Using a General Problem-Solving Strategy to facilitate learning

Abstract

The effectiveness of general problem-solving strategies to create domain-specific knowledge, as well as transfer effects, has been investigated since the information processing age. This research has failed to provide any convincing evidence that they can facilitate learning, because consistent with the research on expertise, the acquisition of domain-specific knowledge requires specific problem-solving strategies to be mastered. This thesis revisits general problem-solving strategies from a new perspective, based on considerations of different types of knowledge and the re-conceptualisation of cognitive load theory with evolutionary biology. A consistent finding of expert-novice research is that when solving problems, experts work forwards using domain-specific knowledge, whereas novices work backwards using general problem-solving strategies, such as means-ends analysis. In this thesis it is argued that such general problem-solving strategies, which are not formally taught, have evolved over time to assist humans in solving problems when no prior knowledge is available. Therefore, it was hypothesised that for novice learners in a domain, a general problem-solving strategy may have significant advantages in initially developing domain-specific knowledge. In contrast, learners who already possess domain-specific knowledge such a strategy would be redundant. Four experiments tested these predictions. Experiments 1 and 2 were conducted in geometry using a general problem-solving strategy based on a goal-free strategy of finding as many unknowns as possible. In both experiments, which compared a general problem-solving strategy with a conventional problem-solving strategy, novice learners benefited from the general problem-solving strategy on far-transfer problems, whereas high prior knowledge learners did not. In Experiments 3 and 4, which investigated learning in economics and geography, new general problem-solving strategies were created by adapting the goal-free approach to the different domains. In both experiments the general problem solver led to higher learning gains than a conventional problem-solving strategy on similar (similar to acquisition) and transfer (dissimilar to acquisition) tests for all learners. All experiments supported the general hypotheses that when prior knowledge is low, learners could benefit from using a general problem-solving strategy during knowledge acquisition. However, when prior knowledge was higher, no benefit was gained.