A Note on the Calculation of Strategies in Concept Attainment
Joe L. Byers · American Educational Research Journal · 1967
Since the publication of A Study of Thinking (1956), in which it is noted that strategies play a significant role in concept attainment (CA), there has been some attention given to various methods of calculating CA strategies (Byers, 1963).2 This paper reports a comparison of two slightly different methods of determining selection strategies that are appropriate for use with conjunctive concepts. The CA task may be briefly described as follows. The S is told what the possible relevant attributes are and that in the concept he is seeking any combination of them may be relevant. He is shown that he can determine the concept the E has in mind by testing stimulus cards and inferring from such tests which attributes are relevant and which are irrelevant. The problem begins when the S is given one exemplar or positive instance (PI) of the concept, called the focus card. The S then begins to test any other stimulus cards he wishes. When the S selects a card for testing, the E informs him whether the card selected is a PI or a negative instance (NI) of the concept. Each S card selection and the subsequent E feedback is called a trial. The E keeps a record of all trials as they occur. The problem terminates when the S states the correct concept which he is supposed to attain in as few trials as possible. The strategy employed in a CA problem is inferred from the characteristics of cards the S selects. In a situation as described above there is an optimum strategy, i.e., a rule for selecting stimulus cards to be tested. This rule is optimum in the sense that it places a minimum strain on the memory of the S and therefore should, on the average, result in efficient CA. When this optimum strategy, called conservative focus-