12.1 EVOLUTION OF COMPUTER-BASED INSTRUCTION: HISTORICAL PERSPECTIVES

While the genesis of computers has been traced to the theoretical work of mathematicians such as Pascal, and events such as the processing of immigrants at Ellis Island, early hardware developments and applications were both extremely costly and predictably narrow in focus. The size, space, and costs associated with computers generally limited initial application to fields where resources were available and needs were significant. Ile computer's information management capabilities made it ideal for large-scale organization and retrieval of information, complex computations, and the like. Apart from simple management and analysis, comparatively little attention was paid to the role of computers in learning,

Once sufficient interest in computers and learning was generated, a host of related problems was encountered. Apart from cost, size, and space barriers, virtually no educational software was available. Available programniing languages were, for the most part, designed to support data manipulation and business applications, rendering them cumbersome and limited in their educational utility. Computer hardware was, by contemporary yardsticks, extremely primitive, requiring batch-mode cards or electronic typewriters for input, remote printers (or, in later cases, low-resolution monochrome monitors) for output, and entire rooms to house the basic central processing unit. T'he decision to test the power of the computer for learning was no small commitment. It required that applications be developed from the ground up, including not only sound lesson design but also the hardware and software environments to support their creation.

Not surprisingly, then, early efforts were limited both in terms of pedagogy and usability. The criticism of early lessons as little more than electronic page-turners was, for the most part, well founded. Typical computerized lessons were often significantly poorer than alternatives in terms of their ability to represent images, provide direct access to various parts of a lesson, highlight and otherwise emphasize important information, and so on. Early efforts to replace existing media and teaching methods were singularly unsuccessful.

Initial research, particularly the Stanford Arithmetic Project in the mid-1960s, helped to legitimize the scientific study of computer-assisted learning. Attempts were made to define variables operationally, isolate learning effects, and otherwise control extraneous factors (within the limitations of the available mathematics software). The systems were primitive in current terms but reflected state-of-theart, both with respect to the hardware available and the pedagogy enabled.

Clearly, a great deal has changed since the heyday of the Stanford Arithmetic Project. The technologies have changed dramatically both in size and form, and the underlying psychological and pedagogical underpinnings have evolved in many unforeseen ways. Still, these efforts were the genesis of present-day research in computer-based learning. Their contributions were due not to their similarity with contemporary approaches but to the commitment to disciplined inquiry. The debate shifted from a purely intellectual activity to one that could be advanced empirically.

A pivotal point in the early evolution of computers and learning came with the emphasis on human-computer interactions, heretofore awkward or impossible with the media and methods of the day. Spurred by the growing influence of behaviorists such as B. F. Skinner, few questioned the importance of overt responding and differential reinforcement to learning. The computer, through which learners could interact with lessons on an individual basis, was seen as a significant technological breakthrough to overcoming many problems associated with programmed instruction (see 2.3.4 and 19.2.1 for related discussions). Predictably, early lessons were influenced heavily by behavioral psychology: relatively small and discrete steps, overt responses, feedback, and differential branching in the form of drills and tutorials. Individual knowledge, in the form of stimulusresponse-reinforcement (S->R->SR) associations, could be engineered in the form of progressively linked chains (see 2.2). Many large-scale research and development efforts, including the University of Illinois PLATO project, as well as institutes at Florida State University and Penn State University, generated courseware on a significant scale based largely on these roots. For the most part, traditional experimental research paradigms dominated.

Beginning in the late 1960s, cognitive perspectives gained increased acceptance. Research and development shifted toward methods and strategies that induced desired cognitive processes. Strategies were designed to assist the leamer in selecting appropriate information, organizing information into internally consistent concepts, and integrating new with existing knowledge, making it personally relevant and meaningful (see 5.3 through 5.5, 20.1). As alternative psychological and pedagogical perspectives emerged and hardware and software capabilities expanded, both the nature of computer-based learning systems and their underlying foundations metamorphosed. Contemporary research and development thrusts are now evident in areas such as user-centered learning, open-ended learning, and hypermedia (e.g., Horwitz & Barowy, 1994).

The evolution of computers in learning has not occurred in isolation. Rather, it is an interactive byproduct of ongoing developments in psychology, pedagogy, and technology (Atkins, 1993; Hannafin, 1992). At times, technology has been at the forefront of these developments, helping to shape our understanding by demonstrating heretofore untested teaching and learning methods. At other times, technology has enabled widely held notions about teaching and learning, perhaps "increasing the horsepower" of conventional methods. Whether leading or following, technology has assumed a prominent role.

To understand research, one must understand more than the myriad of often-conflicting findings resulting from idiosyncratic approaches employed in individual studies. To understand the research on computers, emerging technologies, and learning, one must attempt to assemble and organize, not merely to collect; to interpret rather than to simply report; to relate across studies and specific technologies, not merely to isolate findings. These are the goals of this chapter.