policy stakeholders’ consideration of potential solutions (Sabatier, 1991). Moreover, research findings can serve to define a problem and frame potential solutions (Gold, 2009). The question emerges, therefore: Can public opinion regarding effective road-related laws be impacted by providing research-informed statements about such laws? A recent review of the literature (Debinski et al., 2014) found 26 studies and reports addressing public opinion about legislation pertaining to motor vehicle and bicycle-related topics (e.g., allterrain vehicles, automated enforcement with red-light and speed cameras, distracted driving, drinking and driving, graduated driver licensing, bicycle and motorcycle helmets, and seat belts). Results suggested that public opinion toward injury prevention legislation is generally favorable, but few studies reported on factors that might serve to influence support or opposition for such initiatives. Understanding public perceptions and influences can inform consideration of the discourse about legislation so as to better protect the public from preventable injury – both on the road and more broadly. The present study seeks to address two key gaps in the current road safety literature. Our first aim is to document the magnitude of support and opposition for four laws among a nationally representative sample of U.S. adults. Second, we examine whether support and opposition to the selected laws is impacted by the provision of factual information on risk or efficacy of the given law in relation to injury risk. Addressing these questions will demonstrate the value of research evidence in shaping public opinion, and may facilitate public information efforts focused on road-related risks, responsible road use practices, and effective laws. We explore whether research-informed statements that are supportive of the selected laws are associated with an increased level of public support (and reduced opposition) for that law.
cameras in school zones as a structural/community-level initiative, and laws pertaining to in-motion infotainment system disablement, so as to represent emerging safety concerns. We chose laws that have at least some evidence supporting their effectiveness in reducing injury risk. The resulting survey included questions pertaining to the four specific laws: (1) mandatory ignition interlock installation for all driving under the influence (DUI) offenders even after their first offense; (2) a mandatory bicycle helmet law for all children under 16 years old; (3) allowing communities to install red light cameras in school zones to help improve pedestrian safety; and (4) requiring that in-vehicle infotainment systems ‘lock’ when the car is moving. There have been five prior studies and reports addressing public opinion regarding mandatory installation of ignition interlocks, and one study about bicycle helmets for riders under 16 specifically, but no data have as yet been reported about support for red light cameras in school zones or restrictions to infotainment system operation (Debinski et al., 2014). With regard to the influence of evidence on public opinion, Girasek (2013) found that the public strongly endorsed the concept that they “expect evidence-based traffic safety policies”. For each of the selected laws, we used the most reliable evidence available to provide a relevant research-informed risk or risk reduction statement. These included two statements that addressed the magnitude of the injury risk posed and two that addressed the efficacy of the law in reducing injury risk. Each statement presents injury risk and/or reduction differently. Statistics were obtained from the CDC Injury and Violence Prevention and Control site, the CDC Web-based Injury Statistics Query and Reporting System (WISQARS), and reviews published in the Cochrane Database of Systematic Reviews (see Table 1). 2.2. Data collection Our data are the result of a nationally representative online survey conducted over two days (October 29–30, 2012). We contracted with the polling firm, Harris Interactive Inc. to include a set of road-related questions about specific laws in a survey with U.S. adults (ages 18 and older) via its Harris Poll Quick Query omnibus product. An online format was chosen over a telephone survey to yield a greater number of responses in a more cost-efficient manner. The Harris Poll OnlineSM Panel members are recruited for surveys using a “double opt-in” process, which requires that each registrant confirms participation in a panel by clicking on a link within an email that is sent to their email address, after voluntarily registering to be a part of the Panel. The survey is designed to permit only one respondent per household. Harris Interactive generates a nationally representative sample using a two-stage approach, first balancing the sample that will receive a survey invitation, and then weighting collected data on at least eight demographic factors (Harris Interactive, 2009). In addition to our set of questions, we also collected demographic information about each respondent including residence by region of country, age, sex, marital status, highest completed education level, annual household income, household size, employment status, and whether or not there are any children under 18 in the home.
2. Methods 2.3. Survey items 2.1. Topics We selected four laws that pertain to road safety, two of which (alcohol ignition interlock and bicycle helmets), were selected from the Trust For America’s Health report of state-level injury indicators (Levi et al., 2012). We also selected laws that require red light
All respondents were asked to indicate support or opposition to each of the four sample laws. In the online survey, each law and the questions related to it were presented individually and in turn to each respondent. Each law was introduced through a brief descriptive text accompanied by an illustrative photo
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
111
Table 1 Questions posed and research-informed statement provided. Topic
Question posed
Evidence provided and source
Alcohol ignition interlock
Would you support or oppose a law in your state that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law would include people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car Would you support or oppose a law in your state that requires all children under 16 to wear helmets when riding a bicycle?
Research shows that when an alcohol ignition is on the car, the driver’s risk of being caught again for drunk driving decreases to 1/3 of what the risk would be without the interlock (Willis et al., 2004)
Bicycle helmets
Vehicle infotainment system
Would you support or oppose a law requiring that infotainment systems cannot work when the car is moving?
Red light camera
Would you support or oppose a law in your state that allows communities to install red light cameras in school zones to help improve pedestrian safety?
Research shows that bicycle helmets provide a 63% reduction in head injury and 86% reduction in brain injury in children younger than 15 (Thomas et al., 1994; Thompson et al., 1999) Research shows that in the U.S., 2 out of 10 serious crashes involve distracted driving, resulting in over 5000 deaths and 440,000 injuries per year (NHTSA, 2009; CDC, 2012) Every year there are about 43,000 injuries from motor vehicle crashes with pedestrians younger than 18 (CDC WISQARS, 2012)
1. An alcohol ignition interlock is a device that can be attached to a car to determine if the driver has been drinking alcohol. The alcohol ignition interlock works by having the driver breathe into the device. If a preset level of alcohol is detected, the car will not start.
Photo: Colourbox.com 2. Would you support or oppose a law in your state that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law would include people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car. Strongly oppose Somewhat oppose Neither support nor oppose Somewhat support Strongly support Research shows that when an alcohol ignition is on the car, the driver’s risk of being caught again for drunk driving decreases to 1/3 of what the risk would be without the interlock. Based on this information, would you support or oppose a law in your state that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law would include people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car. Strongly oppose Somewhat oppose Neither support nor oppose Somewhat support Strongly support 3. Does the state that you live in have a law that requires everyone who is found guilty of drunk driving to install an alcohol ignition interlock in their vehicle? This law includes people who have only been convicted of drunk driving once. The state would decide how long the interlock has to stay on the car. Yes No I don’t know Fig. 1. Ignition interlock survey questions.
112
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
(see Fig. 1 for an example of the question presentation). The poll questions and research-informed statements were assessed for face validity through cognitive interviewing with a convenience sample of fifteen participants. Cognitive interviewing is a qualitative method used to identify potential problems with survey questions where the survey is administered and then respondents are probed on various aspects of the questions (Willis, 2005). For our study, interviewees were shown the statements and accompanying photos and were asked to respond to the questions pertaining to each topic, and then asked to orally describe their understanding of the topic, including whether their understanding was supported by the text and/or picture; to interpret the law that was being addressed and explain how they made their decision; and to assess what they thought about the piece of evidence that was provided in the second question of each set. After completion of the cognitive interviews and consideration of issues raised by participants, modifications were made to the survey questions. Modifications included replacing pictures that were not beneficial for understanding the topic, and clarifying that each state would be responsible for specifying details of a given law. After being presented with the introduction and the photo, respondents were asked to indicate whether they would support or oppose each law. Possible responses were on a 5-point Likert scale, ranging from “strongly oppose” to “strongly support.” After responding to the initial question, respondents were prompted to the next screen where they were provided with a research-informed statement about the law (See Table 1). They were once again asked to indicate support or opposition for the same law, and were presented with the same five response options. It was not possible for respondents to go back and change their answers to the initial question after being presented with the research-informed statement. This research was reviewed and approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board.
question was coded as becoming more oppositional. To examine the proportion of participants who became more supportive or less oppositional, we divided the total number of people who changed their response in a positive direction by the total sample and subtracted the number of people who were strongly supportive in response to the initial question. Those who were initially strongly supportive were removed from the denominator because they were not eligible to become any more supportive in response to the evidence informed decision. To examine the proportion of participants who became more oppositional or less supportive, the same procedure was followed, and those who were strongly opposed in response to the first question were eliminated from the denominator. In addition to comparing the proportions of respondents by level of support and opposition after the initial and research-informed statements, we also conducted a paired analysis of the mean scores before and after exposure to the research-informed statement. 3. Results 3.1. Sample The Harris Omnibus Survey (Quick Query) was completed by 2397 respondents in the two-day window that it was available to panel members. Characteristics of the sample weighted to the U.S. population are presented in Table 2. The proportion of male and female respondents was roughly equal. The largest percentage (20%) of respondents was 45–54 years of age, with few older adults (>75 years) participating. More than half (56%) of respondents reported a family income of less than $75,000 and few (8%) reported an income of $150,000 or more. More than a quarter (29%) of respondents had at least a college degree (see Table 2). 3.2. Initial support and opposition
2.4. Statistical analysis Weighted data were analyzed using Stata 12.1 (StataCorp LP, College Station, TX). Harris Interactive assigned each respondent a weight for their contribution to the nationally representative sample (based on March 2011 Current Population Survey data). A survey design was declared for the dataset and the weight provided by Harris Interactive was designated as a probability, or sampling, weight. All Stata estimation commands subsequently utilized the ‘svy’ prefix, which accounts for the complex survey sampling. We report overall support and opposition for each law as well as the levels of strong support and strong opposition for each. We also report the proportion of ‘supportive’ movement and ‘oppositional’ movement for each law following exposure to the research-informed statement. We present data pertaining to each of the five response categories (“strongly oppose” to “strongly support”). Strong opposition was scored as 1, and strong opposition was scored as 5, such that a higher mean indicates higher support. A variable for “net support” and “net oppose” was created by condensing the “strongly” and “somewhat” responses for both support and opposition. A variable to describe movement in responses from the initial question to the research-informed second question was generated using the five original categories. Individuals who provided the same response for both questions were considered to have not moved or changed opinion. Any response to the second, research-informed question that was more supportive or less oppositional than the original response was considered a change in the positive direction. Similarly, any response to the second question that was less supportive or more oppositional than the response to the initial
Table 3 displays the proportion of respondents who initially supported and opposed each of the four laws in the study. Each of the four laws was supported by the majority of respondents, with the highest net support for ignition interlocks and bicycle helmet laws for children (74.4 and 74.8% respectively), followed by vehicle infotainment system laws (61.4%). There was least initial support for red light cameras laws (58.0%). The highest proportion of strong opposition was seen in response to red light camera laws (16.3%) and vehicle infotainment system laws (14.6%), which is more than double the proportions who strongly opposed ignition interlock laws (6.6%) and bicycle helmet laws (5.9%). 3.3. Research-informed support and opposition Table 3 also displays the proportion of respondents who supported and opposed each of the four laws after exposure to the research-informed statement. Each law saw a net increase in support following exposure. The greatest increase was observed for laws that restricted vehicle infotainment systems (30.7% of respondents became more supportive). In addition, exposure to a research-informed statement was associated with increased support for laws allowing communities to introduce red-light cameras among 22.2% of respondents and 20% increased support for laws mandating bicycle helmets for children (see Table 4). While opinion movement about each law was largely positive following exposure to research data, there were also some respondents who became less supportive after such exposure. Decreased support (increased opposition) was most apparent in relation to mandatory ignition interlock laws (7.6% of respondents who were not already strongly opposed became more oppositional), followed
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
113
Table 2 Sample characteristics (weighted to the national sample). Characteristic
U.S. populationa
Weighted sample n
%
n
%
1176 1221
49.05 50.95
Total pop.: 309,349,689 153,267,861 158,324,058
49.2 50.8
Age 18–24 years 25–34 years 35–44 years 45–54 years 55–64 years 65–74 years 75–84 years 85 years and over
310 430 393 486 377 340 60 1
12.93 17.95 16.38 20.29 15.71 14.20 2.49 0.04
Population over 18: 237,681,218 31,209,548 41,540,346 40,827,710 44,653,387 38,065,201 22,489,229 13,197,352 5,698,445
13.1 17.5 17.2 18.8 16.0 9.5 5.6 2.4
Income Less than $34,999 $35,000–$74,999 $75,000–$149,999 $150,000 or more Missing income data
642 706 646 181 222
26.78 29.44 26.95 7.56 9.2
40,973,894 36,631,417 27,367,632 10,018,782 NA
35.7 31.8 23.8 8.7 NA
1017 694 686
42.45 28.94 28.61
87742516 59,838,341 58,890,813
42.5 29.0 28.5
Sex Male Female
Education High school or less Some college OR associate’s degree College grad or more a
Data acquired from American Community Survey 2011 (1-year estimate).
Table 3 Initial and research-informed opinions regarding road safety policies. Ignition interlock
Initial response Net support Strongly support Somewhat support Neither support or oppose Net oppose Somewhat Oppose Strongly Oppose
Bicycle helmet
Vehicle infotainment
Red light camera
n
%
n
%
n
%
n
%
1784 1165 618 272 341 182 159
74.4 48.6 25.8 11.3 14.2 7.6 6.6
1792 1321 471 315 290 149 141
74.8 55.1 19.6 13.1 12.1 6.2 5.9
1473 1033 440 346 578 229 349
61.4 43.1 18.3 14.4 24.1 9.6 14.6
1391 843 549 374 632 241 391
58.0 35.2 22.9 15.6 26.4 10.0 16.3
76.4 49.1 27.3 12.0 11.6 6.3 5.4
1859 1386 473 300 238 121 118
77.5 57.8 19.7 12.5 9.9 5.0 4.9
1717 1204 512 323 357 177 180
71.6 50.2 21.4 13.5 14.9 7.4 7.5
1508 843 549 374 509 241 391
62.9 35.2 22.9 15.6 21.2 10.0 16.3
Response after exposure to research-informed statement Net Support 1832 Strongly support 1177 655 Somewhat support 287 Neither Support or Oppose 278 Net oppose 150 Somewhat oppose 128 Strongly oppose
by laws pertaining to infotainment systems (5.9%), laws permitting community installation of red-light cameras (4.2%), and laws mandating bike helmets for children (2.4%). Consistent with the findings in Table 3, the paired analyses demonstrate significant increases toward supportive opinions of the laws when comparing responses to the initial and research-informed statements (Table 5).
4. Discussion This study adds to the literature by outlining the magnitude of support and opposition for four evidence-based road safety laws among a nationally representative sample of U.S. adults. Although this research focuses on laws in the U.S., the findings are also
Table 4 Describing movement in support and opposition.
Those who were not already strongly supportive (of total sample) People who became more supportive/less oppositional (of not already strongly supportive) Those who were not already strongly opposed (of total sample) People who became more opposed/less supportive (of not already strongly opposed)
Ignition interlock
Bicycle helmet
Vehicle infotainment
Red light camera
n
%
n
%
n
%
n
%
1232
51.4
1076
44.9
1364
56.9
1554
64.8
246
20.0
215
20.0
418
30.7
345
22.2
2238
93.4
2256
94.1
2048
85.4
2006
83.7
170
7.6
53
2.4
122
5.9
84
4.2
114
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
Table 5 Comparison of mean scores between initial and research-informed opinions regarding road safety laws. Topic of laws
Initial mean
Research-informed mean
Difference in means
Test statistic, p valuea
Ignition interlock Bicycle helmet Vehicle infotainment Red light camera
4.022 4.119 3.658 3.505
4.086 4.205 3.994 3.678
0.064 0.086 0.336 0.173
F(1, 2396) = 10.62, p = 0.001 F(1, 2396) = 43.27, p < 0.001 F(1, 2396) = 94.26, p < 0.001 F(1, 2396) = 69.55, p < 0.001
a
Analysis consisted of a weighted paired t-test, using svy command in Stata 12.1.
relevant to other counties. Globally there is a push for the enactment and enforcement of laws that address the key risks factors for road traffic injuries (speed, alcohol, motorcycle helmets, seat belts, and child restraints). The WHO Global Status Report on Road Safety mentions a need for political will to advance these laws, and understanding the public’s support for them, may bolster the political will needed for legislative initiatives. (World Health Organization, 2013) In this research, we consider whether support and opposition to the selected legislative change after exposure to a researchinformed statement either about the risk posed by a particular behavior or the efficacy of the law to reduce injury risk. We found that a majority (58–75%) of those surveyed supported the four presented laws. Moreover, supporters tend to be strong supporters. There was relatively little ambivalence about these laws, as evidenced by only 11–16% of respondents providing neutral responses. While rates of strong opposition to the laws were low, more people were more strongly opposed to laws related to infotainment systems (15%) and red light cameras (16%) than to bicycle helmets (6%) and ignition interlock systems (7%). The mean scores for each of the laws were generally supportive both in response to the initial and research-informed statements (between 3.5 and 4.2 on a 5-point scale). The increases in support seen when comparing the two responses, while statistically significant, were modest in size. For each of the laws, there were some respondents who became less supportive or more oppositional following exposure to the research-informed statement. It may be that the data presented suggested a smaller risk or less efficacious law than respondents had been assuming. We saw the highest level of increase in opposition (7.6%) for the alcohol interlock law. The statement provided included comparative recidivism rates with and without the device, and it may be that the level of risk reduction was not as much as people would have otherwise assumed. It is also possible that a statement related to a change in behavior is less impactful than one about absolute risk. Alternatively, this change could be that asking a second time caused people to consider the question more carefully and they were actually more opposed originally or in general, regardless of the research evidence, or it could simply be simple response variance. The current analysis only allows us to speculate about why research-informed statements would reduce support among a small (but not inconsequential) subset of respondents. Nevertheless, our results provide some potentially useful guidance for injury prevention professionals interested in advancing specific motor vehicle safety legislation. Motor vehicle injury prevention is one of the CDC’s Winnable Battles, and the data collected in this study may inform efforts to promote and implement legislation that not only are supported by evidence, but also have an additional benefit of being supported by the public. This is the first study to our knowledge that examined the impact of providing research evidence on a respondent’s opinion about a specific law. We found substantial benefits in terms of increased public support when presented with this information; 20–30% of U.S. adults became more supportive/less oppositional. Policies are informed by factors beyond research results, including whether a countermeasure is framed in relation to safety or mobility (Bates et al., 2010). Regardless, these results provided actionable information that can be used by practitioners and researchers alike. Our
results suggest there may be great utility in including evidence on risk and risk reduction as part of a legislative advocacy campaign. The findings also should encourage researchers to share their study findings with the public as well as with policymakers as a means of increasing public support for effective legislative solutions. 4.1. Limitations Certain limitations should be considered with regard to this study. All surveys and polls (both those that employ probability sampling and those that do not), are subject to multiple sources of error. Much of this error is difficult (or impossible) to quantify or estimate; sources of error include sampling error, non-response error, error associated with question wording and response options, and post-survey weighting and adjustments. More specific to our study, we acknowledge the limitations associated with generalizing from online surveys that inevitably exclude portions of the population who are not online. According to recent research (Zickuhr, 2013), 15% of Americans do not use the internet, and not being online is associated with being older, having lower education and lower income. Moreover, due to limited resources there were survey design compromises. For instance, the structure of the Quick Query poll did not allow us to manipulate the ordering of the questions so that laws were asked about in varying order. In order to facilitate weighting these data to provide a nationally representative sample, we exposed all respondents to the question about the law, followed by the statement and a follow-up question about the law. Our design may have been more robust had we conducted a between groups design and compared the opinion of those exposed to the data with those who had not been exposed. We would argue, however, that there is value in considering how exposure to risk or legislative efficacy information might change the opinion of a respondent. We are also aware that we cannot isolate the effect of repeated measures on the same individual as all respondents saw all 4 laws; however, any repeated measures bias would likely have the same effect across all four laws so it is valid to compare the proportions who support/oppose the evidenceinformed statements across the policies. Finally, we tested a limited number of laws and research-informed statements; it may be that certain theory-based types of evidence (e.g., personal susceptibility) would be more or less impactful on changing public opinion in a pro-safety direction, and that this may vary by the specific law under consideration.1 5. Conclusions The likelihood of adoption of effective policies is enhanced when we understand current public opinion, as well as what influences opinion change. These data demonstrate considerable support on the part of the U.S. adult population for a set of four road-related laws. Moreover, we established that exposure to the kinds of data produced through injury research can help strengthen public support further, pointing to the potential benefit to be gained from educational campaigns that utilize risk and efficacy data. There is
1
Harris only allows one individual per household to be included in its Online Poll.
K.C. Smith et al. / Accident Analysis and Prevention 73 (2014) 109–115
some work still to do to better understand why a small subset of people become more oppositional when presented with information about risks faced or the law’s impact. However, we should be greatly heartened by the fact that in relation to these laws, injury control starts from a position of strength in terms of public attitudes that can serve to keep people safer when they are on the road. These results bode well for injury prevention efforts that seek to utilize legislation to reduce risk and enhance safety. References Aeron-Thomas, A., Hess, S., 2005. Red-light cameras for the prevention of road traffic crashes. Cochrane Database Syst. Rev. 2, http://dx.doi.org/10.1002/ 14651858.CD003862.pub2, Art. No.: CD003862. Bates, L., Watson, B., King, M., 2010. Mobility and safety are conflicting goals for transport policy makers when making decisions about graduated driver licensing. Int. J. Health Promot. Educ. 48 (2), 46–51. Centers for Disease Control and Prevention, 2012. Injuries and Violence Are Leading Causes of Death: Key Data & Statistics, Retrieved, 2013 from http://www.cdc.gov/injury/overview/data.html Debinski, B., Smith, K.C., Gielen, A.C., 2014. Public opinion on motor vehicle related injury prevention policies: a systematic review of a decade of research. Traffic Inj. Prev. 15 (3), 243–251. Dinh-Zarr, T.B., Sleet, D.A., Shults, R.A., Zaza, S., Elder, R.W., Nichols, J.L., Sosin, D.M., 2001. Reviews of evidence regarding interventions to increase the use of safety belts. Am. J. Prev. Med. 21 (4), 48–65. Girasek, D.C., 2013. Gauging popular support for traffic safety in the United States. Accid. Anal. Prev. 50, 1112–1117. Gold, M., 2009. Pathways to the use of health services research in policy. Health Serv. Res. 44 (4), 1111–1136. Harris Interactive, 2009. ESOMAR 26 Questions and Answers, Retrieved, 2013 from http://www.harrisinteractive.com/vault/HI Corp ESOMAR26 QandA.pdf Hirsch, P., 2003. Adolescent driver risk taking and driver education: evidence of a mobility bias in public policymaking. J. Saf. Res. 34 (3), 289–298. Institute of Medicine (US). Committee on Assuring the Health of the Public in the 21st Century, 2003. The Future of the Public’s Health in the 21st Century. National Academies Press, Washington, DC. Levi, J., Segal, L., Kohn, D., 2012. The Facts Hurt: A State-by-State Injury Prevention Policy Report. Trust for America’s Health, Washington, DC.
115
Liu, B.C., Ivers, R., Norton, R., Boufous, S., Blows, S., Lo, S.K., 2008. Helmets for preventing injury in motorcycle riders. Cochrane Database Syst. Rev. 1, http://dx.doi.org/10.1002/14651858.CD004333.pub3, Art. No.: CD004333. Macpherson, A., Spinks, A., 2008. Cochrane review: bicycle helmet legislation for the uptake of helmet use and prevention of head injuries. Evid. Based Child Health Cochrane Rev. J. 3 (1), 16–32. Murray, C.J., Vos, T., Lozano, R., Naghavi, M., Flaxman, A.D., Michaud, C., Abdalla, S., 2013. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the global burden of disease study 2010. Lancet 380 (9859), 2197–2223. Page, B.I., Shapiro, R.Y., 1983. Effects of public opinion on policy. Am. Polit. Sci. Rev. 77 (1), 175–190. Russell, K.F., Vandermeer, B., Hartling, L., 2011. Graduated driver licensing for reducing motor vehicle crashes among young drivers. Cochrane Database Syst. Rev. 10, http://dx.doi.org/10.1002/14651858.CD003300.pub3, Art. No.: CD003300. Sabatier, P.A., 1991. Toward better theories of the policy process. Polit. Sci. Polit. 24 (2), 147–156. Shults, R.A., Bergen, G., 2013. Attitudes towards requiring ignition interlocks for all driving while intoxicated offenders: findings from the 2010 HealthStyles survey. Inj. Prev. 19 (1), 68–71. Thomas, S., Acton, C., Nixon, J., Battistutta, D., Pitt, W.R., Clark, R., 1994. Effectiveness of bicycle helmets in reducing head injury in children: case-control study. British Medical Journal 308, 173. Thompson, D.C., Rivara, F., 1999. Thompson R. Helmets for preventing head and facial injuries in bicyclists. Cochrane Database of Systematic Reviews 4, http://dx.doi.org/10.1002/14651858, Art. No.: CD001855. Willis, C., Lybrand, S., Bellamy, N., 2004. Alcohol ignition interlock programmes for reducing drink driving recidivism. Cochrane Database Syst Rev. 18 (4), CD004168. Willis, G.B., 2005. Cognitive Interviewing: A Tool for Improving Questionnaire Design. Sage, Thousand Oaks, CA. Wilson, C., Willis, C., Hendrikz, J.K., Le Brocque, R., Bellamy, N., 2010. Speed cameras for the prevention of road traffic injuries and deaths. Cochrane Database Syst. Rev. 11 (10), Article CD004607. World Health Organization, 2013. WHO Global Status Report on Road Safety 2013: Supporting a Decade of Action. World Health Organization, http://www.who.int/violence injury prevention/road safety status/2013/en/ (accessed 10.07.14). Zickuhr, K., 2013. Who’s Not Online and Why. Pew Internet & American Life Project, http://www.pewinternet.org/2000/09/21/whos-not-online/ (accessed 07.10.14).
