A study of engineering college students' use of computer-based semantic networks in a computer programming language class
Abstract
Computer-based semantic network construction has been used as a tool in educational research to probe and reveal cognitive structure. Studies have identified needs to investigate the changes in the quality and quantity of specific knowledge structures constructed by students before, during and after instruction. Further needs include the characterization and measurement of knowledge structures as related to instructional purposes and learners' individual differences. In this study, the use of computer-based semantic network construction was explored for student learning, for teaching, and for cognitive research. More specifically, the impact of building semantic networks on engineering students while they studied and learned C-programming language was investigated. Out of 130 students enrolled in a C-language programming class for electrical and computer engineers, 37 were randomly selected to learn semantic network concepts and regularly generate computer-based semantic networks related to the course content. The 37 students completed personal profile questionnaires, two self-reports that reflect individual differences (Schmeck's Inventory of Learning Processes and Petty and Cacioppo's Need for Cognition Questionnaire), and took one test (Guay Rotations Test). Out of the 37 selected students 12 were videotaped for qualitative data collection. Students who built semantic networks scored better in class achievement than students who did not. However, the results were not statistically significant at an $\alpha$-level of.05. Changes in semantic network variables over time were analyzed along with student feedback. Variables associated with semantic networks increased in number as learning progressed. Students' feedback toward the study had advantages and disadvantages. Roughly 30% of the 37 students thought that they would use computer-based semantic networks if they had access to them, 15% were undecided, and the remaining students indicated they would not use it. A 'quality' measure of a semantic network was defined and analyzed. The quality measure was helpful, yet it needed refining to reflect more accurate results. Relationships between individuals' cognitive and motivational factors and semantic network variables were documented showing the correlation between the Methodical study measure on Schmeck's Inventory of Learning Processes and class grade to be.45. Finally, the use of computer-based semantic networks proved to be an excellent platform to reveal the way students construct knowledge, and the way they articulate knowledge into such a knowledge structure.
Degree
Ph.D.
Advisors
Lehman, Purdue University.
Subject Area
Curricula|Teaching|Engineering|Computer science|Educational software
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.