Statistical Evaluation of a “Boot Camp” Course for Preparing Students for Success in a FORTRAN Programming Course

Evaluation of new educational methods traditionally uses randomly selected treatment and control groups. However, if it is believed that a new method is effective then it may be desirable to allow anyone who wants to participate to do so (i.e., to not withhold the new “treatment” from possibly deserving students). This complicates evaluation of the effectiveness of the new method. We evaluate the effectiveness of a weekend “boot camp” overview course used to prepare students for a three credit hour FORTRAN programming course. Boot camp students did remarkably better in the overall course, but is this because of what was learned in boot camp or is it because the students entering boot camp were “better” to begin with? Using collected survey data and final course grades from both boot camp and nonboot camp students (totaling 256 students of which 73 enrolled in boot camp over two semesters), we attempt to answer this question. We break down the students into groups by age, gender, year, major, transfer status, experience, GPA, course load, and external load. Hypothesis testing shows that boot camp helped (or appears to have helped) most groups of students. We show that students younger than the class average received a greater benefit from the boot camp than did older students. Boot camp also benefits underclassman (freshman and sophomores) more than upperclassman. These observations suggest that the boot camp is most effective for students in the early stages of college, motivating us to encourage new students to take the boot camp. With these methods, we hope that the cause of improvement in final class grade for some student groups can be convincingly stated as being due to boot camp. Certainly, the stated results should be enough to convince any student to enroll in boot camp; the one weekend pain is small compared to the likely (but, not guaranteed) gain.