Levels of Expertise and Instructional Design

Article excerpt


When scientific or technical information is presented to learners, it usually is conveyed via multiple sources. Frequently, instruction involves a combination of textual and diagrammatic information. How the text and diagrams are designed and related can have substantial effects on the material's learnability, and the ideal instructional design may interact with the learner's level of expertise. This paper investigates relations between instructional design and levels of expertise. We begin with a discussion of aspects of cognitive architecture that are relevant to instructional design.

Processing of information occurs within a limited working memory (Baddeley, 1992; Miller, 1956). Only a few items or elements of information can be handled in working memory at one time. From the perspective of instructional design, it is this restriction that provides the major limitation of working memory. Too many elements can overwhelm working memory, decreasing the effectiveness of instruction.

In contrast to working memory, long-term memory can hold an unlimited number of elements in the form of hierarchically organized schemas (Chi, Glaser, & Rees, 1982; Larkin, McDermott, Simon, & Simon, 1980). Schema is defined as a cognitive construct that permits people to treat multiple subelements of information as a single element, categorized according to the manner in which it will be used. For example, everyone has acquired schemas for lines and angles and, through further learning, has acquired higher-order schemas that combine these lines and angles into geometric shapes such as squares. The schema for a square is stored in long-term memory and, despite consisting of multiple subelements, can be transferred to working memory as a single element to be processed. From this analysis it can be seen that schemas have a dual function: storing learned information in long-term memory and reducing the burden on working memory by allowing multiple elements of information to be treated as a single element.

Automation similarly reduces working memory load (Kotovsky, Hayes, & Simon, 1985; Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977). Information can be processed either consciously or automatically, and conscious processing requires more working memory resources than does automatic processing. Schemas are stored in long-term memory with varying degrees of automaticity. A schema can be stored and retrieved from long-term memory either in fully automated form or in a form that requires conscious consideration of each of the elements and their relations. If a schema can be brought into working memory in automated form, it will make limited demands on working memory resources, leaving more resources available to search for a possible problem solution, for example. If a fully automated schema incorporating the problem solution is available in long-term memory for transfer to working memory, solution will proceed easily and smoothly.

Cognitive load theory, which incorporates this architecture, has been used to design a variety of instructional procedures (see Sweller, 1994, for a recent summary) based on the assumptions that working memory is limited and that skilled performance is driven by automated schemas held in long-term memory. The theory assumes that information presented to learners and the activities required of them should be structured to eliminate any avoidable load on working memory and to maximize the acquisition of automated schemas. Two techniques are relevant to the experiments reported in this paper.

The split-attention effect. Multiple sources of information are often directed to learners who find one or more of the sources unintelligible in isolation and who can achieve understanding only by mentally integrating the various sources of information. A geometric proof consisting of a diagram and associated statements is an example: Neither the diagram nor the statement is likely to be intelligible unless the two are mentally integrated. …


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.