The application of language acquisition theory to programming concept instruction: Chunks versus programs from scratch

Elizabeth Anne Gates Applin

Abstract

This study explored the relationship between assignment based teaching methods and achievement in an introductory programming course. Subjects in the study were 42 community college students in south Mississippi. All participants completed the Watson-Glaser Critical Thinking Appraisal and the Productive Environment Preferences Survey. These instruments along with ACT English score, ACT math score, and high school math grade point average formed the pretreatment measures for the subjects. After two identical programming assignments, students in the control group wrote all programming assignments from scratch while students in the experimental group began adding their coded solutions to larger well-documented programs furnished by the researcher/instructor. Various cognitive psychologists tell us that instructional content should be grouped into chunks, delivered in an optimal sequence, and presented in a context at or just beyond the learner's understanding to force deep processing and permanent memory of the content. Advance organizers allow scaffolding of new ideas onto old established concepts and this introduction presented in advance of the actual topic coverage in a classroom setting has proven effective. Language acquisition studies indicate that new vocabulary must be introduced in context to aid understanding. Second language acquisition theories also suggest that new material should be presented in forms that are just beyond the learner's current level of competence in the language. The analysis of covariance statistical procedure was used show effect for group. The adjusted means for the control group and the experimental group on the variable final exam grade were 55.8053 and 62.0232 respectively with an observed significance level of .019. The adjusted means for the variable final course average were 69.8892 and 73.4523 respectively with an observed significance level of .043. Qualitative observations of students from the two groups in CS2 the following semester seem to indicate that the template methodology can have positive long term effects on many characteristics of student programs. The programming style, documentation practices, parameter passing mechanism use, code reuse, and modularization practice of those students who had been part of the experimental group was superior to that of the control group participants.