An empirical comparison of techniques for teaching recursion in introductory computer sciences
John E. Greer · 1987
The present study examined the effects of three approaches to teaching recursion on students' subsequent achievement in recursion. Eighty-four students who were registered in CS315 at the University of Texas at Austin during the fall semester of 1986 were randomly assigned to three instructional treatment groups. One group was taught the concept of recursion by means of an architecture-oriented approach, which focused on the mechanism by which recursion is implemented. A second group was taught by means of a theory-oriented approach which focused on the principle of mathematical induction as the basis for recursion. The third group was taught by means of a task-performance-oriented approach which focused on the underlying goal-hierarchies experts use in coding and analyzing recursive programs. Video-taped lectures developed for each of the three instructional treatments were presented to the three groups. Students' achievement in recursion was measured with a posttest immediately following the instructional treatment and with two follow-up retention tests. No significant difference in achievement in recursion was detected among the three groups. Students identified as having higher general computer science ability did demonstrate higher achievement in recursion, regardless of instructional treatment. Two interview/problem-solving sessions, one immediately after the instructional treatment and the other at the end of the semester were conducted with six students. During the first interview session students from the architecture-oriented group seemed to use plans consistent with sequential and iterative processes learned prior to studying recursion. The high ability student from the theory-oriented group attempted to construct a recursive interpretation of the problem, and nearly constructed a correct solution. During the second interview session students did not use the naive plan of first constructing an iterative algorithm and then making it recursive. The theory-oriented and task-performance-oriented students tended to focus on separation of the terminating and recursive cases to a greater extent than the architecture-oriented students. The present study documented a possible link between approaches to instruction and the evolution of programming plans. Further research that focuses on the acquisition of recursion-programming plans, as well as the relationship between programming plans and success in constructing and analyzing recursive programs, is needed.