Answers are grounded in this paper and cite their sources.
Earlier T. E. Goldsmith, P. J. Johnson, and W. H. Acton (1991) reported that a Pathfinder-based representation of domain knowledge was more predictive of classroom exam performance than the underlying data from which the Pathfinder representation was derived (relatedness ratings of pairs of concepts from the domain). The question is, what affords Pathfinder this predictive advantage? Analyses reveal that whereas the relatedness ratings are based on a 7-point scale, all of the predictiveness is contained in the most highly related pairs. Moreover, Pathfinder's advantage is related to the manner in which it dichotomizes the relatedness data into directly linked and indirectly linked concept pairs
No takes yet. Share an insight, caveat, or question.
Johnson et al. (1994) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: