The Short-term Effectiveness of Written Driver Knowledge Tests

About the report:

Highway Safety Program Standard 5, Driver Licensing, issued by the U. S. Department of Transportation requires the states to test applicants for a renewal of their operator's license on rules of the road at least once every four years. The cost of additional manpower to administer and score these tests, the expense of expanding existing facilities to deal with nearly one million additional applicants per year, and the cost of test materials led the state of Virginia to request a waiver of the testing portion of the standard until evidence could be presented to show that knowledge testing had the desired safety benefit. The state, in requesting this waiver, proposed to conduct a study of the efficacy of knowledge testing as an accident/conviction reduction countermeasure. The test subjects comprised four groups of drivers: a control group receiving no treatment, a group that received only a driver's manual, a group that received a manual and a test to be completed at home, and a group that received a manual and were requested to take a test in the examining station at the time of application for license renewal. Comparisons between groups were made of accidents, major convictions, minor convictions, accidents with an associated conviction, and administrative actions taken as a result of points accumulated under the Driver Improvement Program. For the two groups administered a knowledge test, comparisons involved those who passed, failed, or refused to take the test. (Since Virginia statutes do not require knowledge testing for every renewal applicant, there was the probability of a refusal group.) The study findings are to be presented in two reports. This report covers the first six months of driving exposure for each applicant and deals with the short-term effects of the program. A second report, to be prepared at the end of two year's driving exposure, will deal with long-term effects. The short-term findings of the study can be summarized under two broad categories: comparisons where statistical differences were not proven to exist, and comparisons where a statistical difference between groups did exist. The comparisons within each of these categories are: the control group compared to an experimental group, two experimental groups compared to each other, and when performances on a knowledge test are compared. Of the 135 comparisons carried out, there were no statistical differences which reached significance, p = .05, in 125 of them. Of the 10 comparisons in which a statistical difference was found, 7 involved applicants who refused to take the home test. In each case their driving records were worse than the records of those in the group to which they were compared. These findings for applicants refusing to take the home test do not provide state licensing officials with meaningful data for the implementation of a knowledge retesting program. In addition, 2 of the other 3 comparisons where a significant difference was found involve accident with conviction data where the sample size is very small and thus limits the practical effects of the statistical results. Because of the number and nature of the categories that were different, it is concluded that knowledge testing does not improve short-term driving performances as measured in terms of accidents, convictions, and administrative actions.

Authors