Repeat Offenders and Persistent Drinking Drivers in the U.S.

James Hedlund and James Fell

NHTSA, NTS-20, 400 Seventh St SW, Washington, DC 20590, USA


Substantial progress has been made in the United States (U.S.) in reducing the impaired driving problem. Future gains will be more difficult because of certain high risk groups of drivers who are more resistant to change. Persistent drinking drivers are one of those groups: they drink and drive again and again; their behavior apparently has not been affected by information and education; and many of them have not been deterred by laws and enforcement or even by arrest and punishment. We need to know who this driver is, what he knows and thinks, and how he acts. This paper gives a brief overview of recent data and studies on the persistent drinking driver primarily from the U.S. Programs being initiated to deal with some of these high risk drivers will be discussed.


Motor vehicle crashes involving impaired driving are a significant public health problem in the U.S. In 1994, there were an estimated 16,900 fatalities in alcohol-related crashes, which is the lowest number in many years (NHTSA, 1995). However, it still represents an average of one alcohol-related fatality every 30 minutes. In addition, an estimated 1.2 million people are injured annually in alcohol-related crashes and the burden of these crashes on society is over $46 billion in direct costs alone (Miller and Blincoe, 1994). Almost 2 in every 5 Americans will be involved in an alcohol-related crash at some time in their lives (NHTSA, 1994) Substantial progress has been made, however, since 1982. In that year, 57% of the 43,945 traffic fatalities were estimated to be alcohol-related (i.e. a driver or pedestrian/bicyclist had a positive blood alcohol concentration (BAC)). This translated to 25,170 alcohol-related traffic deaths. In 1994, the 16,900 alcohol-related traffic deaths represented 42% of the total and was a 33% decrease from 1982. While this progress is gratifying, the toll is still unacceptably high and much of it is preventable. Recently, the U.S. Secretary of Transportation held a meeting with over 100 representatives of government, private industry, and citizen activists which resulted in the support of a goal of reducing alcohol-related traffic deaths to 11,000 by the year 2005 (NHTSA, 1995). In order to reach this goal, the meeting representatives agreed that certain high risk drivers needed to be targeted for special program emphasis. One of these groups includes persistent or repeat drinking drivers: the person who drinks and drives again and again, week after week, month after month, year after year; the person whose drinking and driving behavior has apparently not been changed by information and education, who has probably not been deterred by drinking and driving laws and enforcement, and perhaps even by arrest and punishment for drinking and driving law violations. He (almost always he) has been called hard-core, problem drinker, alcoholic, anti-social. He appears periodically in the press after a tragic crash in which he kills innocent victims and prompts repeated calls to "do something" about him. This paper gives a brief overview of recent data and studies on repeat offenders and persistent drinking drivers primarily from the United States (TRB, 1995).


The most objective data on drinking drivers in crashes come from the Fatal Accident Reporting System (FARS). FARS data are limited in several important respects. FARS includes only fatal crashes. FARS contains only information from official sources, such as police reports and driver records, and consequently is silent on many important issues. FARS does contain information on drivers with prior driving while intoxicated (DWI) convictions before they had their fatal crash. This is a narrow definition of the persistent drinking driver: convictions, not arrests, within the past three years only. Alcohol-positive (BAC=.01 g/dl or greater) drivers with a prior DWI conviction probably are persistent drinking drivers, since they have been caught and punished for this offense fairly recently. However, many persistent drinking drivers have not been convicted recently, as as data shown in the next section will indicate.

FARS data indicate that 2,252 drivers with at least one prior DWI were involved in fatal crashes in 1992 while 47,880 other drivers involved in fatal crashes did not have a prior DWI on their record. Analyses indicate:

The prior-DWI drinking drivers are older, more frequently male, drive older vehicles, wear belts less frequently, and have more rural crashes than the no-prior-DWI drinking drivers. Drivers with prior DWI convictions have also been found to be overrepresented in fatal crashes and the relative risk of fatal crash involvement is greater for these repeat offenders according to studies. One study showed that about 3 percent of all licensed drivers had a prior arrest for DWI within the past three years, yet 12 percent of intoxicated drivers involved in fatal crashes had at least one prior DWI conviction in the past three years. In that same study, intoxicated drivers with prior DWI convictions had 4.1 times the risk of being in a fatal crash as intoxicated drivers without prior DWIs (Fell, 1992). Another study showed that this fatal crash risk increases with the number of prior DWI arrests (Brewer, et al, 1994).

Another method of examining persistent drinking drivers in FARS is to analyze drivers with very high BACs -- .20 g/dl and above. Only fatally-injured drivers from the 25 states that recorded BAC levels for over 80% of all fatally-injured drivers in 1992 were examined to eliminate possible bias:

While they overlap considerably, these two groups are far from identical. In particular, most of the high-BAC fatal drivers do not have a prior DWI conviction within the past three years. Compared to the prior-DWI drinking drivers, the very high-BAC fatal drivers are younger, less frequently male, drive newer cars, have more single-vehicle and more nighttime crashes, and wear belts less frequently.

Both FARS data sets give only a partial view of the persistent drinking driver. Drivers with a prior DWI who are involved in a fatal crash are a small subset of all persistent drinking drivers. Very high-BAC drivers involved in a fatal crash may not all be persistent drinking drivers.


As noted above, FARS data underestimate the proportion of drivers in fatal crashes with a prior DWI since FARS considers only convictions within the past three years. Another view of the size of the repeat DWI offender problem comes from state records of DWI arrests or convictions.

A recent survey by the National Highway Traffic Safety Administration (NHTSA) of the repeat DWI offender problem has produced the results presented in Table 1 (Fell, 1995). The extent of the DWI repeat offender problem in these 12 states where data were available ranges from a low of 21% of drivers convicted of DWI in Iowa in 1992 up to a high of 47% of all drivers convicted of DWI in New Mexico in 1990. The median is around 31-32% of arrests and/or convictions. This means that almost a third of all drivers arrested or convicted of DWI each year is a repeat DWI offender.

Table 1
Incidence Of Repeat Offenders In Selected States

StateNo. Drivers Convicted of DWINo. Drivers Convicted of Repeat DWIRepeat Offense Within Past:Percent Repeat DWI OffendersYear(s)
IA18,0003,7806 years21%1992
LA101,16124,9185 years24%1989-1993
NE146,62038,54830 years26%1985-1994
WI169,39052,0735 years31%1984-1988
NC65,71421,0287 years32%1988
OH637,678211,2805 years33%1980-1993
CA216,45372,7287 years34%1991
NM16,1847,63730 years47%1990
StateNo. Drivers Arrested for DWINo. Drivers With Repeat Arrest for DWIRepeat Offense Within Past:Percent Repeat DWI OffendersYear(s)
SD8,8212,0905 years24%1993
CO99,84826,3355 years26%1989-1991
TX352,372125,94110 years36%1987-1990
MN30,71714,03430 years46%1993

Certain cautions in the state data should be noted. Some states report repeat offenders from DWI arrests, most report from DWI convictions. It is possible that those convicted may have a higher probability of having a repeat offense than those who are not. Note in Table 1 how a repeat offense is defined, i.e. how far back in the driver record qualifies as a repeat DWI. For most states it is only within the past 5-7 years; for Minnesota, Nebraska and New Mexico it was 30 years, or the life of the individual driver record. This would explain why some proportions of repeat offenders are higher in some states than others. A report from California is revealing in this aspect. Their study showed that of all the drivers convicted of DUI in 1980, 11% repeated within the first year, 25% repeated within 3 years, 35% within 5 years, and 44% within 10 years (Peck, 1994).

It also should be noted that most states only check their own state records for previous DWIs for any given driver, not other state records or the National Driver Register. It is possible that many drivers arrested or convicted of DWI in one state had other DWI arrests or convictions in another state that never got into the current driver's record. This would render the repeat offender statistics as conservative.

From these data it appears that from one-quarter to one-third of all DWI offenders in a typical state are repeat offenders. Conversely, two-thirds to three-quarters have no prior DWI offense on record.


Several review papers in the last 15 years have summarized available literature on drinking driver characteristics. A recent review of more than 130 original and review papers (Kennedy, 1993) summarizes key information as follows. (Numerical estimates below are approximate ranges from the majority of studies reviewed.)

DWI Demographics

DWI Personality and Attitude (compared to all drivers)

DWI Drinking Behavior

This summary doesn't specifically address the persistent drinking driver. However, it's clear from the description of the drinking behavior that these persons are persistent drinkers, and it's fairly safe to infer that many are repeat DWI offenders most are persistent drinking drivers as well.


A 1993 NHTSA national survey of drinking and driving attitudes and behavior sampled 4010 persons aged 16 and above via a random digit dialing method (Boyle, 1995). A total of only 56 said that they were frequent drinking drivers in that they drank on at least 12 days, and drove after drinking on at least 8 days, in the month before they were surveyed. (By this definition, 1.4% of the adult population admits to being persistent drinking drivers.) With the twin cautions that the sample size is relatively small and the information is self-reported, these 56 frequent drinker-drivers reported the following:

Demographics and Drinking Behavior:

Attitudes about Drinking and Driving:

Attitudes on Drinking and Driving Laws and Enforcement:

These respondents certainly fit the definition of persistent drinking drivers: they drink frequently (almost every day) and drink and drive frequently (more than twice a week). However, they differ in important respects from the crash-involved drinking drivers in FARS and in the literature. They are somewhat older, they drink at home rather than in bars or taverns, and they don't drink as much at one sitting. They believe that drinking and driving is an important highway safety problem and they seem to accept DWI laws at approximately current BAC levels. They also have a far higher expectation of detection, arrest, and sanction if they drink and drive than occurs in practice. Again, these results must be interpreted with caution. They are self-reported data from a small sample in a telephone survey.


These data suggest that persistent drinking drivers are far from homogeneous. Most American adults drink (64% in the NHTSA survey), and 15% admit to driving after drinking in the past month. The data presented above help identify the most persistent of these drinker drivers -- those who admit to driving after drinking almost every other day. But even these very persistent drinker-drivers differ in important respects from drinker-drivers who are involved in fatal crashes.

The persistent drinker-driver as defined in the NHTSA survey fits the intuitive definition precisely: he drinks and drives regularly and repeatedly. But he may not be our most appropriate target. He seems to behave as a generally responsible member of society; in particular, he is rarely convicted of DWI (and, by inference, rarely involved in a crash). We may instead wish to focus on an even higher-risk group, as defined in the FARS data and the literature, who appear to be heavier drinkers, more anti-social, and more difficult to affect through traditional traffic safety measures. This group of persistent drinking drivers are persons who have driven after drinking repeatedly, especially with high BAC levels.

Persistent drinking drivers are likely to be resistant to change, since their behavior has persisted despite drinking and driving prevention and deterrence activities.

Repeat DWI offenders are a subset of persistent drinking drivers who require special attention. They comprise 1/4 to 1/3 of all DWI offenders in a typical state.


About a third of all drivers arrested for DWI are repeat offenders, and about 1 out of 8 intoxicated drivers in fatal crashes have had a prior DWI conviction within the past 3 years. This means that repeat offenders (from official records) do not constitute the majority of the DWI problem in the U.S. The likelihood of arrest for DWI varies from about 1 in 200 instances in some communities to 1 in 2,000 in others (NHTSA, 1990). Therefore, prevention of DWI in the first place and dealing effectively with first time DWI offenders is still a rational and necessary approach to the problem: state laws, enforcement, and public information and education. These methods have been effective in reducing impaired driving and alcohol-related crash deaths (Fell, et al, 1993).

However, the data presented in this paper indicate that many persons who are arrested and convicted of DWI continue to drink and drive. To address this problem, a number of states and communities in the U.S. have initiated programs and sanctions to deal with the repeat offender (Popkin, 1994). These include:

Further research is needed to evaluate the effectiveness of these sanctions. Certainly it seems reasonable that mandatory alcohol problem assessment and the assignment of appropriate treatment based upon that assessment, in addition to sanctions, should be instituted by the states.

While repeat DWI offenders are only a subset of persistent drinking drivers in the U.S., they are at least identifiable. We must find ways to better identify the other persistent drinking drivers. NHTSA is conducting research in this area which may shed more light on the problem. This ongoing research will determine how persistent drinking drivers make their decisions to drink and drive and what values, standards and norms they live by. Possible intervention strategies might then be developed which have a greater chance of success. In order to reach the goal of reducing alcohol-related fatalities by 35% (to 11,000) in the year 2005, this high risk group of persistent drinking drivers must be affected.


Boyle, J. M., National Survey of Drinking and Driving Attitudes and Behavior: 1993, National Highway Traffic Safety Administration, DOT HS 808 202, January 1995.

Brewer, R.D., Morris, P., Cole, T., Watkins, S., Patetta M., Popkin, C., The Risk of Dying in Alcohol-Related Automobile Crashes Among Habitual Drunk Drivers, New Eng. Jour. of Med., Vol. 331, No. 8, August 25, 1994.

Fell, J.C., Repeat DWI Offenders: Their Involvement in Fatal Crashes, National Highway Traffic Safety Administration, Proceedings of 12th International Conference on Alcohol, Drugs and Traffic Safety - T92, Editors: Utzelmann, Berghaus, Kroj, Cologne, Germany, October 1992.

Fell, J.C., Hedlund, J., Vegega, M.E., Klein, T.M., Johnson, D., Reduction in Alcohol-Related Traffic Fatalities - United States, 1990-1992, Morbidity and Mortality Weekly Report, Vol. 42, No. 47, December 3, 1993.

Fell, J.C., Repeat DWI Offenders in the United States, National Highway Traffic Safety Administration, NHTSA Traffic Tech. No. 85, February 1995.

Kennedy, B. P., Characteristics of Drinking Drivers: Literature Review, Injury Control Center, Harvard School of Public Health, 1993.

Miller, T.R. and Blincoe, L.J., Incidence and Cost of Alcohol-Involved Crashes in the United States, Accid. Anal and Prev., Vol. 26, No. 5, pp. 503-591, 1994.

National Highway Traffic Safety Administration (NHTSA), Alcohol and Highway Safety 1989: A Review of the State of Knowledge, U.S. Department of Transportation, Washington, D.C., DOT HS-807-557, March 1990.

National Highway Traffic Safety Administration (NHTSA), Traffic Safety Facts--1993: Alcohol, National Center for Statistics and Analysis, Washington, D.C., July 1994.

National Highway Traffic Safety Administration (NHTSA), Partners in Progress: Impaired Driving Goals and Strategies for 2005, U.S. Department of Transportation, Washington, D.C., March 1995.

Peck, R.C., Quantifying the Net Accident Contribution of Convicted DUI Repeaters: Some Methodological Issues and Preliminary Findings, California Department of Motor Vehicles, 73rd Annual Meeting of Transportation Research Board, Session 123, January 10, 1994.

Popkin, C.L. and Wells-Parker, E., A Research Agenda for the Specific Deterrence of DWI, Journal of Traffic Medicine, Vol. 22, No. 1, 1994.

Transportation Research Board (TRB), Strategies for Dealing with the Persistent Drinking Driver, National Research Council, Transportation Research Circular, Number 437, February 1995.

Return to Contents
Return to T'95 Home Page