HOME-STUDY DRIVER EDUCATION
ability to detect a difference between the means when the difference actually does exist. None of the other home-study courses differed significantly from classroom or from each other (ps > .05). Therefore, with the possible exception of better attitudes for CD students, students who completed home-study driver education courses had attitudes that were similar to those completing classroom courses.
A supplementary analysis comparing correctly assigned non-driver training students in PEN to comparable students in the other courses who were assigned only after PEN was added to the study also did not indicate a significant difference between PEN and any of the other instruction methods (ps > .05).
The supplementary analyses used to determine if the relationships among the courses differed as a function of the provider school that the students attended did not indicate a statistically significant Instruction Method x Provider School interaction on CPI attitudes (p > .05).
First-Attempt DMV Written Knowledge Test Pass Rate Comparisons
The participants were also compared on their pass rates for their first DMV written test attempt. However, recall that teens who simultaneously enroll in behind-the-wheel driver training take their written test before they actually complete their driver education course. Therefore, only students who were not enrolled in driver training and who had actually taken the DMV written knowledge test since completing their driver education course were compared. In addition, not all of the students who did not take driver training had taken a written knowledge test at DMV to apply for their instruction permit by the time of this analysis. The number of non-driver training students missing a written test outcome was 18 (10.1%) for classroom instruction, 10 (18.5%) for the PEN course, 37 (14.6%) for the CD-ROM course, and 34 (10.2%) for the workbook course. Results of a chi square test of independence indicated that the missing written score rates were not significantly different among the four courses, χ2(3, N = 820) = 5.37, p = .146. The written test analyses are therefore based on the 721 students with a valid written test score who did not simultaneously enroll in driver training.
The number of students with a valid first attempt DMV written knowledge test score, and the number and percentage of these students who passed the test, are shown in Table 8 for each instructional method and statistical comparison. The percentages of students passing the courses are also shown in Figure 3. Results of the one-way ANOVA for each level of statistical comparison are presented in Table 9 along with the percentage-point difference between each home-study pass rate and the classroom instruction pass rate. Negative differences would indicate that a higher percentage of the classroom students passed the written knowledge test on their first attempt than did the home-study students in a particular course. Positive differences would indicate that a higher percentage of home-study students passed the written test on their first attempt than did the classroom students. The percentage point differences that were significantly different from classroom according to the Tukey post-hoc tests are asterisked in Table 9.