The use of semantic networking in teaching programming

The authors explored the use of computer-based semantic networks (semnets) as a learning tool in a programming class. They selected 40 students from an introductory C programming language class in the School of Electrical Engineering at Purdue University to build semantic networks on their knowledge of C. Students updated their networks biweekly and gave their feedback on the project at the end of the semester. In order to compare semnet quality with class achievement, the authors devised a quantitative method to evaluate semnets. In applying this method to a small number (five) of semantic networks they found a correlation of 0.55 between semnet weights and class grades. They found that traditional use of semnets in the programming language class is of limited benefit. In comparing `semnet students´ with the remaining students in the class, it was found that the `semnet students´ received higher letter grades but a careful analysis showed smaller differences in actual achievement. The authors conclude with student feedback in using semnets and suggestions for future use of semantic networks in a programming class