Age of drinking onset, driving after drinking, and involvement in alcohol related motor-vehicle crashes.

This study assessed whether persons who begin drinking at younger ages are more likely to report drunk driving and alcohol-related crash involvement over the life course, even after controlling analytically for diagnosis of alcohol dependence, years of drinking alcohol, and other personal characteristics associated with the age respondents started drinking. A national survey asked 42,862 respondents the age that they started drinking, whether they drove after drinking too much, and whether they were in motor-vehicle crashes because of their drinking. This analysis focused on 27,081 (65%), who reported ever drinking in their lifetime. The earlier the age respondents started drinking, the more likely they were to report driving after drinking too much and being in a motor-vehicle crash because of their drinking even after adjusting for current/ever diagnosis of alcohol dependence, number of years respondents had been drinking, and other characteristics and behaviors associated with the age respondents started drinking. Particularly, among persons who were never alcohol-dependent, those who began drinking in each age group under 21, relative to those starting at age 21 or older, were more likely to report "ever" and "in the past year" being in a crash after drinking too much. The traffic safety benefits of delaying drinking may extend well beyond the legal drinking age of 21.     

Less is more: the influence of traffic count on drinking and driving behaviour

Drinking and driving road checks are often organized with either a clear prevention or repression objective in mind. The objective of a prevention strategy is to make as many people as possible believe that police officers are enforcing drinking and driving laws and that drinking drivers will most likely be caught. As such, targeting high traffic count road sites with high-visibility road checks is a priority because it serves to increase awareness of the enforcement activity. An alternative to this prevention approach is the “repression” approach that involves targeting times and places where the highest number of drinking drivers are to be expected. Rather than attempting to affect the subjective chance of getting caught, this approach seeks to increase the objective likelihood of getting caught; the aim is to apprehend as many drinking drivers as possible. Regardless of the chosen strategy, there is a need to understand how traffic count influences drinking and driving behaviour as traffic count may play a role in a police officer's choice of sites for a road check. The objective of this paper is to shed some light on this relationship between drinking and driving behaviour and traffic count. In this paper, data from a roadside survey, carried out in British Columbia in 2003, are used. A two-level logistic regression analysis was carried out with data from 2627 drivers coming from 48 different road sites to replicate a model that was previously obtained with comparable data from a Belgian roadside survey, also carried out in 2003. The present study successfully replicated the findings of the Belgian model, substantiating that the probability for drivers to be drinking and driving significantly decreases with an increasing level of traffic count. This supports the suggestion that drinking drivers avoid high traffic count road sites. The relevance of these findings with respect to organizing preventive or repressive road checks and possible confounding variables are discussed at the end of this paper.     

Trends in alcohol-impaired driving in Canada

While a general decreasing trend in the number of persons killed in a traffic crash involving a drinking driver has occurred in Canada since the 1980s, it is evident that much of this decrease occurred in the 1990s. Since 2002, less progress has been made as the number of persons killed in crashes involving drinking drivers remains high. To better understand the current situation, this paper describes trends in drinking and driving in Canada from 1998 to 2011 using multiple indicators based on data collected for the Traffic Injury Research Foundation's (TIRF) Road Safety Monitor (RSM), the National Opinion Poll on Drinking and Driving, and trends in alcohol-related crashes based on data collected for TIRF's national Fatality Database in Canada. There has been a continued and consistent decrease in the number of fatalities involving a drinking driver in Canada. This remains true when looking at the number of fatalities involving a drinking driver per 100,000 population and per 100,000 licensed drivers. This decreasing trend is also still apparent when considering the percentage of persons killed in a traffic crash in Canada involving a drinking driver although less pronounced. Data from the RSM further show that the percentage of those who reported driving after they thought they were over the legal limit has also declined. However, regardless of the apparent decreasing trend in drinking driving fatalities and behaviour, reductions have been relatively modest, and fatalities in crashes involving drivers who have consumed alcohol remain high at unacceptable levels.     

The effect of the new traffic law on drinking and driving in São Paulo,Brazil

Objective

To present data on drinking and driving in the city of São Paulo, Brazil, and the effects of the new traffic law (Law 11,705) introduced in 2008.

Methods

A cross-sectional study was performed using a questionnaire and passive breath test data to study the prevalence of drinking and driving and the association of drinking and driving with background characteristics and drinking patterns on two separate occasions. The data were gathered from 2007 to 2009 through roadside surveys conducted on busy public roads. Four thousand two-hundred thirty-four (4234) drivers were approached, before and after prohibition, from the south, north, east, and west regions of the city of São Paulo, located in southeastern Brazil, including cars, motorcycles, and utility vehicles. A total of 3854 (91%) consented to participate in the survey and answered the questionnaire. Out of this group, 3229 (84%) agreed to take the passive breathalyzer test.

Results

Logistic regression analyses controlling for gender and age was used to predict a positive breath test (above 0.2 g/l) and the impact of the new law. These analyses indicated that, after the passage of the new traffic law, there was a 45% decrease in driver behavior with positive breathalyzer results. Having a pattern of alcohol consumption of at least once a week and the habit of drinking and driving are risks for a positive breathalyzer.

Conclusions

Despite the decline in the frequency of motorists driving under the influence of alcohol, traffic-related injuries and deaths, after the new law, other measures for a public policy related to alcohol should be considered based on scientific evidence, consistency of action, clear goals, community support, and greater reliability in the laws.     

Validity of surrogate measures of alcohol involvement when applied to nonfatal crashes

Since the 1970s, nighttime fatal crashes have been used as a surrogate measure for alcohol-related fatalities for crashes for which more direct measures were absent. The validity of this approach was confirmed in 1985 but has not been re-evaluated since. Although this measure has also been applied to identify alcohol involvement in nonfatal crashes, its validity when applied to nonfatal cases has never been determined.The objective of this study was to evaluate the appropriateness of using nighttime crashes as surrogate measures for alcohol impairment when applied to fatal and nonfatal injury and property damage only (PDO) crashes. To do so, we used data from a crash-control design study collected at the roadside in two U.S. states between 1997 and 1999, as well as from the 1997-1999 and 2004-2006 Fatality Analysis Reporting System.The outcome of this study confirms the validity of using nighttime crashes as a surrogate measure for alcohol-related fatalities and supports the use of after-midnight crashes for measuring alcohol involvement in nonfatal and PDO crashes when the number of late-night crashes permits.     

The association between opioid analgesics and unsafe driving actions preceding fatal crashes

Currently, most epidemiological research into the impact of opioid analgesics on road safety has focused on the association between opioid use and traffic crash occurrence. Yet, the role of opioid analgesics on crash responsibility is still not properly understood. Therefore, we examined the impact of opioid analgesics on drivers (all had a confirmed BAC = 0) involved in fatal crashes (1993-2006) using a case-control design based on data from the Fatality Analysis Reporting System. Cases had one or more crash-related unsafe driving actions (UDA) recorded; controls had none. We calculated adjusted odds ratios (ORs) of any UDA by medication exposure after controlling for age, sex, other medications, and driving record. Compared to drivers who tested negative for opioid analgesics, female drivers who tested positive demonstrated increased odds of performing an UDA from ages 25 (OR: 1.35; 95% CI: 1.05; 1.74) to 55 (OR: 1.30; 95% CI: 1.07; 1.58). For male drivers this was true from ages 25 (OR: 1.66; 95% CI: 1.32; 2.09) to 65 (OR: 1.39; 95% CI: 1.17; 1.67). The detection of opioid analgesics was not associated with greater risk of an UDA for older drivers. Research is necessary to examine why these age differences exist, and if possible, to ensure that opioid analgesics do not contribute to crashes.     

Effects of enforcement intensity on alcohol impaired driving crashes

Background

Research measuring levels of enforcement has investigated whether increases in police activities (e.g., checkpoints, driving-while-intoxicated [DWI] special patrols) above some baseline level are associated with reduced crashes and fatalities. Little research, however, has attempted to quantitatively measure enforcement efforts and relate different enforcement levels to specific levels of the prevalence of alcohol-impaired driving.

Objective

The objective of this study was to investigate the effects of law-enforcement intensity in a sample of communities on the rate of crashes involving a drinking driver. We analyzed the influence of different enforcement strategies and measures: (1) specific deterrence – annual number of driving-under-the-influence (DUI) arrests per capita; (2) general deterrence – frequency of sobriety checkpoint operations; (3) highly visible traffic enforcement – annual number of traffic stops per capita; (4) enforcement presence – number of sworn officers per capita; and (5) overall traffic enforcement – the number of other traffic enforcement citations per capita (i.e., seat belt citations, speeding tickets, and other moving violations and warnings) in each community.

Methods

We took advantage of nationwide data on the local prevalence of impaired driving from the 2007 National Roadside Survey (NRS), measures of DUI enforcement activity provided by the police departments that participated in the 2007 NRS, and crashes from the General Estimates System (GES) in the same locations as the 2007 NRS. We analyzed the relationship between the intensity of enforcement and the prevalence of impaired driving crashes in 22–26 communities with complete data. Log-linear regressions were used throughout the study.

Results

A higher number of DUI arrests per 10,000 driving-aged population was associated with a lower ratio of drinking-driver crashes to non-drinking-driver crashes (p = 0.035) when controlling for the percentage of legally intoxicated drivers on the roads surveyed in the community from the 2007 NRS. Results indicate that a 10% increase in the DUI arrest rate is associated with a 1% reduction in the drinking driver crash rate. Similar results were obtained for an increase in the number of sworn officers per 10,000 driving-age population.

Discussion

While a higher DUI arrest rate was associated with a lower drinking-driver crash rate, sobriety checkpoints did not have a significant relationship to drinking-driver crashes. This appeared to be due to the fact that only 3% of the on-the-road drivers were exposed to frequent sobriety checkpoints (only 1 of 36 police agencies where we received enforcement data conducted checkpoints weekly). This low-use strategy is symptomatic of the general decline in checkpoint use in the U.S. since the 1980s and 1990s when the greatest declines in alcohol-impaired-driving fatal crashes occurred. The overall findings in this study may help law enforcement agencies around the country adjust their traffic enforcement intensity in order to reduce impaired driving in their community.     

A population-based study on the prevalence and correlates of drinking and driving in Hong Kong

To examine the prevalence and correlates of drinking and driving in Hong Kong, an anonymous, random telephone survey was conducted on 9860 Chinese adults (18-70 years of age) from April to June 2006. Trained interviewers administered a structured interview consisting of questions on socio-demographic information, drinking pattern, drink-driving, and motor vehicle accidents. The census age-standardized past-year prevalence of driving within 2 h of drinking was 5.2% among males and 0.8% among females. The prevalence across age showed an inverted U-shaped trend for males peaking at 8.2% between 41 and 45 years. For females the prevalence was fairly stable between the ages of 20 and 55. The past-year prevalence of alcohol-related motor vehicle accidents was 0.1%, with the majority being in the 26-30 age group. For males who drank, the prevalence of drinking and driving was 5.0% among those without problem drinking, 14.8% among binge drinkers, 37.1% among alcohol abusers and 22.4% among the alcohol dependent. For females who drank, the corresponding figures were all lower at 1.2%, 6.9%, 12.1% and 12.5%, respectively. Higher socio-economic status, weekly drinking, binge drinking and alcohol abuse were independently associated with higher likelihood of drinking and driving in both genders. Among drinking drivers, having a job that required drinking was the only predictor of having had a motor vehicle accident. The elevated prevalence of drinking and driving among alcohol abusers, binge drinkers and the alcohol dependent may portend higher population-level rates of alcohol-related motor accidents in the future since the prevalence of problem drinking has previously been noted to be increasing rapidly in Hong Kong.     

Alcohol-related predictors of adolescent driving: Gender differences in crashes and offenses

Demographic and alcohol-related data collected from eighth-grade students (age 13 years) were used in logistic regression to predict subsequent first-year driving crashes and offenses (age 17 years). For young men's crashes and offenses, good-fitting models used living situation (both parents or not), parents' attitude about teen drinking (negative or neutral), and the interaction term. Young men who lived with both parents and reported negative parental attitudes regarding teen drinking were less likely to have crashes and offenses. For young women's crashes, a good-fitting model included friends' involvement with alcohol. Young women who reported that their friends were not involved with alcohol were least likely to have crashes. No model predicting young women's offenses emerged.     

The mixed effects of precipitation on traffic crashes

PURPOSE: This paper investigates the relationship between precipitation and traffic crashes in the US during the period 1975-2000. Traffic crashes represent the leading cause of death and injury for young adults in the US, and the ninth leading cause of death for the overall population. Prior studies have found that precipitation raises the risk of traffic crashes significantly. METHODS: A negative binomial regression approach is employed. Two different units of analysis are examined: state-months and state-days. The sample includes all 48 contiguous states. RESULTS: A surprising negative and significant relationship between monthly precipitation and monthly fatal crashes is found. However, in the daily level analysis, a strong positive relationship is estimated, as in prior studies. The source of the contrasting results appears to be a substantial negative lagged effect of precipitation across days within a state-month. In other words, if it rained a lot yesterday, then on average, today there are fewer crashes. Additional analysis shows that the risk imposed by precipitation increases dramatically as the time since last precipitation increases. For example, 1cm of precipitation increases the fatal crash rate for a state-day by about 3% if exactly 2 days have passed since the last precipitation and by about 9% if more than 20 days have passed. This basic pattern holds for non-fatal crashes as well. CONCLUSIONS: The lagged effects of precipitation across days may be explained by the clearing of oil that accumulates on roads during dry periods or by the conditioning of people to drive more safely in wet conditions. Either way, policy interventions that prepare drivers more adequately for the risks of precipitation following dry periods are likely to be beneficial.     

Correlation between modes of drinking and modes of driving as reported by students at two American universities

This paper examines the correlates and predictors of driving under the influence behaviors (DUIBs) during the past month by college students. Measures of heavy episodic drinking, monthly drinking frequency, monthly drinking variance, monthly drinks per occasion and reported marijuana use are compared as predictors net of other predictive factors. A cross-sectional telephone survey was conducted with college students by a university social science research laboratory. Respondents attended two large public universities located in the southwestern part of the US. Participants included 803 randomly selected college students. The interview schedule included items from the Core Alcohol and Drug Survey and the College Alcohol Risk Assessment Guide. Several additional last-drinking-event items were also developed for the interview. Bivariate analyses indicate that marijuana use (past year), heavy episodic drinking, reports of DUIBs (driving under the influence or riding with a driver who is under the influence) in the past year, monthly frequency of drinking, the average number of drinks consumed when drinking and age are correlates of DUIBs during the past month. Multivariate analyses indicate past year DUIBs, monthly frequency of drinking and monthly marijuana use predicted recent DUIB.     

Identifying social mechanisms for the prevention of adolescent drinking and driving

This study identifies social mechanisms that might help prevent youth from being involved in driving under the influence of alcohol (DUI) and riding with drinking drivers (RWDD). Data collected through telephone surveys with 1534 adolescents and young adults aged 15-20 years (mean=17.6, S.D.=1.6) in California, USA, were analyzed. Structural equation modeling analyses showed that DUI and RWDD were strongly related to drinking in unstructured situations, modeling of DUI by peers and parents, and perceived peer approval or disapproval of DUI. DUI outcome expectancies were indirectly related to DUI and RWDD through situational drinking. Parental monitoring and DUI law enforcement were also indirectly related to DUI and RWDD through DUI expectancies and other mechanisms. The findings, overall, suggest that parental influence remains important even through late adolescence. Parental monitoring, in particular, might help to reduce unstructured socializing with peers, drinking, and affiliation with peers who engage in DUI. Parental monitoring may also foster beliefs about the risks of DUI. Conversely, parents' own DUI behavior may normalize drinking and DUI behaviors, thus countering monitoring efforts.     

Trajectories of kinematic risky driving among novice teenagers

Objectives

Elevated gravitational force event rates are associated with the likelihood of a crash or near crash and provide an objective measure of risky driving. The purpose of this research is to examine the patterns over time of kinematic measures of risky driving among novice teenage drivers.

Methods

Driving data were collected from 42 newly licensed teenage drivers during the first 18 months of licensure. Data recording systems installed in participants’ vehicles provided information on driving performance and driver characteristics. Latent class and logistic regression models were used to analyze trajectories of elevated gravitational-force (g-force) event rates, called kinematic risky driving, with respect to risk groups and associated factors.

Results

Kinematic risky driving over the 18-month study period was best characterized as two classes, a higher-risk and a lower-risk class. The rate of kinematic risky driving during the first 6 months generally maintained over 18 months. Indeed, of those classified by latent class analysis as higher risk, 88.9%, 94.4% and 94.4% had average event rates above the median in the 1st, 2nd, and 3rd 6-month periods, respectively, indicating substantial tracking over time. Friends’ risky driving, friends’ risky behavior, self-reported risky driving, and perceptions about risky driving and driving privileges were associated with trip-level rates of kinematic risky driving. However, none of these factors was associated with trip-level rates after stratifying by overall risk in a latent class model, although friend's risky driving was marginally significant.

Conclusion

Kinematic risky driving tended to track over time within the lower and higher risky driving groups. Self-reported risky driving and having risky friends were predictors of kinematic risky driving rates, but these variables did not explain the heterogeneity within higher and lower classes of risky drivers.     

Macroscopic spatial analysis of pedestrian and bicycle crashes

This study investigates the effect of spatial correlation using a Bayesian spatial framework to model pedestrian and bicycle crashes in Traffic Analysis Zones (TAZs). Aggregate models for pedestrian and bicycle crashes were estimated as a function of variables related to roadway characteristics, and various demographic and socio-economic factors. It was found that significant differences were present between the predictor sets for pedestrian and bicycle crashes. The Bayesian Poisson-lognormal model accounting for spatial correlation for pedestrian crashes in the TAZs of the study counties retained nine variables significantly different from zero at 95% Bayesian credible interval. These variables were – total roadway length with 35 mph posted speed limit, total number of intersections per TAZ, median household income, total number of dwelling units, log of population per square mile of a TAZ, percentage of households with non-retired workers but zero auto, percentage of households with non-retired workers and one auto, long term parking cost, and log of total number of employment in a TAZ. A separate distinct set of predictors were found for the bicycle crash model. In all cases the Bayesian models with spatial correlation performed better than the models that did not account for spatial correlation among TAZs. This finding implies that spatial correlation should be considered while modeling pedestrian and bicycle crashes at the aggregate or macro-level.     

The contribution of alcohol to night time crash risk and other risks of night driving

Many studies show that driving at night is more risky in terms of crash involvements per distance travelled than driving during the day. The reasons for this include the more prevalent use of alcohol by drivers at night, the effects of fatigue on the driving task and the risk associated with reduced visibility. Although the consumption of alcohol prior to driving occurs most commonly at night, drink-driving is not inherently a night time risk factor. This study decomposes the New Zealand risk of driving at night into risk associated with alcohol and risk associated with inherently night time factors. The overall risk associated with alcohol use by drivers was shown to decrease with increasing age for the most risky situation analysed (male drivers on weekend nights). Given the levels of drinking and driving on weekend nights, the overall effect of alcohol was shown to contribute almost half of weekend night time risk for drivers aged under 40 on lower volume roads, but to contribute little to overall risk on higher-volume roads, consistent with other research showing that higher-volume roads are not favoured by drinking drivers. Risk at night relative to risk during the day (excluding risk associated with drinking and driving) was shown to decrease with age. Roads with illumination at night are less risky at night relative to during the day than roads without illumination. The risks estimated in this paper reflect the behaviour of the road users studied and their prevalence on the roads under the conditions analysed.     

The relationship of underage drinking laws to reductions in drinking drivers in fatal crashes in the United States

This study reports on an effort to evaluate and interrelate the existence and strength of two core laws and 14 expanded laws designed to (a) control the sales of alcohol, (b) prevent possession and consumption of alcohol, and (c) prevent alcohol impaired driving by youth aged 20 and younger. Our first analysis determined if the enactment of the possession and purchase laws (the two core minimum legal drinking age laws) was associated with a reduction in the ratio of drinking to nondrinking drivers aged 20 and younger who were involved in fatal crashes controlling for as many variables as possible. The ANOVA results suggest that in the presence of numerous covariates, the possession and purchase laws account for an 11.2% (p=0.041) reduction in the ratio measure. Our second analysis determined whether the existence and strength of any of the 16 underage drinking laws was associated with a reduction in the percentage of drivers aged 20 and younger involved in fatal crashes who were drinking. In the regression analyses, making it illegal to use a false identification to purchase alcohol was significant. From state to state, a unit difference (increase) in the strength of the False ID Use law was associated with a 7.3% smaller outcome measure (p=0.034).     

Prevalence and characteristics of red light running crashes in the United States

About 40% of motor vehicle crashes occur at intersections. In recent years, the number of crashes at traffic signals has increased considerably. A major cause of such crashes is drivers disregarding traffic signals. Despite concerns about the frequent occurrence of red light violations and the significant crash consequences, relatively little is known about the overall prevalence and characteristics of red light running crashes. The present study examines the prevalence of red light running crashes on a national basis and identifies the characteristics of such crashes and the drivers involved. Cities with especially high rates of fatal red light running crashes are identified. Countermeasures to reduce red light running crashes based on collision patterns and characteristics of drivers involved are discussed. It was estimated that about 260 000 red light running crashes occur annually in the United States, of which approximately 750 result in fatalities. Comparisons were made between red light running drivers and drivers deemed not to have run red lights in these same crashes. As a group, red light runners were more likely than other drivers to be younger than age 30, male, have prior moving violations and convictions for driving while intoxicated, have invalid driver’s licenses, and have consumed alcohol prior to the crash. Comparisons also were made between characteristics of red light runners involved in daytime and nighttime crashes. Nighttime red light runners were more likely than daytime runners to be young, male, and have more deviant characteristics, 53% having high blood alcohol concentrations.     

Drinking and driving among Mexican American and non-Hispanic white males in Long Beach,California

Although drinking and driving in the United States has declined substantially during the past two decades, this trend has not been seen among Hispanic drivers. Higher rates of driving while impaired (DWI) arrests and alcohol-related crashes, particularly among Mexican Americans, also have been noted. The extent to which this reflects a lack of understanding of DWI laws rather than a disregard for them is unknown. A survey was conducted among Mexican American and non-Hispanic white male DWI arrestees in Long Beach, California, to ascertain alcohol use, attitudes toward drinking and drinking and driving, and knowledge of DWI laws. The findings were compared with those of Mexican American and non-Hispanic white males recruited from the local community. Mexican American males, both DWIs and those from the community, reported heavier drinking than non-Hispanic white males. All four groups of respondents tended to underestimate the number of drinks needed to achieve the blood alcohol concentration (BAC) threshold at or above which it is illegal to drive under California law. Estimations were around 2-3 drinks rather than a more realistic estimate of 4-5 drinks. However, Mexican American DWIs and their comparison group vastly overestimated the number of drinks to make them unsafe drivers (8- 10 drinks). Furthermore, fewer than half were aware of the BAC threshold in California (0.08%) compared with between 60 and 78% of non-Hispanic whites. This study is limited in scope and needs to be replicated in other communities and with other racial/ethnic groups. However, the clear lack of knowledge of the DWI law in California and a lack of understanding of the relationship between number of drinks and BAC point to the need for culturally sensitive programs that are developed and implemented within the Mexican American community.     

Exploring the effects of attitudinal and perception characteristics on drinking and driving non-compliant behaviour

The objective of this paper is to examine a number of factors (observed and latent) that might have a causal effect on drinking and driving (D&D) behaviour. Face-to-face surveys were conducted among patrons at bars and cafeterias and 305 valid questionnaires were filled. A confirmatory factor analysis was performed so as to identify the latent constructs and a mixed structural equation model was developed. From the analysis it came up that non-compliant behaviour of D&D is limited at older ages, also associated with high levels of income and car availability. Though men are consuming more alcohol, women seem to be more prone in driving under the influence (DUI) of alcohol. Furthermore, it was found that people who strongly support the examined interventions in the study (e.g. better enforcement, more traffic safety campaigns, stricter penalties) are more unlikely to drive after drinking compare to those who have some objections. Finally, it was not found any statistically significant relation between individuals’ level of awareness and D&D behaviour.     

Influences during adolescence on perceptions and behaviour related to alcohol use and unsafe driving as young adults

OBJECTIVE: To investigate whether aggression, parent and peer influences, and previous traffic-related experiences at ages 15 and 18 impacted on (a) differences between the perceived safe and estimated legal alcohol consumption limit, and (b) driving while impaired (DWI) behaviour at age 21. METHOD: The study population was a birth cohort involved in a longitudinal investigation of health and development. At the ages of 15 and 18, study members completed questionnaires assessing parent and peer attachment, experience travelling with an alcohol impaired adult or youth, aggression, and previous crash experience. At age 21, study members were questioned about how much alcohol they perceived they could drink and still drive safely, and whether they had driven after 'perhaps consuming too much alcohol'. For each participant their legal alcohol consumption limit was estimated using their height and weight. Path analysis was used to determine whether variables measured at ages 15 and 18 predicted differences between the perceived safe and estimated legal alcohol consumption limit and driving while impaired, both measured at age 21. RESULTS: Insufficient females drove while impaired at age 21, who also had complete data on all other variables, to conduct path analysis for this outcome. For males, aggression at ages 15 and 18, travelling with an impaired youth at age 18, and previous crash experience at age 18 predicted DWI behaviour at age 21. Only aggression at age 15 predicted the difference between perceived safe and estimated legal alcohol consumption limit for the males. For females, aggression at ages 15 and 18, and travelling with an impaired adult at age 15 predicted the difference between perceived safe and estimated legal alcohol consumption limit. CONCLUSION: The results show that aggressive behaviours and adult and/or youth modelling of drink driving behaviours in mid- to late-adolescence are related to differences between perceived safe and estimated legal alcohol consumption limit for both genders and driving while impaired for males.