The effects of segmented multimedia worked examples and self-explanations on acquisition of conceptual knowledge and problem-solving performance in an undergraduate engineering course

The study investigated the effects of non-segmented multimedia worked examples (NS-MWE), segmented multimedia worked examples (S-MWE), and segmented multimedia worked examples enhanced with self-explanation prompts (S-MWE-SE) on acquisition of conceptual knowledge and problem solving performance in an undergraduate engineering course. In addition, the study examined learner-generated self-explanations for possible interrelationships between the type and quality of self-explanations and learning outcomes.;A total of 62 engineering students from a 300-level civil engineering course participated in the study as a part of their course. After the initial introduction of the concepts by the instructor in class, they completed the treatment and the conceptual knowledge test online as a part of their homework assignment, and completed a problem-solving test in class as a part of a course quiz. Participants' homework logs were retrieved to examine the generated self-explanations, and determine time on task and amount of access to the multimedia files.;A one-way ANOVA test did not reveal any significant differences between any of the groups on either outcome measure. Time on task also did not reveal any significant differences between the groups. However, in-treatment problem solving performance scores showed a significant difference between the S-MWE-SE and the NS-MWE groups with the S-MWE-SE group performing significantly better.;There were several positive linear relationships found between the principle-based type of explaining, quality of explaining, and problem solving performance, as well as between correctness of self-explanations and acquisition of conceptual knowledge. Multiple regression results revealed that quality of self-explanations was a significant predictor of performance on the problem-solving test for the S-MWE-WE group.;Multimedia access logs from the groups that received segmented multimedia example treatments were examined for possible patterns using a two-step clustering procedure. As a result, three distinct groups were identified based on their amount of access to each of the 7 segments: the most frequent viewers, moderate viewers, and the least frequent viewers. One-way ANOVA revealed that the most frequent viewers performed significantly worse on the problem-solving test than the other two groups. The results followed a similar pattern on the conceptual knowledge test score, but the differences did not reach the significance level. However, all three types of viewers received very similar scores on the in-treatment performance.;It was also discovered from the logs that several participants may have collaborated on the treatment. Both collaborating and non-collaborating groups were examined to determine whether any possible differences in performance existed. There was a significant difference in performance on the conceptual knowledge test with the collaborating group performing significantly better. Multiple regression results revealed that principle-based explaining was a significant predictor of performance on the problem-solving test for non-collaborating participants in the S-MWE-WE group.