This article was downloaded by: [Umeå University Library] On: 01 April 2015, At: 12:28 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Traffic Injury Prevention Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/gcpi20
Effectiveness of Social Host and Fake Identification Laws on Reducing Underage Drinking Driver Fatal Crashes a
a
a
James C. Fell , Michael Scherer , Sue Thomas & Robert B. Voas
a
a
Pacific Institute for Research and Evaluation, Calverton, Maryland Published online: 11 Oct 2014.
Click for updates To cite this article: James C. Fell, Michael Scherer, Sue Thomas & Robert B. Voas (2014) Effectiveness of Social Host and Fake Identification Laws on Reducing Underage Drinking Driver Fatal Crashes, Traffic Injury Prevention, 15:sup1, S64-S73, DOI: 10.1080/15389588.2014.928929 To link to this article: http://dx.doi.org/10.1080/15389588.2014.928929
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Traffic Injury Prevention (2014) 15, S64–S73 C Taylor & Francis Group, LLC Copyright ISSN: 1538-9588 print / 1538-957X online DOI: 10.1080/15389588.2014.928929
Effectiveness of Social Host and Fake Identification Laws on Reducing Underage Drinking Driver Fatal Crashes JAMES C. FELL, MICHAEL SCHERER, SUE THOMAS, and ROBERT B. VOAS Pacific Institute for Research and Evaluation, Calverton, Maryland
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Received 17 March 2014, Accepted 24 May 2014
Objective: The public generally assumes that the minimum legal drinking age of 21 (MLDA-21) legislation in the United States is embodied in a single law and therefore all states have the same law. Actually, the MLDA-21 state laws consist of multiple provisions that support the core MLDA-21 laws and include a family of policies directed at controlling underage drinking and underage drinking and driving. Because social host and fake identification laws have recently garnered interest by policy makers in the states, this study was designed to determine their effectiveness. Methods: The effective dates for 2 types of social host laws and 3 fake identification laws were documented using the Alcohol Policy Information System (APIS), the Substance Abuse and Mental Health Services Administration’s (SAMHSA) 2011 Report to Congress on the Prevention and Reduction of Underage Drinking, and legal research tools. These laws include social host prohibitions (SHPs) and social host civil liability (SHCL), the use of fake identification (FID), retailer support for FID, and transfer/production of FID. We used a pre–post design to evaluate the influence on underage drinking-and-driving fatal crashes of these 5 laws using the Fatality Analysis Reporting System (FARS) data set for the years 1982 through 2010. The data were analyzed using structural equation modeling (SEM) controlling for as many variables as possible. Results: For those drivers younger than aged 21 years, FID supplier laws were associated with significant decreases in FARS ratios after states adopted these laws (−1.0%, P = .030). Conclusions: The 24 states that have adopted FID supplier laws are saving an estimated 14 lives per year in the United States. An additional 16 lives could be saved if the remaining states adopted this law. FID supplier laws prohibit the production of a FID or transfer of an ID or FID to another person. The more stringent the law (i.e., whether a state prohibits only one element [weaker law] compared to both transferring and manufacturing a FID [stronger]) the more effective a deterrent it becomes to supplying a minor with a FID. States without FID supplier laws should consider adopting them. Keywords: minimum legal drinking age 21 (MLDA-21), fake identification (FID), social host civil liability (SHCL), social host prohibition (SHP), underage drinking drivers, fatal crashes
Introduction Underage Drinking Problem In 1984, federal legislation in the United States was adopted that gave states a strong incentive to either keep or raise their drinking age to 21 years (or lose a substantial amount of federal highway construction funding). This legislation was adopted by the U.S. Congress as a result of recommendations by a President’s Commission on Drunk Driving and Mothers Against Drunk Driving. There was strong evidence that youth in states with a drinking age of 21 were driving to a neighboring state where the drinking age was 18 (or 19) in order to drink and then getting involved in alcohol-related crashes on their way back home. Consequently, by 1988, the minimum legal age to purchase and/or possess alcohol beAssociate Editor Kathy Stewart oversaw the review of this article. Address correspondence to James C. Fell, Senior Research Scientist, Pacific Institute for Research and Evaluation, 11720 Beltsville Drive, Suite 900, Calverton, MD 20705. E-mail: [email protected]
came 21 in all states and the District of Columbia. Surveys indicate, however, that almost half of 8th graders and about three fourths of high school seniors in 2007 reported that they had consumed alcohol and more than half reported being drunk within the past year (Johnston et al. 2008). More recently, in 2012, 13% of 8th graders, 27% of 10th graders, and 40% of 12th graders reported drinking alcohol within the last 30 days. In that same survey, 6% of 8th graders, 15% of 10th graders, and 22% of 12th graders reported heavy episodic drinking (5 or more alcoholic drinks) in the past 30 days (Johnston et al. 2013). Almost a third of the youths in this country reported drinking alcohol at age 16 or younger, with two thirds reporting initiating drinking by age 18 (Hingson et al. 2001). Youths aged 12 to 17 who reported drinking alcohol within the past month were more than twice as likely to get involved in school fights than nondrinkers (Office of Applied Studies 1999). Nearly half of all college students reported participating in high-risk (or binge) drinking (5 or more drinks per session) during 2000 (Hingson et al. 2002; Wechsler et al. 2002). Alcohol has been shown to be a contributor in more than 40% of all college student academic
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Reducing Underage Drinking Driver Fatal Crashes problems and 28% of all college dropouts (National Institute on Alcohol Abuse and Alcoholism [NIAAA] 1997). Underage drinking is related to youth crimes, suicides, rapes, assaults, alcohol poisonings, and unintentional injuries (Bonnie and O’Connell 2004). These underage drinking consequences cost society close to $62 billion annually in direct costs (e.g., medical care) and indirect costs (e.g., pain and lost quality of life; Miller et al. 2006). The relative risk of being killed in a single-vehicle crash for young drivers (aged 16–20) at blood alcohol concentrations (BACs) as low as 0.020 to 0.049 g/dL is almost 4 times that of sober drivers of the same age and is 12 times the risk at BACs from 0.05 to 0.079 g/dL (Voas et al. 2012). According to the NHTSA, minimum legal drinking age of 21 (MLDA-21) laws have reduced traffic fatalities involving youth by 13% and have saved 28,765 lives since 1975 when states began raising their drinking age (Arnold 1985; NHTSA 2008; Womble 1989). Even without strict enforcement, MLDA-21 laws still saved an estimated 533 lives in 2011 (NHTSA 2013a). From 1988 to 1995, alcohol-related traffic fatalities for youth aged 15 to 20 declined from 4,187 to 2,212, a 47% decrease. In 2002, however, 23% of 15- to 20-year-old drivers in fatal crashes had been drinking (BAC ≥ 0.01 g/dL), and in 2011, that proportion was 24%. This stagnation has occurred despite the passage by all states of zero tolerance (ZT) laws for underage drivers (no alcohol while driving), which were designed to strengthen MLDA-21 laws. Toomey and her colleagues (1996) provided examples of how effective the MLDA-21 laws have been without apparent enforcement.
65 The National Academy of Sciences and Institute of Medicine published a comprehensive report in 2004 on strategies to reduce underage drinking in the United States (Bonnie et al. 2004). The report responded to a congressional request for a review of a broad range of federal, state, and nongovernmental programs and policies aimed at reducing or preventing underage drinking. The policies and some of the National Academy of Sciences and Institute of Medicine recommendations are embodied in an expanded set of 17 MLDA-21 laws that we are continuing to evaluate for their effectiveness (see Table 1 for a listing of each law, abbreviations, and brief descriptions). Only one state has enacted all 17 laws (i.e., Utah). Comparatively, Kentucky has enacted only 7 of the 17 laws. State legislatures with an interest in controlling the use of alcohol by their citizens tend to have many underage drinking and impaired driving laws. States with other priorities tend to have fewer of these laws. Moreover, many states that have these law components provide for important exceptions to them. There is a need to determine which of the 17 laws relating to under age 21 drinking and under age 21 impaired driving have been effective in the states and to determine whether strengthening these laws and/or increasing resources devoted to their enforcement will increase their overall effectiveness. Previous Study We documented the effective dates for 6 of the 17 underage drinking law components in a previous study: possession, purchase, use and lose, ZT, graduated driver licensing
Table 1. Minimum legal drinking age 21 (MLDA-21) law components and abbreviations MLDA-21 law components CORE LAWS THAT APPLY TO YOUTH: Possession Purchase
EXPANDED LAWS THAT APPLY TO YOUTH: Consumption Use and lose Use of fake identification APPLY TO YOUTH DRIVING: Zero tolerance Graduated driver licensing with night restrictions APPLY TO PROVIDERS Furnishing or selling Age of on-premise sellers/servers Age of off-premise sellers/servers Keg registration Responsible beverage service training Retailer support provisions for fake identification Social host civil liability Social host prohibition APPLY TO MANUFACTURERS OR SUPPLIERS OF FAKE IDENTIFICATION Transfer/production of fake identification
Abbreviation (if any) and comment Illegal for youth under age 21 to possess alcohol Illegal for youth under age 21 to purchase or attempt to purchase alcohol. Possession and purchase were treated together as the laws that met federal standards by 1988 Illegal for youth under age 21 to consume alcohol Alcohol citation for youth under age 21 results in driver’s license suspension FID minor—illegal for youth under age 21 to use a fake identification to purchase alcohol ZT—Illegal for a driver under age 21 to have any alcohol in their system when driving GDL—Youth with intermediate or provisional license prohibited from driving without an adult in the vehicle past a certain hour at night Illegal to furnish or sell alcohol to a youth under age 21 Minimum age set for selling/serving alcohol Minimum age set for selling/serving alcohol Identification number for beer keg and purchaser required RBS—Responsible beverage training mandatory or voluntary FID retailer—Provisions to assist retailers in avoiding sales to youth under age 21 SHCL—Allows suing social hosts for injuries caused by underage drinking guests SHP—Prohibits social hosting of underage 21 drinking parties FID supplier—Prohibits manufacturing and/or supplying fake identification to youth for the purposes of buying alcohol
APPLY TO STATES CONCERNING CONTROL OF ALCOHOL DISTRIBUTION State control of alcohol sales State control—A state-run retail distribution system of one or more of the alcohol beverage types
Downloaded by [Umeå University Library] at 12:28 01 April 2015
66 (GDL) with night restrictions, and keg registration statutes and determined their effectiveness (Fell et al. 2009). For those aged 20 and younger, the results of that study indicated that 4 of the underage drinking and underage impaired-driving laws were associated with significant decreases in the ratio of drinking to nondrinking drivers in fatal crashes using the Fatality Analysis Reporting System (FARS) data set for the years 1982 through 2004. Specifically, significant decreases in the ratio of underage drinking drivers to underage nondrinking drivers in fatal crashes, the crash incidence ratio (CIR), were associated with the implementation of possession and purchase laws treated together as the initial core MLDA-21 laws (−16%, P < .001), the ZT law (−5%, P = .015), and the use and lose law (−5%, P = .026). The GDL law with night restrictions did not emerge as significantly related to the underage fatal CIR. Contrary to expectations, keg registration was associated with a significant increase in the fatal CIR (12%, P < .001). Three of the laws targeting the overall driving public were also associated with reductions in the drinking-driver fatal CIR for youth: a secondary enforcement seat belt usage law or adding a primary enforcement law (−3%, P = .041), the administrative license revocation (ALR) law (−5%, P = .024), and the 0.08 BAC per se limit law (−8%, P = .002). There was a marginally significant downward trend associated with the 0.10 BAC per se law for the underage group (−7%, P = .065). The illegal BAC limit for drivers aged 21 and older has changed in the United States from 0.15 g/dL to 0.10 g/dL and currently to 0.08 g/dL. Current Study Social Host Laws Prohibitions on hosting underage parties laws involve statutory prohibitions enforced by state or local governments, generally through criminal proceedings, that can lead to sanctions such as fines or imprisonment. In contrast, social host civil liability laws involve actions by a private party seeking monetary damages for injuries that result from permitting underage drinking on the host’s premises. Wagenaar et al. (2001) conducted a survey of public attitudes toward social host liability and found that 93% agreed that liability of the host was at least equal to the liability of the underage drinking driver causing the injury (Wagenaar et al. 2001). Wagoner and colleagues (2013) found that communities with social host policies concerning underage drinking parties had adolescents with a lower odds of drinking in large peer groups compared to adolescents in communities without such policies. Dills (2010) analyzed national survey data and FARS data among 18- to 20-year-olds. The study found that social host civil liability laws were associated with a 3% decrease in reported heavy drinking, a 4% decrease in reported drinking and driving, and a 5%–9% decrease in alcohol-related fatal crashes. Fake Identification Laws In our earlier research, using a different methodology and approach, we found that from state to state, a unit increase in the strength of the fake identification use law (fake identification [FID] minor) was associated with a 7.3% smaller outcome measure (P = .034; Fell et al. 2008). The difference between
Fell et al. the weakest (no license sanction) and the strongest (administrative license sanction) FID minor law represented a 14.1% decrease in the under age 21 driver FARS ratio. Given the apparent importance of these social host and fake ID laws, we sought to determine the effectiveness of each. Our pre- and postlaw design used data covering 29 years from 1982 through 2010 (6 additional years than the prior study).
Methods We followed the NIAAA’s APIS General Protocol titled Conducting Legal Research on Activity 6 Policies, General Protocols (NIAAA working manuscript) methodology and used the online legal research tools Westlaw and HeinOnline in order to conduct original legal research to document the effective dates of 11 additional MLDA-21 law components that have been enacted in various states. Westlaw data were fundamental tools for this effort in providing the effective dates. However, it was not always possible to obtain this information for statutes in each state over time. Therefore, our legal researchers relied on other published sources, such as HeinOnline, as well information obtained directly from the states, as necessary, to complete the data set. Our research design utilized a 29-year period in the FARS data set from 1982 through 2010. We selected 5 of the 11 additional underage drinking laws for analysis. We selected 3 FID laws, (1) FID for minors, (2) FID for suppliers, (3) FID for retailer support, as well as 2 social host laws (4) social host prohibition (SHP) and (5) social host civil liability (SHCL). The FID laws (1) for minors prohibited the use of false identification cards by minors (i.e., minors under age 21); (2) for suppliers prohibited a person from lending, transferring to, or producing false identification for an underage person (i.e., suppliers); and (3) for retailer support established policies that allowed one or more of the following for retailers of alcohol outlets: the use of identification scanners, distinctive licenses for those under age 21, seizure of identification cards deemed false, or the right to sue an underage person that uses false identification to purchase alcohol (i.e., retailer support). For example, Arkansas adopted an FID for minors law (Ark. Code Ann. s. 5-27-504) and an FID for suppliers law (Ark. Code Ann. s. 5-27-502) in 1991. Later, Arkansas adopted an FID retailer support law (Ark. Code Ann. s. 27-16-801) in 1997. A total of 50 states and D.C. have FID for minors laws, 24 states have FID for suppliers laws, and 45 states have FID retailer support laws. The SHCL laws involve the civil liability faced by noncommercial alcohol servers for injuries or damages caused by their intoxicated or underage drinking guests. In states with SHCL laws, injured third parties are able to sue social hosts (as well as the minor who caused the crash) for monetary damages. SHCL is distinguished from underagefurnishing and host party policies (which we refer to here as social host prohibitions [SHP]), which can result in penalties imposed by the state (e.g., fines, jail, etc.). SHCL law is established by statute or by a state court through common law. We collected effective dates on both statutory and common law (case law) for the SHCL policy topic and, in jurisdictions with both statutes and common law (rather than one or the other), we selected the earliest effective date for our analyses. For example, in Arizona, the state court issued a decision in
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Reducing Underage Drinking Driver Fatal Crashes 1996 (Knoell v. Cerkvenik-Anderson Travel, Inc., 185 Ariz. 546, 917 P.2d 689) allowing a third party to sue a social host for injuries suffered as a result of allowing an underage person to consume alcohol on his or her property. A total of 28 states have social host prohibition laws and 33 states have social host civil liability laws. We then used the 3 laws with very strong evidence for regulating drinking and driving including the adoption of the 0.10 BAC limit for driving, lowering the BAC limit for driving to 0.08, and ALR laws. These 3 laws have been shown to be effective in reducing impaired-driving fatal crashes (Klein 1989; Shults et al. 2001; Voas et al. 2000; Wagenaar and Maldonado-Molina 2007; Wagenaar et al. 2007). We also selected laws designed to promote general driving safety, which would likely impact traffic fatalities including sobriety checkpoints designed to cut down on impaired drivers (Fell et al. 2003; Lacey et al. 1999; Peek-Asa 1999) and mandatory seat belt laws (primary and secondary enforcement; Voas et al. 2007). For the current study, we partially replicated a design utilized in previous research by Fell et al. (2009), which found that additional variables assessing the economic strength of the state may contribute to the ratio of drinking drivers to nondrinking drivers involved in fatal crashes. These variables included per capita beer consumption and indications of economic strength in the state including employment rates and vehicle-miles traveled (VMT). Sources of Data Traffic Fatalities Annual traffic fatality data for each state was derived from the NHTSA’s FARS database from 1982 to 2010 (NHTSA 2013b). The FARS database is a census of motor vehicle traffic crashes that resulted in the death of an individual within the 30 days following the crash. To be included in the FARS database, the crash must have (1) occurred on U.S. public roadways and (2) been attended to and reported by police. The involvement of alcohol in the crash is derived from positive BAC data from tests of drivers involved in the crashes. When BACs for drivers are missing, the BACs are imputed to obtain a more accurate picture of the involvement of alcohol in the crash (Subramanian 2002). Thus, since 1982 every driver in FARS has a BAC from testing by the police or medical examiners or an imputed BAC based upon other crash data (e.g., time of crash, police assessment that driver had been drinking, age and gender of the driver, number of vehicles in the crash, etc.). This information was available in FARS for all years examined in the current study (1982 through 2010) and deemed appropriate for these analyses because many of the MLDA-21 laws were adopted in this 29-year period. It is feasible that any number of variables could have a potential impact on rates of drinking-driver-related crashes (i.e., road conditions, weather, non-alcohol-related policies, etc.). As such, it would be ideal for any study exploring factors related to drinking-driver-related crashes to account for and control for each of these variables. Even if it were possible to know all of the potential factors that may contribute to drinking-driver-related crashes, obtaining accurate operational measures for each would be extremely difficult if not impossible. However, because many of these unknown factors
67 would likely also impact non-drinking-driver-related crashes, the utilization of a control “non-drinking driver” condition should wash out many of these potential effects. For the purposes of this study, the outcome variables were computed from FARS data for each year by state. Subjects were characterized by age (younger than age 21 or aged 26 years and older) and then the number of drinking-driver-related crashes was divided by the number of non-drinking-driver-related crashes. This gave us the ratio of positive to negative BACs for those younger than age 21 and for those aged 26 years and older (drivers with BACs > 0.01 as a ratio to drivers with BACs = 0.00). As was the case in Fell et al. (2009), we also avoided using data on drivers between the ages of 21 and 25. By doing so, we were able to examine 2 groups of drivers without concern of potential carryover effects. False Identification and Social Host Prohibitions and Social Host Civil Liability Laws The current study utilized similar methods to Fell et al. (2009) in measurement of the presence of laws. That is, years in which the law was effective from January through December were coded as 1 and years in which the law was not present at any time throughout the year were coded as 0. The years in which laws became effective were coded as a proportion to reflect the time of year in which a law was in force. For example, if a law was in effect in July where it would be present for only half of that year, it was coded as 0.50; if a law was in effect in April where it would be present three quarters of the year it was coded as 0.75, and so on. The complete years leading up to the effective dates of a law were coded as 0 and the complete years following the effective date of a law were coded as 1. Finally, data were gathered to determine and code the strength of each of these laws. The strength coding for the 3 FID laws (i.e., minor, retailer support, and supplier) and SHP laws was accomplished in our earlier research (Fell et al. 2008). However, the strength coding for SHCL was not available in our earlier research and subsequently was computed in a similar fashion for the current study. FID minor laws were given a strength based on whether the states had administrative and/or judicial sanctions in this law. Because all states prohibit minors from using FIDs, a score ranging from 1 (which would indicate that no license sanctioning procedure is in place) to 3 (indicating a law with administrative or both administrative and judicial sanction procedures in place) was used. FID retailer support laws were given a strength based on the presence of incentives for retailers to use scanners, having distinctive drivers licenses issued for easier identification of those under age 21, seizure of suspicious ID by retailers, and the right to sue a minor. These laws lost points if the state law allowed for a general affirmative defense in court. An affirmative defense example would be that the furnisher inspected the youth’s identification and came to a reasonable conclusion based on the appearance that it was valid. Another example would be that the furnisher came to a good faith or reasonable decision that the alcohol receiver was aged 21 or older. Subsequently, scores for FID retailer support may range from 0, indicating no retailer support provisions, to 5, indicating all provisions except a general affirmative defense in court. FID supplier laws were given a strength based on whether the (a) lending, transfer, or sale of FID and the (b) manufacturing
68
Fell et al.
Table 2. Point values in determining strength of fake identification laws Scoring criteria
Weight point values
Fake identification—minor Fake identification law License sanction procedure
+1.0 point for presence of the law +2.0 points for administrative sanctions OR +2.0 points for administrative and judicial sanctions OR +1.0 points for judicial sanctions only Fake identification—supplier +1.0 point if yes +1.0 point if yes
Lending, transfer, or sale of false IDs criminalized Manufacturing and distribution of false IDs criminalized
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Fake identification—retailer support +1.0 point if yes +2.0 points if yes +1.0 point if yes +1.0 point if yes −1.0 point for general affirmative defense 0.0 points for specific affirmative defense or none
Incentives for retailers to use scanners Use of distinctive licenses Seizure of suspicious ID by retailer permitted Right to sue minor Affirmative defense
Source: Fell et al. (2008).
and distribution of FID is criminalized. Subsequently, scores for FID supplier laws may range from 0, indicating no law against providing FIDs, to 2, indicating that both the lending and manufacturing of FIDs is criminalized (see Table 2). The SHP laws were given a strength score based on the statutes addressing underage actions, property types, and knowledge levels. Specifically, states were given a strength score using the following point values described in Table 3. As such, states may have had a SHP strength score ranging from 0 (indicating no SHP law) to 10 (indicating a general statute covering all underage actions, all property types, with negligence as the knowledge standard and no exceptions to the law). In a court case against a social host, the prosecution must present sufficient evidence to meet a knowledge standard to establish a violation. The more demanding the standard in terms of the evidence required, the lower the likelihood of a successful prosecution. The following categories are ranked by the level of evidence required, from most demanding to least demanding: • Overt act—The host must have actual knowledge of specific aspects of the party and must commit an act that contributes to its occurrence. • Knowledge—The host must have actual knowledge of specific aspects of the party; no action is required. • Recklessness—The host may not have acted with actual knowledge of the party but must act with intentional disregard for the probable consequences of her or his actions. • Negligence—The host knew or should have known of the event’s occurrence (in legal terminology this is referred to as “constructive knowledge”). The SHCL laws were scored in a similar fashion by examining established statutes and/or by state court through common law. A designation of social host civil liability signifies that liability may be imposed for the negligent provision of alcohol to a minor. The law is limited by narrowing the scope of who may be sued (i.e., such as only individuals over the age of 21 as is evident in some states) and by the necessity of proving that the social host knew that the guest was underage and/or a need to provide evidence beyond a reasonable
doubt. With a base score of 3.0 points allotted for having a SHCL law, scores can range from 0, indicating no law, to 3, indicating either or both common and statutory laws with no limitations. Driving Safety and Drinking and Driving Laws Driving safety laws were composed of the strength of sobriety checkpoints and a composite score of primary and secondary seat belt usage laws. Data concerning the strength of sobriety checkpoints was derived from previous research by Fell et al. (2003), who interviewed state representatives in 2000 to determine the number of checkpoints conducted in the previous year. Similar to Fell et al. (2009), we used reported frequency Table 3. Point values in determining strength of social host laws Scoring criteria Social host prohibitions Type of statute
Underage guest actions triggering violation
Property type covered by law
Knowledge standard
Preventive actions available to offender Exceptions to law
Social host civil liability Common law or common law and statutory law Statutory law only Limitations to the law
Weight point values +2.0 points for general OR +1.0 point for specific +1.0 point for possession +1.0 point for consumption +1.0 point for intention to possess or consume +1.0 point for residence +1.0 point for outdoor area +1.0 point for other areas +2.0 points for negligence +1.0 point for knowledge 0 points for overt act −1.0 if yes −1.0 point for family −1.0 point for resident of household −1.0 point for other +3.0 points +3.0 points −1.0 point for limitation on who may be sued −1.0 point for limitation on elements or standards of proof
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Reducing Underage Drinking Driver Fatal Crashes of sobriety checkpoints and coded them as 0 if checkpoints are illegal or otherwise not conducted, 1 if they were conducted infrequently, and 2 if they were frequently conducted such as on a monthly or weekly basis. Similarly, seat belt use laws were determined by whether states had laws dictating mandatory seat belt use. If states did not have this law (i.e., New Hampshire), they were coded as 0, if they had secondary seat belt laws they were coded as 1, and if they had secondary seat belt laws upgraded to primary seat belt laws they received a code of 2. As such, in both sobriety checkpoints and seat belt laws, higher scores indicated stronger laws. Drinking and driving laws were primarily composed of the establishment of 0.08 and 0.10 illegal limit laws and ALR laws, all of which were coded as 0 if the law was absent and 1 if the law was present. As was done with the previous laws, if seat belt laws, illegal BAC limit laws, or ALR laws were changed in the middle of the year, scores were altered to reflect the proportion of the year the law was in effect. Beer Consumption Per capita beer consumption rates were obtained for individuals aged 15 years and older by year and state. These data were obtained from the annual publication of the NIAAA’s Alcohol Epidemiologic Data System. Of note, beer consumption rates were only available as general numbers by state and year. Ideally, it would have been preferable to have these numbers available by age range to allow for a more accurate representation of the role of beer consumption on FARS statistics for those younger than aged 21 years and those aged 26 years and older. However, because these data were not available, per capita beer consumption by state and year was deemed sufficient for our analyses (Fell et al. 2009). Economic Strength Similar to Fell et al. (2009), economic strength was determined by state unemployment rates and VMT. VMT was derived from the Federal Highway Administration data. The Federal Highway Administration produces an annual estimate of total VMT by state and year. Unemployment statistics were derived from the Bureau of Labor Statistics (BLS), which publishes monthly employment statistics by state (Fell et al. 2009). The method of data collection utilized by the BLS underwent 2 notable changes during the years examined in the current study. In 2007, the BLS data included a seasonal adjustment that is designed to eliminate the influences of holidays, weather, and other recurring seasonal events. In 2010, the data also included a population control designed to account for vital statistics on births, deaths, migration, and other population-related variables. The Hypothesized Model The hypothesized model used in this study is illustrated in the path diagram (Figure 1). The model is composed of (1) 3 FID laws (i.e., minor, retailer support, and supplier); (2) 2 versions of the social host laws (i.e., SHP and SHCL laws); (3) 3 alcoholimpaired driving laws (0.08 legal limit, 0.10 legal limit, and ALR); (4) 2 driving safety laws (use of sobriety checkpoints
69 0.08 law 0.10 law
Alcohol Impaired Driving Laws
ALR law < 21 FARS Ratio (drinking to nondrinking drivers)
FID - minor FID - supplier
Beer consumption
FID - retailer
SHP SHCL
Checkpoints
Driving Safety Laws
> 26 FARS Ratio (drinking to nondrinking drivers)
SB laws
Employment
Economic Strength
VMT
Fig. 1. Full structural equation model (SEM) for underage drinking laws.
and seat belt use laws); (5) 2 variables determining economic strength (employment rate and VMT by state and year); (6) per capita beer consumption by state and year; (7) ratio of under age 21 drinking drivers to underage 21 nondrinking drivers in fatal crashes; and (8) ratio of drinking drivers to nondrinking drivers in fatal crashes among drivers aged 26 years and older as a comparison. The model assumes that all 3 FID laws and both types of social host laws directly affect both the younger than aged 21 years FARS ratio and, potentially, beer consumption. Similarly, we assumed that both the latent variables measuring drinking and driving laws and economic strength directly affect beer consumption and both younger than aged 21 years and aged 26 years and older FARS ratios, but latent variables predicting driving safety laws only directly affect both FARS ratios. We also assumed no covariance between latent variables. Further, we have evidence that beer consumption also directly affects both FARS ratios. Because beer consumption has previously been identified as highly correlated to the FARS alcohol ratios, its inclusion in the model was deemed appropriate (see Voas et al. 2007). It was also deemed that FID laws could have a potentially important impact on beer consumption, which, in turn, would also impact FARS ratios. Therefore, in this model, beer consumption is both a predictor and an intermediate outcome, and FARS ratios among individuals younger than aged 21 years and older than aged 26 years are both outcome variables. Furthermore, though we would expect that the use of a control condition could wash out many of the unmeasured variables that may impact crash rates, some residual data will likely still be evident. As such, it was expected that the residual data from both the younger than aged 21 and aged 26 and older FARS ratios would likely be covariates (NHTSA 2013b). We ran the model incorporating the strength of all 5 laws: FID minor, FID retailer, FID supplier, SHCL, and SHP. Strengths of these laws were adapted from Fell et al. (2008) where possible and as described
70
Fell et al.
earlier; otherwise, they were coded in a similar fashion when data were not available in previous research as was the case with the SHCL laws.
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Data Analysis Structural Equation Modeling The data were analyzed using structural equation modeling (SEM) techniques with Analysis of Moment-Based Structures (AMOS v.21), an SPSS-based package (IBM SPSS Inc., Chicago, IL). SEM is a statistical technique frequently utilized to estimate causal relationships based on qualitative assumptions represented in a path diagram. SEM allows for confirmatory and exploratory modeling of both observed variables (rectangles in Figure 1) or latent variables (ovals in Figure 1) derived from combinations of other observed vari¨ ables (Joreskog 1966, 1969, 1967). The use of SEM has gained notable popularity among researchers both for its utility in exploring relationships beyond what is possible with simple analyses of variance or multiple regression analyses and for its applicability to a variety of functions. SEM generally requires a large number of data points (>200; Kline 2005). Because the current study had 1,479 data points (51 states × 29 years),
SEM was deemed appropriate. Further, SEM was used in the current analysis to more accurately account for simultaneous effects of laws on multiple outcomes (i.e., FARS ratios and beer consumption) and because beer consumption was modeled as an intervening variable rather than only a predictor of FARS ratios.
Results The model incorporating the strengths of the laws indicated that one of the FID laws had a significant negative impact on the FARS ratios of drivers who were drinking to drivers who were not drinking. Specifically, for those younger than aged 21 years, only the FID supplier law was associated with significant decreases in FARS ratios (−1.0, P = .030; see Table 4 and Figure 2). However, FID laws dealing with minors and suppliers and SHP laws were found to have a negative but nonsignificant impact on FARS ratios for drivers younger than aged 21 (Table 4 and Figure 2). Beer consumption, which we also hypothesized to be impacted by FID and both social host laws, however, displayed different patterns. SHP laws—not SHCL laws—were found to have a significant impact on annual per
Table 4. Regression weights and significance level for direct effects including strength of fake identification and social host MLDA-21 lawsa Model including strength of FID and SHP laws Predictor
Estimate
SE
P value
0.019 0.007 0.003 0.006 0.005 0.003 0.001 0.003 0.022
Traffic Injury Prevention Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/gcpi20
Effectiveness of Social Host and Fake Identification Laws on Reducing Underage Drinking Driver Fatal Crashes a
a
a
James C. Fell , Michael Scherer , Sue Thomas & Robert B. Voas
a
a
Pacific Institute for Research and Evaluation, Calverton, Maryland Published online: 11 Oct 2014.
Click for updates To cite this article: James C. Fell, Michael Scherer, Sue Thomas & Robert B. Voas (2014) Effectiveness of Social Host and Fake Identification Laws on Reducing Underage Drinking Driver Fatal Crashes, Traffic Injury Prevention, 15:sup1, S64-S73, DOI: 10.1080/15389588.2014.928929 To link to this article: http://dx.doi.org/10.1080/15389588.2014.928929
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Traffic Injury Prevention (2014) 15, S64–S73 C Taylor & Francis Group, LLC Copyright ISSN: 1538-9588 print / 1538-957X online DOI: 10.1080/15389588.2014.928929
Effectiveness of Social Host and Fake Identification Laws on Reducing Underage Drinking Driver Fatal Crashes JAMES C. FELL, MICHAEL SCHERER, SUE THOMAS, and ROBERT B. VOAS Pacific Institute for Research and Evaluation, Calverton, Maryland
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Received 17 March 2014, Accepted 24 May 2014
Objective: The public generally assumes that the minimum legal drinking age of 21 (MLDA-21) legislation in the United States is embodied in a single law and therefore all states have the same law. Actually, the MLDA-21 state laws consist of multiple provisions that support the core MLDA-21 laws and include a family of policies directed at controlling underage drinking and underage drinking and driving. Because social host and fake identification laws have recently garnered interest by policy makers in the states, this study was designed to determine their effectiveness. Methods: The effective dates for 2 types of social host laws and 3 fake identification laws were documented using the Alcohol Policy Information System (APIS), the Substance Abuse and Mental Health Services Administration’s (SAMHSA) 2011 Report to Congress on the Prevention and Reduction of Underage Drinking, and legal research tools. These laws include social host prohibitions (SHPs) and social host civil liability (SHCL), the use of fake identification (FID), retailer support for FID, and transfer/production of FID. We used a pre–post design to evaluate the influence on underage drinking-and-driving fatal crashes of these 5 laws using the Fatality Analysis Reporting System (FARS) data set for the years 1982 through 2010. The data were analyzed using structural equation modeling (SEM) controlling for as many variables as possible. Results: For those drivers younger than aged 21 years, FID supplier laws were associated with significant decreases in FARS ratios after states adopted these laws (−1.0%, P = .030). Conclusions: The 24 states that have adopted FID supplier laws are saving an estimated 14 lives per year in the United States. An additional 16 lives could be saved if the remaining states adopted this law. FID supplier laws prohibit the production of a FID or transfer of an ID or FID to another person. The more stringent the law (i.e., whether a state prohibits only one element [weaker law] compared to both transferring and manufacturing a FID [stronger]) the more effective a deterrent it becomes to supplying a minor with a FID. States without FID supplier laws should consider adopting them. Keywords: minimum legal drinking age 21 (MLDA-21), fake identification (FID), social host civil liability (SHCL), social host prohibition (SHP), underage drinking drivers, fatal crashes
Introduction Underage Drinking Problem In 1984, federal legislation in the United States was adopted that gave states a strong incentive to either keep or raise their drinking age to 21 years (or lose a substantial amount of federal highway construction funding). This legislation was adopted by the U.S. Congress as a result of recommendations by a President’s Commission on Drunk Driving and Mothers Against Drunk Driving. There was strong evidence that youth in states with a drinking age of 21 were driving to a neighboring state where the drinking age was 18 (or 19) in order to drink and then getting involved in alcohol-related crashes on their way back home. Consequently, by 1988, the minimum legal age to purchase and/or possess alcohol beAssociate Editor Kathy Stewart oversaw the review of this article. Address correspondence to James C. Fell, Senior Research Scientist, Pacific Institute for Research and Evaluation, 11720 Beltsville Drive, Suite 900, Calverton, MD 20705. E-mail: [email protected]
came 21 in all states and the District of Columbia. Surveys indicate, however, that almost half of 8th graders and about three fourths of high school seniors in 2007 reported that they had consumed alcohol and more than half reported being drunk within the past year (Johnston et al. 2008). More recently, in 2012, 13% of 8th graders, 27% of 10th graders, and 40% of 12th graders reported drinking alcohol within the last 30 days. In that same survey, 6% of 8th graders, 15% of 10th graders, and 22% of 12th graders reported heavy episodic drinking (5 or more alcoholic drinks) in the past 30 days (Johnston et al. 2013). Almost a third of the youths in this country reported drinking alcohol at age 16 or younger, with two thirds reporting initiating drinking by age 18 (Hingson et al. 2001). Youths aged 12 to 17 who reported drinking alcohol within the past month were more than twice as likely to get involved in school fights than nondrinkers (Office of Applied Studies 1999). Nearly half of all college students reported participating in high-risk (or binge) drinking (5 or more drinks per session) during 2000 (Hingson et al. 2002; Wechsler et al. 2002). Alcohol has been shown to be a contributor in more than 40% of all college student academic
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Reducing Underage Drinking Driver Fatal Crashes problems and 28% of all college dropouts (National Institute on Alcohol Abuse and Alcoholism [NIAAA] 1997). Underage drinking is related to youth crimes, suicides, rapes, assaults, alcohol poisonings, and unintentional injuries (Bonnie and O’Connell 2004). These underage drinking consequences cost society close to $62 billion annually in direct costs (e.g., medical care) and indirect costs (e.g., pain and lost quality of life; Miller et al. 2006). The relative risk of being killed in a single-vehicle crash for young drivers (aged 16–20) at blood alcohol concentrations (BACs) as low as 0.020 to 0.049 g/dL is almost 4 times that of sober drivers of the same age and is 12 times the risk at BACs from 0.05 to 0.079 g/dL (Voas et al. 2012). According to the NHTSA, minimum legal drinking age of 21 (MLDA-21) laws have reduced traffic fatalities involving youth by 13% and have saved 28,765 lives since 1975 when states began raising their drinking age (Arnold 1985; NHTSA 2008; Womble 1989). Even without strict enforcement, MLDA-21 laws still saved an estimated 533 lives in 2011 (NHTSA 2013a). From 1988 to 1995, alcohol-related traffic fatalities for youth aged 15 to 20 declined from 4,187 to 2,212, a 47% decrease. In 2002, however, 23% of 15- to 20-year-old drivers in fatal crashes had been drinking (BAC ≥ 0.01 g/dL), and in 2011, that proportion was 24%. This stagnation has occurred despite the passage by all states of zero tolerance (ZT) laws for underage drivers (no alcohol while driving), which were designed to strengthen MLDA-21 laws. Toomey and her colleagues (1996) provided examples of how effective the MLDA-21 laws have been without apparent enforcement.
65 The National Academy of Sciences and Institute of Medicine published a comprehensive report in 2004 on strategies to reduce underage drinking in the United States (Bonnie et al. 2004). The report responded to a congressional request for a review of a broad range of federal, state, and nongovernmental programs and policies aimed at reducing or preventing underage drinking. The policies and some of the National Academy of Sciences and Institute of Medicine recommendations are embodied in an expanded set of 17 MLDA-21 laws that we are continuing to evaluate for their effectiveness (see Table 1 for a listing of each law, abbreviations, and brief descriptions). Only one state has enacted all 17 laws (i.e., Utah). Comparatively, Kentucky has enacted only 7 of the 17 laws. State legislatures with an interest in controlling the use of alcohol by their citizens tend to have many underage drinking and impaired driving laws. States with other priorities tend to have fewer of these laws. Moreover, many states that have these law components provide for important exceptions to them. There is a need to determine which of the 17 laws relating to under age 21 drinking and under age 21 impaired driving have been effective in the states and to determine whether strengthening these laws and/or increasing resources devoted to their enforcement will increase their overall effectiveness. Previous Study We documented the effective dates for 6 of the 17 underage drinking law components in a previous study: possession, purchase, use and lose, ZT, graduated driver licensing
Table 1. Minimum legal drinking age 21 (MLDA-21) law components and abbreviations MLDA-21 law components CORE LAWS THAT APPLY TO YOUTH: Possession Purchase
EXPANDED LAWS THAT APPLY TO YOUTH: Consumption Use and lose Use of fake identification APPLY TO YOUTH DRIVING: Zero tolerance Graduated driver licensing with night restrictions APPLY TO PROVIDERS Furnishing or selling Age of on-premise sellers/servers Age of off-premise sellers/servers Keg registration Responsible beverage service training Retailer support provisions for fake identification Social host civil liability Social host prohibition APPLY TO MANUFACTURERS OR SUPPLIERS OF FAKE IDENTIFICATION Transfer/production of fake identification
Abbreviation (if any) and comment Illegal for youth under age 21 to possess alcohol Illegal for youth under age 21 to purchase or attempt to purchase alcohol. Possession and purchase were treated together as the laws that met federal standards by 1988 Illegal for youth under age 21 to consume alcohol Alcohol citation for youth under age 21 results in driver’s license suspension FID minor—illegal for youth under age 21 to use a fake identification to purchase alcohol ZT—Illegal for a driver under age 21 to have any alcohol in their system when driving GDL—Youth with intermediate or provisional license prohibited from driving without an adult in the vehicle past a certain hour at night Illegal to furnish or sell alcohol to a youth under age 21 Minimum age set for selling/serving alcohol Minimum age set for selling/serving alcohol Identification number for beer keg and purchaser required RBS—Responsible beverage training mandatory or voluntary FID retailer—Provisions to assist retailers in avoiding sales to youth under age 21 SHCL—Allows suing social hosts for injuries caused by underage drinking guests SHP—Prohibits social hosting of underage 21 drinking parties FID supplier—Prohibits manufacturing and/or supplying fake identification to youth for the purposes of buying alcohol
APPLY TO STATES CONCERNING CONTROL OF ALCOHOL DISTRIBUTION State control of alcohol sales State control—A state-run retail distribution system of one or more of the alcohol beverage types
Downloaded by [Umeå University Library] at 12:28 01 April 2015
66 (GDL) with night restrictions, and keg registration statutes and determined their effectiveness (Fell et al. 2009). For those aged 20 and younger, the results of that study indicated that 4 of the underage drinking and underage impaired-driving laws were associated with significant decreases in the ratio of drinking to nondrinking drivers in fatal crashes using the Fatality Analysis Reporting System (FARS) data set for the years 1982 through 2004. Specifically, significant decreases in the ratio of underage drinking drivers to underage nondrinking drivers in fatal crashes, the crash incidence ratio (CIR), were associated with the implementation of possession and purchase laws treated together as the initial core MLDA-21 laws (−16%, P < .001), the ZT law (−5%, P = .015), and the use and lose law (−5%, P = .026). The GDL law with night restrictions did not emerge as significantly related to the underage fatal CIR. Contrary to expectations, keg registration was associated with a significant increase in the fatal CIR (12%, P < .001). Three of the laws targeting the overall driving public were also associated with reductions in the drinking-driver fatal CIR for youth: a secondary enforcement seat belt usage law or adding a primary enforcement law (−3%, P = .041), the administrative license revocation (ALR) law (−5%, P = .024), and the 0.08 BAC per se limit law (−8%, P = .002). There was a marginally significant downward trend associated with the 0.10 BAC per se law for the underage group (−7%, P = .065). The illegal BAC limit for drivers aged 21 and older has changed in the United States from 0.15 g/dL to 0.10 g/dL and currently to 0.08 g/dL. Current Study Social Host Laws Prohibitions on hosting underage parties laws involve statutory prohibitions enforced by state or local governments, generally through criminal proceedings, that can lead to sanctions such as fines or imprisonment. In contrast, social host civil liability laws involve actions by a private party seeking monetary damages for injuries that result from permitting underage drinking on the host’s premises. Wagenaar et al. (2001) conducted a survey of public attitudes toward social host liability and found that 93% agreed that liability of the host was at least equal to the liability of the underage drinking driver causing the injury (Wagenaar et al. 2001). Wagoner and colleagues (2013) found that communities with social host policies concerning underage drinking parties had adolescents with a lower odds of drinking in large peer groups compared to adolescents in communities without such policies. Dills (2010) analyzed national survey data and FARS data among 18- to 20-year-olds. The study found that social host civil liability laws were associated with a 3% decrease in reported heavy drinking, a 4% decrease in reported drinking and driving, and a 5%–9% decrease in alcohol-related fatal crashes. Fake Identification Laws In our earlier research, using a different methodology and approach, we found that from state to state, a unit increase in the strength of the fake identification use law (fake identification [FID] minor) was associated with a 7.3% smaller outcome measure (P = .034; Fell et al. 2008). The difference between
Fell et al. the weakest (no license sanction) and the strongest (administrative license sanction) FID minor law represented a 14.1% decrease in the under age 21 driver FARS ratio. Given the apparent importance of these social host and fake ID laws, we sought to determine the effectiveness of each. Our pre- and postlaw design used data covering 29 years from 1982 through 2010 (6 additional years than the prior study).
Methods We followed the NIAAA’s APIS General Protocol titled Conducting Legal Research on Activity 6 Policies, General Protocols (NIAAA working manuscript) methodology and used the online legal research tools Westlaw and HeinOnline in order to conduct original legal research to document the effective dates of 11 additional MLDA-21 law components that have been enacted in various states. Westlaw data were fundamental tools for this effort in providing the effective dates. However, it was not always possible to obtain this information for statutes in each state over time. Therefore, our legal researchers relied on other published sources, such as HeinOnline, as well information obtained directly from the states, as necessary, to complete the data set. Our research design utilized a 29-year period in the FARS data set from 1982 through 2010. We selected 5 of the 11 additional underage drinking laws for analysis. We selected 3 FID laws, (1) FID for minors, (2) FID for suppliers, (3) FID for retailer support, as well as 2 social host laws (4) social host prohibition (SHP) and (5) social host civil liability (SHCL). The FID laws (1) for minors prohibited the use of false identification cards by minors (i.e., minors under age 21); (2) for suppliers prohibited a person from lending, transferring to, or producing false identification for an underage person (i.e., suppliers); and (3) for retailer support established policies that allowed one or more of the following for retailers of alcohol outlets: the use of identification scanners, distinctive licenses for those under age 21, seizure of identification cards deemed false, or the right to sue an underage person that uses false identification to purchase alcohol (i.e., retailer support). For example, Arkansas adopted an FID for minors law (Ark. Code Ann. s. 5-27-504) and an FID for suppliers law (Ark. Code Ann. s. 5-27-502) in 1991. Later, Arkansas adopted an FID retailer support law (Ark. Code Ann. s. 27-16-801) in 1997. A total of 50 states and D.C. have FID for minors laws, 24 states have FID for suppliers laws, and 45 states have FID retailer support laws. The SHCL laws involve the civil liability faced by noncommercial alcohol servers for injuries or damages caused by their intoxicated or underage drinking guests. In states with SHCL laws, injured third parties are able to sue social hosts (as well as the minor who caused the crash) for monetary damages. SHCL is distinguished from underagefurnishing and host party policies (which we refer to here as social host prohibitions [SHP]), which can result in penalties imposed by the state (e.g., fines, jail, etc.). SHCL law is established by statute or by a state court through common law. We collected effective dates on both statutory and common law (case law) for the SHCL policy topic and, in jurisdictions with both statutes and common law (rather than one or the other), we selected the earliest effective date for our analyses. For example, in Arizona, the state court issued a decision in
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Reducing Underage Drinking Driver Fatal Crashes 1996 (Knoell v. Cerkvenik-Anderson Travel, Inc., 185 Ariz. 546, 917 P.2d 689) allowing a third party to sue a social host for injuries suffered as a result of allowing an underage person to consume alcohol on his or her property. A total of 28 states have social host prohibition laws and 33 states have social host civil liability laws. We then used the 3 laws with very strong evidence for regulating drinking and driving including the adoption of the 0.10 BAC limit for driving, lowering the BAC limit for driving to 0.08, and ALR laws. These 3 laws have been shown to be effective in reducing impaired-driving fatal crashes (Klein 1989; Shults et al. 2001; Voas et al. 2000; Wagenaar and Maldonado-Molina 2007; Wagenaar et al. 2007). We also selected laws designed to promote general driving safety, which would likely impact traffic fatalities including sobriety checkpoints designed to cut down on impaired drivers (Fell et al. 2003; Lacey et al. 1999; Peek-Asa 1999) and mandatory seat belt laws (primary and secondary enforcement; Voas et al. 2007). For the current study, we partially replicated a design utilized in previous research by Fell et al. (2009), which found that additional variables assessing the economic strength of the state may contribute to the ratio of drinking drivers to nondrinking drivers involved in fatal crashes. These variables included per capita beer consumption and indications of economic strength in the state including employment rates and vehicle-miles traveled (VMT). Sources of Data Traffic Fatalities Annual traffic fatality data for each state was derived from the NHTSA’s FARS database from 1982 to 2010 (NHTSA 2013b). The FARS database is a census of motor vehicle traffic crashes that resulted in the death of an individual within the 30 days following the crash. To be included in the FARS database, the crash must have (1) occurred on U.S. public roadways and (2) been attended to and reported by police. The involvement of alcohol in the crash is derived from positive BAC data from tests of drivers involved in the crashes. When BACs for drivers are missing, the BACs are imputed to obtain a more accurate picture of the involvement of alcohol in the crash (Subramanian 2002). Thus, since 1982 every driver in FARS has a BAC from testing by the police or medical examiners or an imputed BAC based upon other crash data (e.g., time of crash, police assessment that driver had been drinking, age and gender of the driver, number of vehicles in the crash, etc.). This information was available in FARS for all years examined in the current study (1982 through 2010) and deemed appropriate for these analyses because many of the MLDA-21 laws were adopted in this 29-year period. It is feasible that any number of variables could have a potential impact on rates of drinking-driver-related crashes (i.e., road conditions, weather, non-alcohol-related policies, etc.). As such, it would be ideal for any study exploring factors related to drinking-driver-related crashes to account for and control for each of these variables. Even if it were possible to know all of the potential factors that may contribute to drinking-driver-related crashes, obtaining accurate operational measures for each would be extremely difficult if not impossible. However, because many of these unknown factors
67 would likely also impact non-drinking-driver-related crashes, the utilization of a control “non-drinking driver” condition should wash out many of these potential effects. For the purposes of this study, the outcome variables were computed from FARS data for each year by state. Subjects were characterized by age (younger than age 21 or aged 26 years and older) and then the number of drinking-driver-related crashes was divided by the number of non-drinking-driver-related crashes. This gave us the ratio of positive to negative BACs for those younger than age 21 and for those aged 26 years and older (drivers with BACs > 0.01 as a ratio to drivers with BACs = 0.00). As was the case in Fell et al. (2009), we also avoided using data on drivers between the ages of 21 and 25. By doing so, we were able to examine 2 groups of drivers without concern of potential carryover effects. False Identification and Social Host Prohibitions and Social Host Civil Liability Laws The current study utilized similar methods to Fell et al. (2009) in measurement of the presence of laws. That is, years in which the law was effective from January through December were coded as 1 and years in which the law was not present at any time throughout the year were coded as 0. The years in which laws became effective were coded as a proportion to reflect the time of year in which a law was in force. For example, if a law was in effect in July where it would be present for only half of that year, it was coded as 0.50; if a law was in effect in April where it would be present three quarters of the year it was coded as 0.75, and so on. The complete years leading up to the effective dates of a law were coded as 0 and the complete years following the effective date of a law were coded as 1. Finally, data were gathered to determine and code the strength of each of these laws. The strength coding for the 3 FID laws (i.e., minor, retailer support, and supplier) and SHP laws was accomplished in our earlier research (Fell et al. 2008). However, the strength coding for SHCL was not available in our earlier research and subsequently was computed in a similar fashion for the current study. FID minor laws were given a strength based on whether the states had administrative and/or judicial sanctions in this law. Because all states prohibit minors from using FIDs, a score ranging from 1 (which would indicate that no license sanctioning procedure is in place) to 3 (indicating a law with administrative or both administrative and judicial sanction procedures in place) was used. FID retailer support laws were given a strength based on the presence of incentives for retailers to use scanners, having distinctive drivers licenses issued for easier identification of those under age 21, seizure of suspicious ID by retailers, and the right to sue a minor. These laws lost points if the state law allowed for a general affirmative defense in court. An affirmative defense example would be that the furnisher inspected the youth’s identification and came to a reasonable conclusion based on the appearance that it was valid. Another example would be that the furnisher came to a good faith or reasonable decision that the alcohol receiver was aged 21 or older. Subsequently, scores for FID retailer support may range from 0, indicating no retailer support provisions, to 5, indicating all provisions except a general affirmative defense in court. FID supplier laws were given a strength based on whether the (a) lending, transfer, or sale of FID and the (b) manufacturing
68
Fell et al.
Table 2. Point values in determining strength of fake identification laws Scoring criteria
Weight point values
Fake identification—minor Fake identification law License sanction procedure
+1.0 point for presence of the law +2.0 points for administrative sanctions OR +2.0 points for administrative and judicial sanctions OR +1.0 points for judicial sanctions only Fake identification—supplier +1.0 point if yes +1.0 point if yes
Lending, transfer, or sale of false IDs criminalized Manufacturing and distribution of false IDs criminalized
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Fake identification—retailer support +1.0 point if yes +2.0 points if yes +1.0 point if yes +1.0 point if yes −1.0 point for general affirmative defense 0.0 points for specific affirmative defense or none
Incentives for retailers to use scanners Use of distinctive licenses Seizure of suspicious ID by retailer permitted Right to sue minor Affirmative defense
Source: Fell et al. (2008).
and distribution of FID is criminalized. Subsequently, scores for FID supplier laws may range from 0, indicating no law against providing FIDs, to 2, indicating that both the lending and manufacturing of FIDs is criminalized (see Table 2). The SHP laws were given a strength score based on the statutes addressing underage actions, property types, and knowledge levels. Specifically, states were given a strength score using the following point values described in Table 3. As such, states may have had a SHP strength score ranging from 0 (indicating no SHP law) to 10 (indicating a general statute covering all underage actions, all property types, with negligence as the knowledge standard and no exceptions to the law). In a court case against a social host, the prosecution must present sufficient evidence to meet a knowledge standard to establish a violation. The more demanding the standard in terms of the evidence required, the lower the likelihood of a successful prosecution. The following categories are ranked by the level of evidence required, from most demanding to least demanding: • Overt act—The host must have actual knowledge of specific aspects of the party and must commit an act that contributes to its occurrence. • Knowledge—The host must have actual knowledge of specific aspects of the party; no action is required. • Recklessness—The host may not have acted with actual knowledge of the party but must act with intentional disregard for the probable consequences of her or his actions. • Negligence—The host knew or should have known of the event’s occurrence (in legal terminology this is referred to as “constructive knowledge”). The SHCL laws were scored in a similar fashion by examining established statutes and/or by state court through common law. A designation of social host civil liability signifies that liability may be imposed for the negligent provision of alcohol to a minor. The law is limited by narrowing the scope of who may be sued (i.e., such as only individuals over the age of 21 as is evident in some states) and by the necessity of proving that the social host knew that the guest was underage and/or a need to provide evidence beyond a reasonable
doubt. With a base score of 3.0 points allotted for having a SHCL law, scores can range from 0, indicating no law, to 3, indicating either or both common and statutory laws with no limitations. Driving Safety and Drinking and Driving Laws Driving safety laws were composed of the strength of sobriety checkpoints and a composite score of primary and secondary seat belt usage laws. Data concerning the strength of sobriety checkpoints was derived from previous research by Fell et al. (2003), who interviewed state representatives in 2000 to determine the number of checkpoints conducted in the previous year. Similar to Fell et al. (2009), we used reported frequency Table 3. Point values in determining strength of social host laws Scoring criteria Social host prohibitions Type of statute
Underage guest actions triggering violation
Property type covered by law
Knowledge standard
Preventive actions available to offender Exceptions to law
Social host civil liability Common law or common law and statutory law Statutory law only Limitations to the law
Weight point values +2.0 points for general OR +1.0 point for specific +1.0 point for possession +1.0 point for consumption +1.0 point for intention to possess or consume +1.0 point for residence +1.0 point for outdoor area +1.0 point for other areas +2.0 points for negligence +1.0 point for knowledge 0 points for overt act −1.0 if yes −1.0 point for family −1.0 point for resident of household −1.0 point for other +3.0 points +3.0 points −1.0 point for limitation on who may be sued −1.0 point for limitation on elements or standards of proof
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Reducing Underage Drinking Driver Fatal Crashes of sobriety checkpoints and coded them as 0 if checkpoints are illegal or otherwise not conducted, 1 if they were conducted infrequently, and 2 if they were frequently conducted such as on a monthly or weekly basis. Similarly, seat belt use laws were determined by whether states had laws dictating mandatory seat belt use. If states did not have this law (i.e., New Hampshire), they were coded as 0, if they had secondary seat belt laws they were coded as 1, and if they had secondary seat belt laws upgraded to primary seat belt laws they received a code of 2. As such, in both sobriety checkpoints and seat belt laws, higher scores indicated stronger laws. Drinking and driving laws were primarily composed of the establishment of 0.08 and 0.10 illegal limit laws and ALR laws, all of which were coded as 0 if the law was absent and 1 if the law was present. As was done with the previous laws, if seat belt laws, illegal BAC limit laws, or ALR laws were changed in the middle of the year, scores were altered to reflect the proportion of the year the law was in effect. Beer Consumption Per capita beer consumption rates were obtained for individuals aged 15 years and older by year and state. These data were obtained from the annual publication of the NIAAA’s Alcohol Epidemiologic Data System. Of note, beer consumption rates were only available as general numbers by state and year. Ideally, it would have been preferable to have these numbers available by age range to allow for a more accurate representation of the role of beer consumption on FARS statistics for those younger than aged 21 years and those aged 26 years and older. However, because these data were not available, per capita beer consumption by state and year was deemed sufficient for our analyses (Fell et al. 2009). Economic Strength Similar to Fell et al. (2009), economic strength was determined by state unemployment rates and VMT. VMT was derived from the Federal Highway Administration data. The Federal Highway Administration produces an annual estimate of total VMT by state and year. Unemployment statistics were derived from the Bureau of Labor Statistics (BLS), which publishes monthly employment statistics by state (Fell et al. 2009). The method of data collection utilized by the BLS underwent 2 notable changes during the years examined in the current study. In 2007, the BLS data included a seasonal adjustment that is designed to eliminate the influences of holidays, weather, and other recurring seasonal events. In 2010, the data also included a population control designed to account for vital statistics on births, deaths, migration, and other population-related variables. The Hypothesized Model The hypothesized model used in this study is illustrated in the path diagram (Figure 1). The model is composed of (1) 3 FID laws (i.e., minor, retailer support, and supplier); (2) 2 versions of the social host laws (i.e., SHP and SHCL laws); (3) 3 alcoholimpaired driving laws (0.08 legal limit, 0.10 legal limit, and ALR); (4) 2 driving safety laws (use of sobriety checkpoints
69 0.08 law 0.10 law
Alcohol Impaired Driving Laws
ALR law < 21 FARS Ratio (drinking to nondrinking drivers)
FID - minor FID - supplier
Beer consumption
FID - retailer
SHP SHCL
Checkpoints
Driving Safety Laws
> 26 FARS Ratio (drinking to nondrinking drivers)
SB laws
Employment
Economic Strength
VMT
Fig. 1. Full structural equation model (SEM) for underage drinking laws.
and seat belt use laws); (5) 2 variables determining economic strength (employment rate and VMT by state and year); (6) per capita beer consumption by state and year; (7) ratio of under age 21 drinking drivers to underage 21 nondrinking drivers in fatal crashes; and (8) ratio of drinking drivers to nondrinking drivers in fatal crashes among drivers aged 26 years and older as a comparison. The model assumes that all 3 FID laws and both types of social host laws directly affect both the younger than aged 21 years FARS ratio and, potentially, beer consumption. Similarly, we assumed that both the latent variables measuring drinking and driving laws and economic strength directly affect beer consumption and both younger than aged 21 years and aged 26 years and older FARS ratios, but latent variables predicting driving safety laws only directly affect both FARS ratios. We also assumed no covariance between latent variables. Further, we have evidence that beer consumption also directly affects both FARS ratios. Because beer consumption has previously been identified as highly correlated to the FARS alcohol ratios, its inclusion in the model was deemed appropriate (see Voas et al. 2007). It was also deemed that FID laws could have a potentially important impact on beer consumption, which, in turn, would also impact FARS ratios. Therefore, in this model, beer consumption is both a predictor and an intermediate outcome, and FARS ratios among individuals younger than aged 21 years and older than aged 26 years are both outcome variables. Furthermore, though we would expect that the use of a control condition could wash out many of the unmeasured variables that may impact crash rates, some residual data will likely still be evident. As such, it was expected that the residual data from both the younger than aged 21 and aged 26 and older FARS ratios would likely be covariates (NHTSA 2013b). We ran the model incorporating the strength of all 5 laws: FID minor, FID retailer, FID supplier, SHCL, and SHP. Strengths of these laws were adapted from Fell et al. (2008) where possible and as described
70
Fell et al.
earlier; otherwise, they were coded in a similar fashion when data were not available in previous research as was the case with the SHCL laws.
Downloaded by [Umeå University Library] at 12:28 01 April 2015
Data Analysis Structural Equation Modeling The data were analyzed using structural equation modeling (SEM) techniques with Analysis of Moment-Based Structures (AMOS v.21), an SPSS-based package (IBM SPSS Inc., Chicago, IL). SEM is a statistical technique frequently utilized to estimate causal relationships based on qualitative assumptions represented in a path diagram. SEM allows for confirmatory and exploratory modeling of both observed variables (rectangles in Figure 1) or latent variables (ovals in Figure 1) derived from combinations of other observed vari¨ ables (Joreskog 1966, 1969, 1967). The use of SEM has gained notable popularity among researchers both for its utility in exploring relationships beyond what is possible with simple analyses of variance or multiple regression analyses and for its applicability to a variety of functions. SEM generally requires a large number of data points (>200; Kline 2005). Because the current study had 1,479 data points (51 states × 29 years),
SEM was deemed appropriate. Further, SEM was used in the current analysis to more accurately account for simultaneous effects of laws on multiple outcomes (i.e., FARS ratios and beer consumption) and because beer consumption was modeled as an intervening variable rather than only a predictor of FARS ratios.
Results The model incorporating the strengths of the laws indicated that one of the FID laws had a significant negative impact on the FARS ratios of drivers who were drinking to drivers who were not drinking. Specifically, for those younger than aged 21 years, only the FID supplier law was associated with significant decreases in FARS ratios (−1.0, P = .030; see Table 4 and Figure 2). However, FID laws dealing with minors and suppliers and SHP laws were found to have a negative but nonsignificant impact on FARS ratios for drivers younger than aged 21 (Table 4 and Figure 2). Beer consumption, which we also hypothesized to be impacted by FID and both social host laws, however, displayed different patterns. SHP laws—not SHCL laws—were found to have a significant impact on annual per
Table 4. Regression weights and significance level for direct effects including strength of fake identification and social host MLDA-21 lawsa Model including strength of FID and SHP laws Predictor
Estimate
SE
P value
0.019 0.007 0.003 0.006 0.005 0.003 0.001 0.003 0.022
