Instructional control strategies and content structure as design variables in concept acquisition using computer-based instruction.

Two instructional design variables directly related to concept learning were investigated. The first variable, management strategy, tested the hypothesis that advising students of their learning need in relationship to acquisition of a task at a given criterion would be more effective than either adaptive control or learner control strategies. Data analysis showed that for college students, the advisement condition resulted in better performance (j < .001) than the learner control and needed less instructional time (j <,.005) and fewer instructional instances (p < .001) than the adaptive control. The second variable contrasted two forms of content structure used in learning coordinate concepts. Students given concepts simultaneously performed better on the posttest (p < .001) and needed less instruction (á < .005) than those who received concepts successively. Instructional Control Strategies and Content Structure as Design Variables in Concept Acquisition Using Computer-Based Instruction One of the concerns of research in the control of the learning environment is adaptive education or the adaptability of the instructional presentation to the individual differences of learners. The term adaptive instruction, however, has frequently been used to describe quite different things (Tennyson & Rochen, 1979). According to Glaser (1977), it combines the development of an individual's initial competence with alternative environments matched to different styles of learning. Landa (1976) defines adaptive instruction as a diagnostic process aimed at adjusting the instructional context to the unique learning needs of each student. For Rochen and Tennyson (1978), it implies a diagnostic process which assesses a variety of learner indices (such as general aptitudes related to the learning task and prior task achievement) and characteristics of the learning task (such as difficulty level, content structure and conceptual attributes) in order to prescribe an initial instructional program which can be adjusted continuously to student os-task learning needs. In contrast to these structured adaptive instructional methods, in which the amount and sequence of the instructional stimuli are made without strategy inputs from the student, the learner control method allows the student to be responsible for the learning strategy. Although learner control, in this way, seems to exhibit the elements of adaptability, instructional research dealing with variables of learner control (using rather large on complex learning tasks) have failed to demonstrate that students can make and carry out decisions of content element selection and personal learning assessment (DiVesta, 1973). In experimental learning tasks which required minimal prior knowledge and included a simple content structure, the learner control stragegy usually resulted in less time on-task (with equivalent performance) than did a form of. program control (Steinberg, 1977). However, in tasks consisting of a. complex content structure and demanding greater prerequisite knowledge, the outcomes were almost entirely in favor of program control (see Tennyson & Rothen, 1979, for a complete review) . Given the current inadequacies of learner control methods of instructional management (especially for computer-based instruction), Tennyson and his associates designed (Rothen & Tennyson, 1978; Tennyson & Park, in press; Tennyson & Rothen, 1979) and tested (Tennyson & Rothen, 1977; C. Tennyson, R. Tennyson, & Rothen, in press; Tennyson, Park, & Rothen, in press; Lau & Tennyson, Note 1; Tennyson & Buttrey, Note 2; Tennyson & Jassal, Note 3) an adaptive instructional management strategy'. This management strategy uses a Bayesian statistical method to integrate (a) assignment of a specific treatment based upon a premeasure of cognitive ability; (b) an initial amount of instructional support based upon a pretest measure of prior achievement; and (c) adjusted amount of instructional support and sequence based upon on-task learning need. The studies cited have demonstrated the effectiveness of this Bayesian probabilistic adaptive instruc-' tional strategy in selecting the appropriate amount and sequence of instruction for learning concepts. However, if a student is provided meaningful information on-task about his or her own learning development, the cognitive strategy used by the student may further refine the diagnosis and prescription made by an adaptive management system. In the present study, I propose an extension of the learner control management strategy by combining learner control with diagnostic and prescriptive information generated from the Bayesian adaptive instructional strategy. Operationally, this implies several things. First, at the start of their instruction students are advised of (a) their initial level of concept attainment compared to the desired learning Criterion (diagnosis) and (b) the amount and sequence of instruction necessary for them to obtain the objective (prescription). Second, students are continuously advised while 'on-task of their learning development (updated diagnosis) and the instructional needs (updated prescription) necessary for concept mastery. Finally., since this is a learner control management strategy, students make decisions on both amount and sequence of instruction. For the independent variable of management strategy, I hypothesized not only that the learner-adaptive control strategy . (the condition using advisement) would be as effective in student acquisition of the learning task as the adaptive strategy (that is, students in both these strategy conditions would surpass the criterion of mastery), but also that it would be more efficient in terms of student on-task learning time. I furthermore. hypothesized that both of these strategies (learner-adaptive control and adaptive control) would be more effective and efficient than the learner control strategy. Psychological experiments on inductive concept formation which manipulate stimulus symbols such as nonsense syllables, pictures, and colors have already dealt with the presentation order of examples dealing with related concepts (Brown, 1974; Bruner, Goodnow, & Austin, 1956; Dominowsky, 1974; Milward & Wickens, 1974, Roach, 1975; Sanders, DiVesta, & Gray, 1972). However, as a design strategy for actual concept teaching, the presentation order of examples for related (coordinate) concepts has not been examined until quite recently. Investigating thé learning of contextually similar rules, k. Tennyson and C. Tennyson (1975) ,showed that presenting rules simultaneously by pairing instances according to matched variable features resulted in significantly better performance than, either a successive presentation of rules or a random presentation More recently, C. Tennyson, R. Tennyson, and Rothen (in press) extended the presentation order of rule examples to the presentation order and content structure of coordinate concepts. The results of their study showed that presentation of concepts according to their coordinate relationships facilitates concept acquisition; students learn to discriminate between such concepts when given examples of each concept concurrently within rational sets. The second purpose of this study was to replicate the C. Tennyson et al. study by applying the content structure variable to another discipline and populatidn. While the C. Tennyson et al. study used four coordinate concepts from psychology, I selected coordinate concepts from physics; and while the earlier study's participants were high school students, I selected young adult college students of 19 to 21 years of age. Given the findings of R. Tennyson and C., Tennyson (1975) and C. Tennyson et al.(in press), I hypothesized that presentation of concepts according to their coordinate relationships would facilitate concept acquisition as contrasted to a design strategy that presents concepts