A statistical incursion on rule learning

Abstract

The objective of this study was to look for a relationship between intelligent behavior and patterns of responses likely to appear in abscence of conceptual behavior. A multiple choice conceptual test was used for that purpose, this test was administered to 107 female and 30 mate college students. It consisted of 15 examples that each participant had to classify as belonging to one, two or three groups of stimuli. Each group was composed by 10 tetragrams, a tetragram was a series of 4 letters, tetragrams belonging to each group had something in common among them. Subjects' task was to find the characteristic defining each group. For group "A" the common feature was to have two consecutive letters from the alphabet, but in inverse order; for group "B" the feature was to have four consonants; and for group "C" that characteristic was to have three consonants and a vowel. The pattern of responses corresponding to ignorance was classified in a scale of "caution-boldness" depending on whether subjects responded in leaving blancks, responded randomly or responded in excess marking more than one answer. To be under the control of a key rule meant that a subject had to score higher than the statistical criteria for randomness, for answers "A", "B" or "C", according to the indexes resulting from dividing number of correct responses by the total of responses the subject made for each group. No significant differences were recorded between men and women in conceptual behavior, in both groups 43 and 44 % of the subjects behaved randomly, and only 20 % of them responded under the control of the three key rules. Women exhibited a light non significant tendency for "caution". In both groups the level of "boldness" varied inversely to the degree of conceptual behavior. The interrelationship between measurements of intelligent behavior and "behavior styles" is discussed.