Goodwin and Sanati compare two approaches to teaching an introductory Pascal course: the “traditional” approach (four lectures a week, plus a two-hour lab) and the “Paslab” approach (two lectures a week, plus a three-hour session with the Paslab learning package developed by the authors). The Paslab package allows the student to step through example programs line by line; Paslab displays the input and output data, as well as the contents of program variables.
The authors began their study with the following hypothesis: “Under the new learning conditions, as compared with traditional conditions, background characteristics of students will have much less impact on final grades, while attitudes will have much greater impact.” They attempted to validate this hypothesis by collecting information from students taking either the traditional course or the Paslab course. Students filled out questionnaires at the beginning of the course and midway through the course. The questions concerned both the students’ backgrounds (e.g., SAT scores, knowledge of BASIC) and their attitudes towards computers and programming. The authors then used multiple regression to determine which factors best predicted each student’s grade in the course.
Some of the resulting statistics support the authors’ hypothesis. For students in the traditional Pascal course, previous knowledge of BASIC was the factor most strongly correlated with success in the course. With Paslab, the authors found no such correlation. On the other hand, other statistics seem to contradict their hypothesis. With Paslab, previous knowledge of Pascal became a factor in student performance (for students in the traditional course, it was not).
In any event, the validity of the authors’ results is questionable. Worcester Polytechnic Institute’s three-point grading scale is too coarse for this kind of study; the authors’ method of expanding the scale (by having three levels of failure) is dubious. Also, most of the factors they studied were not very successful at predicting grades for either the traditional course or the Paslab course.
The authors’ statistics, flawed though they may be, are intriguing. In the traditional Pascal course, for example, the number of high school physics classes taken by a student had a strong negative correlation with the student’s grade. (Knowledge of physics can be hazardous to your Pascal grade?)