State Alcohol Estimates
Traffic Safety Facts 1996

U.S. Department of Transportation
National Highway Traffic Safety Administration

The following data provide estimates of alcohol involvement in fatal crashes for the United States and individually for the 50 states, the District of Columbia, and Puerto Rico (not included in the national totals). These estimates are based on data from NHTSA's Fatality Analysis Reporting System (FARS). Unfortunately, known BAC test results are not available for all drivers and nonoccupants involved in fatal crashes. Missing data can result for a number of reasons, the most frequent of which is that persons are not always tested for alcohol.

To address the missing data issue, NHTSA has developed and employs a statistical model to estimate the likelihood that a fatal crash-involved driver or nonoccupant was sober (BAC of zero), had some alcohol (BAC of 0.01-0.09), or was intoxicated (BAC of 0.10 or greater) at the time of the crash. The statistical model is based on important characteristics of the crash including crash factors (e.g., time of day, day of week, type of crash, location), vehicle factors (e.g., vehicle type and role in the crash), and person factors (e.g., age, sex, restraint use, previous driving violations), and whether or not the state had a 21-year-old minimum drinking age law.

The statistical model was developed using all available known data in the aggregate (that is, at the national level) and applied to each individual driver and nonoccupant with an unknown BAC test result. The estimates presented include a mix of both known and estimated BACs.

Great caution should be exercised in comparing the levels of alcohol involvement among states. Differences in alcohol involvement can be due to any number of factors not necessarily directly related to a state's alcohol traffic safety program. Factors affecting alcohol involvement in fatal crashes include:

Population demographics and the economic environment (older drivers and female drivers exhibit lower levels of alcohol involvement, drivers of older vehicles exhibit higher levels of alcohol involvement, pedestrian fatalities as a group exhibit high levels of alcohol involvement);

Degree of urbanization (alcohol involvement in single- and multi-vehicle crashes tends to be greater in urban fatal crashes, while alcohol involvement in nonoccupant fatal crashes is higher in rural areas);

Types of vehicles (motorcycle drivers exhibit high levels of alcohol involvement followed by drivers of light trucks/vans; drivers of medium and heavy trucks exhibit the lowest levels of alcohol involvement).

One of the major differences among states is in the degree of testing for driver and nonoccupant BACs. These differences in testing affect the accuracy and reliability of the estimates presented, and for 1996 range from a low of 15 percent known BACs to a high of 75 percent known BACs. States with higher rates of known BACs yield estimates of fatal crash alcohol involvement with greater accuracy and precision.

A driver or nonoccupant involved is considered alcohol-related if he/she is involved in the fatal crash and exhibits a BAC of 0.01 or greater (the last column on the right in the tables). A fatality is considered alcohol-related if any driver or nonoccupant involved in the crash had a positive BAC. Estimates are presented for four categories:

(1) BAC of 0.00 (no alcohol),

(2) BAC of 0.01-0.09 (low alcohol),

(3) BAC of 0.10 or greater (high alcohol), and

(4) BAC of 0.01 or greater (any alcohol, the sum of (2) and (3)).

Nationwide in 1996, alcohol was involved in 40.9 percent of the traffic fatalities (8.9 percent low alcohol and 32.0 percent high alcohol), translating to 17,126 alcohol-related fatalities.

Estimates of Alcohol-Involved Drivers

The following tables estimate alcohol involvement for the nation and on a state-by-state basis for 1996 and 1982 using NHTSA's discriminant function model as applied to the FARS data. This model estimates BACs of drivers and nonoccupants when their BAC is not available. The estimates presented represent the combination of known and estimated BACs.

A driver or nonoccupant involved is considered alcohol-related if he/she is involved in the fatal crash and exhibits a BAC of 0.01 or greater (the last column on the right in the tables). Estimates are presented for four categories:

(1) BAC of 0.00 (no alcohol),

(2) BAC of 0.01-0.09 (low alcohol),

(3) BAC of 0.10 or greater (high alcohol), and

(4) BAC of 0.01 or greater (any alcohol, the sum of (2) and (3)).

Nationwide in 1996, alcohol was present in 24.9 percent of the drivers involved in fatal crashes (6.2 percent low alcohol and 18.8 percent high alcohol).

Availability of Known BAC Test Results

The following tables present the percentage of drivers and nonoccupants involved in fatal crashes where a BAC test was given and the results were in the FARS file. Individual tables are presented for all drivers/nonoccupants, fatally injured drivers and surviving drivers.

Nationwide in 1996, a total of 16,689 fatally injured drivers had BAC test results out of a total of 24,456, or 68.2 percent. For surviving drivers, BAC test results were known on 8,073 out of 32,337 drivers, or 25.0 percent. Overall in 1996, FARS contained BAC test results on a total of 24,762 drivers out of 56,793 involved in fatal crashes, or 43.6 percent. Statistics for the base year 1982 are also presented. Any individual state proportion greater than the national percentage is considered good. The higher the proportion of drivers with known BAC test results, the more reliable the state estimate.

State-Level Estimates on a Regional Basis

For the states in the various NHTSA regions, the following graphics summarize the estimated percentages of fatalities that were alcohol related (BAC of 0.01 or greater) in 1982 and 1996 and the estimated percentage of drivers involved in fatal crashes with BAC of 0.10 or greater (high alcohol).