ABSTRACT
The value of goal structure information was shown by Smith and Goodman (1984). They provided subjects with instructions for a circuit assembly task that were either step-by-step directions with little explanation (the linear condition), or that had additional explanatory material. This material was either pure goal structure knowledge (the structural condition), or a mixture of goal structure and how-it-works information (the functional condition). They found that reading time and errors were worst for step-bystep instructions, and that the structural condition showed the best recall and transfer to a similar circuit, with the functional condition elose behind. This implies that the goal structure information is the most useful for procedure construction, although elearly these results are too sketchy to resolve the issue of the merits of goal structure versus how-it-works information. Konoske and Ellis (1986) performed similar experiments in which they provided subjects with step-by-step instructions for the assembly of a model crane that seem to have been either with no explanation or with goal structure explanations. Subjects with goal structure information performed better both initially and after one month. However, in one experiment, the subjects were V.S. Navy personnel with mechanical experience, and for these subjects, there was no advantage for the explanatory information. This suggests that subjects who have the requisite domain knowledge benefit less from explanatory material.
