Factors predicting drunk driving -self-reported behavior before and after raising a BAC limit Hans-Peter KRUGER & Rebecca LOBMANN, Center for traffic safety, Psychologisches Institut der Universitat W urzburg, Roentgering 11, D-97070 W urzburg, Germany ABSTRACT Drinking drivers and sober controls stopped during a roadside survey in East and West Germany (1992-1994) were subsequently interviewed on the phone (n=2368). The interview covered alcohol consumption, attitudes towards DUI and factors of deterrence. Drunk driving could be predicted more from factors of individual moral commitment than from deterrence. Attitudes differed between East and West Germans. After raising the legal BAC limit in East Germany changes in attitudes and self-reported behavior were observed. INTRODUCTION After the re-unification of Germany the BAC limit in East Germany was raised from 0.00% to 0.08% in January 1993. In the present study corresponding changes in attitude and behavior of drivers were observed in a pre-after design and the process of adapting behavior to the new law was examined. W hereas a study of the consequences of lowering a BAC limit cannot distinguish effects of changing the law from effects of inner moral standards, the present study allows to estimate how strong DUI is influenced by social norms and moral commitment. A natural experiment like the rise of a BAC limit should reveal those protective factors which lead to sober driving despite a more permissive law. -441 -
M ETHOD AND SAMPLE The German Roadside Survey (GRSS) was conducted in two adjacent states of Germany: Unterfranken (West Germany) and Thueringen (East Germany). More than 21.000 drivers were stopped by the police. After a short control the drivers were directed to a separate checkpoint where they were interviewed and asked to supply a breath sample and their phone number for a subsequent interview. The method of the GRSS and its results are described in detail by Kruger et al. (1995). It was intended to interview every drunk driver and a sober control (matched for age, gender and day time of driving). O f the 1899 drunk drivers found in the GRSS 72.4% could be contacted, of these 98.5% answered the questions resulting in a total of 2368 interviews. Table 1 shows the distributions of subjects according to gender, age and region. The high percentage of male drivers and drivers aged 25 to 49 years is due to the fact that most DUI drivers are young to middle-aged males. Table 1: Sample of telephone interviews. Thueringen Unterfranken sum male 18-24 318 257 575 25-49 498 645 1143 50+. 129 163 292 missing 1 4 5 sum 946 1069 2015 8 8,2 % 82,5% 85,1% female 18-24 55 75 130 25-49 6 6 135 2 0 1 50+. 5 17 2 2 sum 126 227 353 1 1,8 % 17,5% 14,9% sum 1072 1296 2368 1 0 0,0 % 1 0 0,0 % 1 0 0,0 % - 442-
RESULTS Effects of raising a BAC limit Interview data of drivers are available for time periods before and after the change (winter 1992, spring 1993, spring 1994). A variety of questions about the acceptance of DUI revealed that East German drivers were less permissive towards drunk driving than W est German drivers in all three components of the Roadside Survey. For example, 57.9% of the drivers in Thueringen, but only 23.8% in Unterfranken preferred a limit of zero-bac. Acceptance of drunken driving was still lower for older East German subjects. No differences between East and West Germans were found in subjective certainty of detection and severeness of punishment, but East Germans had a lesser knowledge about the law. W hen the attitudes of East German drivers were examined over time, the effects in variables measuring acceptance of DUI differed: In some items East German drivers stayed with their opinion (measures against DUI, preferred BAC limit). Other answers indicate a more permissive view over time. In the same time period, West German drivers became more restrictive (perception of friends reaction, amount of alcohol for still being able to drive safely). Figure la and lb illustrate these results for two items. The percentage of East Germans preferring a zero-bac limit remained constant while in 1994 more West German drivers favored it than before. On the other hand drivers in Thueringen forwarded higher amounts of alcohol with which they could still drive safely over time (Figure lb). Changes in self-reported behavior can be described as follows: In the component following the raise of the BAC limit (Spring 1993) East German drivers were asked if they sometimes drove with alcohol. 51.5% agreed (this large value is partly due to the selected sample with as much drunk as sober drivers). O f this group another 49.1% said they were only doing this since the BAC limit had been raised. Of drivers between 18 and 24 years of age a higher percentage admitted driving only recently with alcohol (63.6%). Moreover, drivers showing DUI only after the change of the law had a lower mean BAC in the GRSS as older drinking drivers (0.031% vs 0.043%). -443 -
Figure a and lb: Changes in the acceptance of DUI in Thueringen and Unterfranken. Which BAC limit do you prefer? percentage preferring a 0.0-limit How much alcohol can you drink 1992 1993 1994 In a similar question posed in 1994, subjects were asked whether they had driven under the influence of alcohol in the former GDR, after the re-unification but before the rise of the BAC limit and after the rise of it. Figure 2 displays the answers for DUI drivers (straight lines) and sober drivers (pointed lines) of the GRSS. The Figure illustrates that in the former GDR already 33.3% of the East German subjects drove with alcohol. Drivers classified as drinking drivers because of their BAC in the GRSS had shown this behavior more often before (44.5%) than controls classified as sober. Nevertheless, 29.9% o f the controls still admitted having driven with alcohol some time. In both groups behavior changes took place, but the drinking drivers reacted already upon the re-unification whereas the sober drivers waited for the law to change. -444-
Figure 2: Percentages of self-reported drunk driving at different points in time. in the GDR after the re-unif. after the 0.8? Towards a model of drunk driving To describe attitudes towards DUI and knowledge about the law three factors were used: alcohol consumption, exposure to traffic and drunk driving. A descriptive analysis showed that drunk drivers, heavy drinkers and frequent drivers (high exposure to traffic) have a more permissive attitude towards drunk driving and perceive social norms as less restrictive. Moreover, drunk drivers with a mileage of more than 20.000 kilometers per year assume a lower probability of detection than sober drivers. Mobility therefore, is an important modifiying factor for the effect of certainty of detection. No differences were found in subjective severeness of punishment and knowledge about the law. For a m ultivariate analysis a logistic regression was conducted in which variables of attitude and behavior were used to predict driving with alcohol. Figure 3 shows the odd-ratios of the predictors. These indicate that with a permissive attitude towards drunk driving the probability for DUI is three times as high as with a strong moral commitment. High alcohol comsumption and permissive social norms lead to an increase of DUI driving by 1.5 times. No effect was found for the deterrence factors. -445 -
Figure 3: Results of a logistic regression predicting drunk driving. Predictors are listed on the left side of the figure. Significances are indicated by asterixes (** p<.01; * p<.05). Reading example: A driver with a permissive attitude has a risk for drunk driving which is 3.1 times as high as that of a driver who is strongly agains DUI. T1,0 own norm attitude: perm issive : social norm s friends : perm issive so cie ty: p e rm is s iv e H 1.7 * * BAC lim it: p u n ish m e n t: know n know n d e te rre n c e p o lice contro ls: o fte n p ro b a b ility of d e te ctio n : low 1 1.1 1.0 1 1.1 1 1 1 severeness of punishm ent: low 0.9 d riv in g fre q u e n cy: h ig h Z 3.8 b eh a vio ra l d ata,,, a lco h o l c o n s u m p tio n : h ig h 1.6 * * s o c io d e m o g ra p h ic a l d ata a ge : 25 y.+ 1 1.1 tim e o f ro a d sid e day: w o rk in g d a y co n tro l tim e: n ig h t = 1 1.2 1 1.2 1.0 0.5 5.0 odd -ra tio s (log.) DISCUSSION AND CONCLUSION Moral commitment, social norms and alcohol consumption predict drunk driving. This result is consistent with the findings of several other studies (e.g., Berger und Snortum, 1986; Brown, 1993). W hen drivers have a high exposure to traffic certainty of detection has an effect, too. Comparing East and West Germans the former were more restrictive on drunk driving - especially the people over the age of 50. Despite the rise of the legal limit from 0.00% to 0.08% East Germans remained very conservative in their opinion towards drunk driving when questions covered themes discussed in public, e.g., preferred BAC limit, measures against DUI. If the questions dealt with more private topics (e.g.. How m uch alcohol can you drink and still drive safely?) East German drivers showed an increasingly permissive attitude over time. On the other hand, West Germans disapproved drunk driving more strongly at the last timepoint than before. In short, the citizens of the two states became more - 446 -
alike in their opinion. There were also self-reported behavioral changes of East German drivers: 51.5% reported driving with alcohol, half of which admitted, they were doing so only since 1993. Especially the young reacted upon the rise of the BAC limit which points at the susceptibility of young people towards changes in social norms. Subjects who started driving with alcohol after the change of the law still had low BACs in the GRSS. Nevertheless, the number of potential DU1 drivers has obviously increased. Future research has to evaluate if the increase of DUI driving in East Germany will be permanent. R E FE R EN C ES Berger, D. E. & Snortum, J. R. (1986). A structural model of drinking and driving: alcohol consumption, social norms, and moral commitments. Criminology, 24 (1), 139-151. Brown, S.L. (1995). Associations between peer attitudes and behaviour, random breath testing experience and drink driving in a population-representative sample of south Australia. In C. N. Kloeden & A. J. McLean (Eds.), Proceedings o f the 13th International Conference on Alcohol, Drugs and Traffic Safety. (Vol 1, pp. 159-164). Adelaide, Australia: NHMRC Road Accident Research Unit. The University of Adelaide. Kruger, H.-P., ReiB, J.. Hilsenbeck. T., Kazenwadel, J, Vollrath, M. & Krause, W. (1995). The German Roadside Survey 1992-1994: Aims, Methods and Results. In C. N. Kloeden & A. J. McLean (Eds.), Proceedings o f the 13th International Conference on Alcohol, Drugs and Traffic Safety. (Vol 1, pp. 172-178). Adelaide, Australia: NHMRC Road Accident Research Unit. The University of Adelaide. - 447 -
-4 4 8 -