UBC Theses and Dissertations
Simple and contingent biconditional problem solving in three concept learning paradigms Hartman, Bryan Douglas
The structure of a classification appears to consist of two components: (a) relevant attributes and (b) the classification rule which combines the relevant attributes to describe the classification. An experiment was conducted to separate attribute identification (Al) from rule learning (RL) and compare these with the complete learning (CL) of a classification which requires learning both components. The comparison was conducted for two biconditional classification rules, a two attribute, simple biconditional rule (SB), and a three attribute, contingent biconditional rule (CB), and for two sets of solution strategy instructions, an intra-stimuli (RA) strategy involving classification according to the combination of relevant attributes on each card, and an inter-stimuli (ER) strategy involving classification according to the number of relevant attribute discrepancies between each stimulus card and an exemplar focus card. These three experimental factors were combined with two hypothesized control factors, sex and problem order, in a 3x2x2x2x2 factorial design. Each of 48 grade ten Ss (24 male and 24 female) completed both biconditional problems in one of two counterbalanced problem orders. Performance was recorded on six dependent variables: (1) Trials; (2) Errors; and (3) Seconds - all to a criterion of 27 consecutive correct responses; as well as the post-criterion variables, (4) Classifications, the number of correctly classified cards for a withheld subset of the stimulus population used for original learning; (5) Verbalization, a verbal response which describes a classification rule that separates the cards into two mutually exclusive and exhaustive categories; and (6) Strategy, the classification of the verbal response as implying the RA or ER strategy according to whether reference was made to relevant attribute combinations or relevant attribute discrepancies. In general, the results were as follows. First, the obtained order of paradigm difficulty for the Trials, Errors, Seconds, Classifications, and Verbalization variables was CL > Al > RL, This result was interpreted as support for the Al and RL component approach of Haygood and Bourne (1965), and as an extension of this approach to SB and CB rules. Second, the obtained order of rule difficulty for the Trials, Errors, and Seconds variables was CB > SB, This result was interpreted as support for the rule results of Shepard, Hovland, and Jenkins (1961), and as an extension of their results to include four dimension, bivariate stimuli. Third, the obtained order of difficulty for the strategy instructions was RA > ER, But, only for the Seconds variable was this result significant. Consideration of the results for the Strategy variable supported the conclusion that the instruction treatment was not sufficient to overcome the tendency of Ss to choose their own strategy. Consequently, several suggestions for a more effective instruction treatment were offered. Fourth, the obtained correlations between the Classifications and Verbalization variables were .89 and .79 for the SB and CB problems respectively. This result was interpreted as an indication that further investigation of the Classification variable as a method of determining concept attainment would be worthwhile. Finally, the educational emplications of this study were discussed.
Item Citations and Data