Prompting and Feedback Variables in Concept Programming

A standard practice in teaching concepts is to have students interact with definitions of the concepts to be taught, sets of conceptual problems, and answers to the conceptual problems (Coldeway & Merrill, Note 1 : Keenan & Grant, Note 2: Klausemeir, Ghatala, & Frayer, 1974; Markle & Tiemann, 1970; Merrill & Tennyson, 1977; Miller & Weaver, 1976; Tennyson & Park, 1980). However, there are many ways in which students can interact with materials designed to promote conceptual learning. Two common means are prompting and feedback strategies. Generally, in a prompting strategy, the student is supplied with the correct answer or solution to the learning task before performing the task. In a feedback or confirmation strategy the student is required to perform the task before being supplied with the correct answer. Although both behavior-analytic (Skinner, 1954, 1958, 1961, 1968) and cognitive (Kulhavy, 1977) theorists have tended to favor feedback methods, a substantial amount of the accumulated data SUDKJO~~S ~ r o m ~ t i n a (Aiken & Lau, 1967; Hall & Borman, 1973;'0'~ay, 1971). jn their review of the literatures of paired-associate verbal learning, perceptual learning, and signal monitoring, Aiken and Lau found prompting to be generally superior to feedback, either in improving recall and retention or in reducing learning time. In instructional applications, both O'Day (1971) and Hall and Borman (1973) found prompting to be more effective than feedback, especially in improving initial learning. Other research in concept learning has pointed to the importance of the type of information presented in answers to conceptual problems. Both Tennyson, Steve, and Boutwell (1975) and Coldeway and Merrill (Note 1) compared answers which merely identified conceptual problems as examples or nonexamples of particular concepts with analysis statement information., This analysis statement information involved explaining why the problem qualified as an example or nonexample in terms of the presence or absence of the critical properties of the concept. In both experiments, analysis statement answers were found to be more effective than example-nonexample information. The purpose of the present experiment was to jointly examine (a) the effects of prompting and feedback conditions on learning from concept-programmed material and (b) the effects of type-amount of information presented in the answers to conceptual problems. In the prompting condition, students first read the answer to a conceptual problem, then read the problem itself, and finally responded by writing the correct answer. In the feedback condition, students first read the conceptual problem, then responded to it by writing their answer, and finally read the correct answer. The students used three types of answer-information: no answer information, example-nonexample information, and analysis statement information.