HOME-STUDY DRIVER EDUCATION
compare the students completing the four courses on their drive test pass rates because a drive test outcome was available only for 61 (4.6%) of the 1,321 students. This low number is likely due to the 6-month waiting period required by California’s graduated licensing law prior to teens being able to take a behind-the-wheel driving test. The analysis of variance (ANOVA) technique was used to determine if there were any statistically significant differences between the four instruction methods on each remaining outcome measure. Tukey post-hoc multiple comparison tests were used as follow-up tests whenever an ANOVA was statistically significant, to determine exactly which of the course means differed from the others. An alpha level of .05 was used to determine the statistical significance of each omnibus ANOVA and the Tukey post-hoc tests. (An alpha level of .05 means that the differences are deemed significant if the probability of their occurrence by chance alone is less than 5 times in 100.) However, any mean difference with an observed significance probability level that was less than .10 was considered to be a “suggestive” finding.
Two different statistical comparisons were conducted for the study exit examination outcomes and also for the DMV written test. The first analyses, referred to as “Overall,” compared the students in the four instruction methods as assigned by the driving schools, regardless of whether or not the assignment was done correctly (i.e., randomly). For whatever reason, a total of 193 students (14.6%) were incorrectly assigned by the schools to the wrong instruction method based on their day of birth. In addition, the correctness of assignment could not be determined for an additional 48 students (3.6%) for whom the information necessary to make this determination was missing. Because of the potential for self-selection bias when random assignment is not done correctly, a second set of analyses called “Correctly Assigned” is presented that compares only the 1,080 students in the four instruction methods who were correctly randomly assigned by the provider schools.
Two additional sets of analyses based on correctly-assigned students were conducted for the exit test knowledge and attitude outcomes which further divide the study subjects into groups based on whether they were simultaneously enrolled in driver training “Driver Training” or not “No Driver Training.” These analyses were conducted to determine the influence of on-the-road training on the study results. Students enrolled in driver training were not compared on DMV written test pass rates because they take their written test before completing driver education instruction.
Two final sets of analyses were conducted for study exit examination knowledge and attitudes and also for the DMV written test. In these analyses, the students who did not simultaneously enroll in driver training were further divided into those who had access to a computer when they were assigned to an instruction method (and hence could have potentially been assigned to the CD-ROM program) called “Computer,” and those who did not have access to a computer called “No Computer.” These sub-analyses of the non-driver training students were conducted to determine if the pattern of performance across the courses was the same for students who did or did not have access to a computer. In particular, the students with access to a computer may have come from higher socio-economic status households and with more highly educated parents, and hence may have performed better, regardless of the instruction method, than the students who did not have computer access. Obviously, only the three