Search

Subjects have more difficulty in correctly testing a conditional rule stated in abstract form than the same rule given in "thematic" form. The "thematic" form of the rule involves both concrete terms and a concrete relation. Four experimental conditions were formed by combining concrete abstract terms and relations. Fifty subjects were tested in each condition. The results indicated that concrete-ness of the terms is the main factor leading to improved performance with "thematic" as against abstract material in the rule testing task.

A hypothesis development(HD) model is described for concept learning with nvalues per dimension. On the basis of the HD model it was predicted that response times(RTs) would increase, while inspection times(ITs) on errors would decrease and ITs on corrects would increase as the number of values per dimension (v)increased. It was also predicted that RTs and ITs would be positively correlated on correct trials but negatively correlated on errors. These predictions were tested by running an experiment with three levels of v(2, 4, and 6) in which RTs, ITs and success–error results were recorded. There was no systematic ordering of latencies over levels of v. Evidence supporting the prediction of positive RT/IT correlations on corrects was obtained but the prediction of negative RT/IT correlations on errors was not upheld.

Subjects have more difficulty in correctly testing a conditional rule stated in abstract form than the same rule given in “thematic” form. The “thematic” form of the rule involves both concrete terms and a concrete relation. Four experimental conditions were formed by combining concrete abstract terms and relations. Fifty subjects were tested in each condition. The results indicated that concrete-ness of the terms is the main factor leading to improved performance with “thematic” as against abstract material in the rule testing task.