Search

Based on cognitive load theory, this paper reports on two experiments investigating the variability effect that occurs when learners' exposure to highly variable tasks results in superior test performance. It was hypothesised that the effect was more likely to occur using high rather than low levels of guidance and testing more knowledgeable than less knowledgeable learners. Experiment 1, which tested 103 adults studying pre-university mathematics, showed no interaction between levels of variability (high vs. low) and levels of instructional guidance (worked examples vs. unguided problem solving). The significant main effect of variability indicated a variability effect regardless of levels of instructional guidance. Experiment 2, which tested another group of 56 adults enrolled in the same mathematics program, showed an interaction between levels of variability (high vs. low) and levels of learner expertise (novices vs. experts). More experienced learners learned more from high rather than low variability tasks demonstrating the variability effect, while less experienced learners learned more from low rather than high variability tasks demonstrating a reverse variability effect. It was suggested that more experienced learners had sufficient available working memory capacity to process high variability information while less experienced learners were overwhelmed by high variability and learned more using low variability information. Subjective ratings of difficulty supported the assumptions based on cognitive load theory. The major educational implication is that learners should initially be presented with low variability or easier tasks, and as they gain more experience in the task domain, variability or task difficulty should increase.

According to cognitive load theory, using worked examples is an effective and efficient instructional strategy for initial cognitive skill acquisition for novice learners, as it reduces cognitive load and frees up cognitive resources to build task competence. Contrary to this, productive failure (as well as invention learning, desirable difficulties and other problem-solving-first) frameworks suggest that scaffolded problem-solving activities (a generation phase) preceding explicit instruction enhance learners' performance. The reported experimental study investigated the effectiveness of different levels of instructional guidance provided to students during the learning phase preceding explicit instruction in standard solution procedures in enhancing students' engagement and transfer problem-solving skills. Specifically, the study compared partially-guided or unguided attempts at generating problem solutions as opposed to comprehensive guidance, in the form of a worked example. Levels of experienced cognitive load and learner motivation were evaluated in addition to delayed post-test performance scores. There were no differences between the three groups on the transfer post-test outcomes, even though the condition with fully-guided worked examples prior to the explicit instruction expectedly reduced cognitive load relative to the conditions without such guidance. There were also differences between the conditions on some sub-dimensions of the motivation scale. In general, the findings indicate that similar overall outcomes (delayed transfer performance) could be achieved by different sequences of instructional tasks aimed at achieving different sets of specific goals.