II. METHOD
Site Selection
Study states were sought that had Use and Lose laws applicable to alcohol offenses covering persons under the age of 21. The NHTSA regional offices were first canvassed regarding states with such laws that might be willing and able to participate. This, in turn, led to contacts within various states about the study. The main issue was whether a state could provide an adequate sample of persons arrested on a Use and Lose charge and linkage to driver record files. The two states that emerged from this process were Pennsylvania and Missouri.
Missouri
The Law
The Missouri law, titled "Abuse and Lose", is found in Chapter 577.500 of the state's statutes (see Appendix B). The law calls for the suspension of driving privileges of persons under the age of 21 who plea or are found guilty of:
1) Any alcohol related traffic offense; or
2) Possession or use of alcohol, committed while operating a motor vehicle; or
3) Possession or use of a controlled substance; or
4) Alteration, modification or misrepresentation of a license to operate a motor vehicle; or
5) A second offense of possession or use of alcohol by persons under the age of 18.
The suspension period for a first offense is 90 days and one year for subsequent offenses.
Data
Conviction records used in the study came from the Missouri State Highway Patrol's Traffic Arrest System/Alcohol and Drug Offense Records System. These data covered persons under the age of 21 who were arrested in the years 1995, 1996, and 1997 who pled or were found guilty of one of the charges just noted. Arrests and convictions for alcohol related traffic offenses were statewide data while arrests and convictions for the other charges were only those made by the Highway Patrol. The initial file contained 4,843 convictions.
Driver record data were then sought from the Division of Motor Vehicles and Driver Licensing of the Missouri Department of Revenue. Missouri is one of a few states that does not include motor vehicle crash involvements in its driver record files. The resulting data, therefore, only included records of traffic law convictions and actions taken against the drivers' licenses, not crashes. The data request was processed in August 1999.
Among the 4,843 data requests, there were 606 instances where no driver's license number was available. Of these, 417 were matched using name and date of birth and 189 could not be matched and were thus unusable. In addition, there were 387 instances of multiple convictions for the same individual. In these cases, the record with the earliest arrest date was used. The final analysis, therefore, was based on 4,267 cases where a person was convicted of a charge that could expose them to a Use and Lose license action for whom driver history data were available. That is: 4,843 (original) - 189 (not matched) - 387 (multiple convictions) = 4,267 (for analysis).
Pennsylvania
The Law
The Pennsylvania Use and Lose law targeted toward youth is found in Sections 6307- 6313 of the state's Crimes Code (see Appendix B). The law calls for license suspension or delay in licensing of persons under the age of 21 convicted of:
1) Purchase, consumption, possession or transportation of liquor, malt or brewed beverages; or
2) Misrepresenting age to obtain alcohol; or
3) Carrying a false identification card.
The suspension period or delay in licensing is 90 days for a first offense, one year for a second offense and two years for subsequent offenses.
Data
The Pennsylvania Commission on Crime and Delinquency (PCCD) provided data on all cases filed in the state's District Courts during 1995, 1996, and 1997 involving one of the charges just noted. This initial file contained 6,822 cases. The city of Philadelphia has a municipal court system and the city of Pittsburgh has a magistrate court system. Cases from these courts were not available for the study.
The PCCD file was transmitted to the Bureau of Driver Licensing of the Pennsylvania Department of Transportation which did a name/date of birth search of its driver record files to obtain data on the crashes and motor vehicle law violations of the persons involved. In 1,132 cases, no match was obtained. Driver records for 5,690 cases, therefore, were available for analysis.
Analytic Approach
The data from the two states were analyzed separately. The general approach was to group the cases based on the specific Use and Lose charges and, within these, whether or not licensing actions had taken place. Subsequent driving events (violations in Missouri, crashes and violations in Pennsylvania) were then tallied for the resulting subgroups. With the Missouri data, the chi-square statistic was used to test the subsequent violation records of those who did and did not undergo license actions.
With the Pennsylvania data, the subsequent driving performance of those who did and did not undergo license actions were first compared using logistic regression and odds-ratios. Significance of parameter estimates was tested with -2 log likelihood statistic and of the odds ratio with chi-square statistic (p-value <.05). Survival analysis was then employed to estimate the likelihood of subsequent first violation or first crash over time between different groups of license actions. In the survival analysis models, subjects were tracked until their first post input arrest traffic event or until the last date available in the driver history files, but limited to 48 months. Survival was computed by the Kaplan-Meier method. Differences in the survival parameters were tested for significance using the log-rank test. The statistical significance of each parameter was first tested in univariate Cox regression analysis, then significant predictors were entered into Cox proportional hazards multiple regression models. The survival analysis was conducted with the SPSS 10.0 software. A possible interaction between age and gender was tested and found to be not significant. Appendix C provides a summary of this result.
|