Learning a troubleshooting strategy: The roles of domain-specific knowledge and general problem-solving strategies.

Abstract

This research investigated how college students learned a general-purpose troubleshooting strategy, dependency-directed backtracking (DDB). The subjects' task was to find the broken components in networks that were similar to digital circuits. Subjects could make tests of the information flowing along particular network lines and replace network components. With only minimal training in this task, subjects usually used a strategy of backtracking from the incorrect network output. They generally did not use the more sophisticated DDB strategy, which involves backtracking from the bad output, but also eliminating components that lead into good network outputs. Computer simulation modeling suggested that in order for subjects to induce the DDB strategy on their own, they needed to apply (1) certain key domain-specific knowledge about how the components worked, and (2) the general reductio-ad-absurdum (RAA) problem-solving strategy. In an experiment, subjects in five different conditions were given different kinds of extra training beyond the minimum necessary to do the task. The types of extra training were: none (baseline); relevant domain-specific knowledge (i.e., that highlighted by the model); other, irrelevant, domain-specific knowledge; relevant domain-specific knowledge and the RAA strategy; and irrelevant domain-specific knowledge and the RAA strategy. DDB use was measured by the type of tests subjects made and the timing of the tests. The model's predictions were supported. The group that received both relevant domain-specific knowledge and the RAA strategy showed a large and significant increase in DDB use. The other groups showed very little improvement. These results suggest that students can use general problem-solving skills such as RAA to learn other general skills such as DDB, but only when they have the relevant knowledge about domains in which these skills can be applied.