Expert adolescent programmers: the cognitive consequences of programming
Grace Cheng Lu · 1992
The purpose of this study is to investigate whether expert adolescent programmers spontaneously generate effective problem solving strategies and to determine whether the problem solving skills acquired from programming were transferable. Four general problem solving skills, namely procedural (decomposing) skill, systematic planning and self-monitoring skill, evaluating skill, and ability to solve complex problems acquired from programming were examined for transferability. Previous research data were based primarily on subjects with low-level programming knowledge. In contrast, this study examined ability-related and age-related differences between expert adolescent programmers and gifted nonprogrammers in problem solving. The subjects consisted of twenty-four programmers and twenty-four nonprogrammers from the age of thirteen to eighteen years. Twelve of the subjects in each group were enrolled in grades nine and ten and twelve were enrolled in grades eleven and twelve. This study included a structured interview with the programmers and two problem solving tasks for all subjects. Multiple Analysis of Variance and a separate analysis of problem solving behaviors were performed to assess the significant differences in problem solving skills among the four groups. The River-Crossing Puzzle of five missionaries and five cannibals is a well-defined problem developed by Simon and Reed (1976). The variables were selected to measure evaluation and decision processes. The Yearbook-Planning Task is an ill-defined problem developed by this author. The variables were selected to measure abilities in planning and dealing with complex problems. The findings demonstrated that the programmers solved the problems more efficiently, developed higher structured plans, systematically monitored thinking processes, and evaluated more often while processing than the nonprogrammers. Younger programmers who had more programming experiences did better than the other three groups in solving the River-Crossing Puzzle. Age-related differences were present in the Yearbook-Planning Task. In conclusion, metacognitive skills gained through programming are continuous in development and transferability of the acquired skills increases with experience in programming.