5.5 Cognitive theory and educational technology

Educational technology has for some time been influenced by developments in cognitive psychology. Up until now, we have focused mainly on research that has fallen outside the traditional bounds of our field. We have referred to sources in philosophy, psychology, computer science, and so on. In this section, we review the work of those who bear the title "educational technologist" who have been primarily responsible for bringing cognitive theory to our field. We are, again, of necessity selective, focusing on the applied side of our field, instructional design. We begin with some observations about what scholars consider design to be. We then examine the assumptions that underlay behavioral theory and practice at the time when instructional design became established as a discipline. We then argue that research in our field has helped the theory that designers use to make decisions about how to instruct keep up with developments in cognitive theory. However, design procedures have not evolved as they should have. We conclude with some implications about where design should go.

5.5.1 Theory, Practice, and Instructional Design

At the beginning of this chapter we noted that the discipline of educational technology hit its stride during the heyday of behaviorism. This historical fact was entirely fortuitous. Indeed, our field could have started equally well under the influence of Gestalt or of cognitive theory. However, the consequences of this coincidence have been profound and to some extent troublesome for our field. To explain why, we need to examine the nature of the relationship between theory and practice in our field. (Our argument is equally applicable to any discipline.)

The purpose of any applied field, such as educational technology, is to improve practice. The way in which theory guides that practice is through what Simon (1981) and Glaser (1976) call design. The purpose of design, seen this way, is to select the alternative from among several courses of action that will lead to the best results. Since these results may not be optimal, but the best one can expect given the state of our knowledge at any particular time, design works through a process Simon (198 1) calls satisficing.

The degree of success of our activity as instructional designers relies on two things: first, the validity of our knowledge of effective instruction in a given subject domain and, second, the reliability of our procedures for applying that knowledge. Here is an example. We are given the task of writing a computer program that teaches the formation of regular English verbs in the past tense. To simplify matters, let us assume that we know the subject matter perfectly. As subject-matter specialists, we know a procedure for accomplishing the task: Add ed to the infinitive, and double the final consonant if it is immediately preceded by a vowel. Would our instructional strategy therefore be to do nothing more than show a sentence on the computer screen that says, "Add ed to the infinitive, and double the final consonant if it is immediately preceded by a vowel"? Probably not (though such a strategy might be all that is needed for students who already understand the meanings of infinitive, vowel, and consonant). If we know something about instruction, we will probably consider a number of other strategies as well. Maybe the students would need to see examples of correct and incorrect verb forms. Maybe they would need to practice forming the past tense of a number of verbs. Maybe they would need to know how well they were doing. Maybe they would need a mechanism that explained and corrected their errors. The act of designing our instructional computer program in fact requires us to choose from among these and other strategies the ones that are most likely to "satisfice" the requirement of constructing the past tense of regular verbs.

Knowing subject matter and something about instruction are therefore not enough. We need to know how to choose among alternative instructional strategies. Reigleuth (1983) has pointed the way. He observes that the instructional theory that guides instructional designers' choices is made up of statements about relations among the conditions, methods, and outcomes of instruction. When we apply prescriptive theory, knowing instructional conditions and outcomes leads to the selection of an appropriate method. For example, an instructional prescription might consist of the statement, "To teach how to form the past tense of regular English verbs (outcome) to advanced students of English who are familiar with all relevant grammatical terms and concepts (conditions), present them with a written description of the procedure to follow (method)." All the designer needs to do is learn a large number of these prescriptions and all is well.

There are a number of difficulties with this example, however. First, instructional prescriptions rarely, if at all, consist of statements at the level of specificity as the previous one about English verbs. Any theory gains power by its generality. This means that instructional theory contains statements that have a more general applicability, such as "to teach a procedure to a student with a high level of entering knowledge, describe the procedure." Knowing only a prescription at this level of generality, the designer of the verb program needs to determine whether the outcome of instruction is indeed a procedure-it could be a concept, or a rule, or require problem solving-and whether or not the students have a high level of knowledge when they start the program.

A second difficulty arises if the designer is not a subject-matter specialist, which is often the case faced by designers. In our example, this means that the designer has to find out that "forming the past tense of English verbs" requires adding ed and doubling the consonant.

Finally, the prescription itself might not be valid. Any instructional prescription that is derived empirically, from an experiment or from observation and experience, is always a generalization from a limited set of cases. It could be that the present case is an exception to the general rule. The designer needs to establish whether or not this is so.

These three difficulties point to the requirement that instructional designers know how to perform analyses that lead to the level of specificity required by the instructional task. We all know what these are. Task analysis permits the instructional designer to identify exactly what the student must achieve in order to attain the instructional outcome. Learner analysis allows the designer to determine the most critical of the conditions under which instruction is to take place. And the classification of tasks, described by task analysis, as facts, concepts, rules, procedures, problem solving, and so on links the designer's particular case to more general prescriptive theory. Finally, if the particular case the designer is working on is an exception to the general prescription, the designer will have to experiment with a variety of potentially effective strategies in order to find the best one, in effect inventing a new instructional prescription along the way.

Even from this simple example, it is clear that, in order to be able to select the best instructional strategies, the instructional designer needs to know both instructional theory and how to do task and learner analysis, to classify learning outcomes into some theoretically sound taxonomy, and to reason about instruction in the absence of prescriptive principles. Our field, then, like any applied field, provides to its practitioners both theory and procedures through which to apply the theory. These procedures are predominantly, though not exclusively, analytical.

Embedded in any theory are sets of assumptions that are amenable to empirical verification. If the assumptions are shown to be false, then the theory must be modified or abandoned as a paradigm shift takes place (Kuhn, 1970). The effects of these basic assumptions are clearest in the physical sciences. For example, the assumption in modem physics that it is impossible for the speed of objects to exceed that of light is so basic that, if it were to be disproved, the entire edifice of physics would come tumbling down. What is equally important is that the procedures for applying theory rest on the same set of assumptions. The design of everything from cyclotrons to radio telescopes relies on the inviolability of the "light barrier."

It would seem reasonable, therefore, that both the theory and procedures of instruction should rest on the same set of assumptions and, further, that should the assumptions of instructional theory be shown to be invalid, the procedures of instructional design should be revised to accommodate the paradigm shift. In the next section, we show that this was the case when instructional design established itself within our field within the behavioral paradigm. However, we do not believe that this is the case today.

5.5.2 The Legacy of Behaviorism

The most fundamental principle of behavioral theory is that there is a predictable and reliable link between a stimulus and the response it produces in ' a student. Behavioral instructional theory therefore consists of prescriptions for what stimuli to employ if a particular response is intended (see 2.2.1.3). The instructional designer can be reasonably certain that with the right sets of instructional stimuli all manner of learning outcomes can be attained. Indeed, behavioral theories of instruction can be quite intricate (Gropper, 1983) and can account for the acquisition of quite complex behaviors. This means that a basic assumption of behavioral theories of instruction is that human behavior is predictable. The designer assumes that if an instructional strategy, made up of stimuli, has had a certain effect in the past, it will probably do so again.

The assumption that behavior is predictable also underlies the procedures that instructional designers originally developed to implement behavioral theories of instruction (Andrews & Goodson, 1981; Gagn6, Briggs & Wager 1988; Gagn6 & Dick, 1983). If behavior is predictable, then all the designer needs to do is to identify the subskills the student must master that, in aggregate, permit the intended behavior to be learned, and select the stimulus and strategy for its presentation that builds each subskill. In other words, task analysis, strategy selection, try-out, and revision also rest on the assumption that behavior is predictable. The procedural counterpart of behavioral instructional theory is therefore analytical and empirical, that is, reductionist. If behavior is predictable, then the designer can select the most effective instructional stimuli simply by following the procedures described in an instructional design model. Instructional failure is ascribed to the lack of sufficient information, which can be corrected by doing more analysis and formative testing.

5.5.3 Cognitive Theory and the Predictability of Behavior

The main theme of this chapter has been cognitive theory. We have argued that cognitive theory provides a much more complete account of human learning and behavior because it considers factors that mediate between the stimulus and the response, such as mental processes and the internal representations that they create. We have documented the ascendancy of cognitive theory and its replacement of behavioral theory as the dominant paradigm in educational psychology and technology. However, the change from behavioral to cognitive theories of learning and instruction has not been accompanied by a parallel change in the procedures of instructional design through which the theory is implemented.

You might well ask why a change in theory should be accompanied by a change in procedures for its application. The reason is that cognitive theory has essentially invalidated the basic assumption of behavioral theory, that behavior is predictable. Since the same assumption underlies the analytical, empirical, and reductionist technology of instructional design, the validity of instructional design procedures is inevitably called into question,

Cognitive theory's challenges to the predictability of behavior are numerous and have been described in detail elsewhere (Winn, 1987, 1990, 1993). The main points may be summarized as follows:

1. Instructional theory is incomplete. This point is trivial at first glance. However, it reminds us that there is not a prescription for every possible combination of instructional conditions, methods, and outcomes. In fact, instructional designers frequently have to select strategies without guidance from instructional theory. This means that there are often times when there are no prescriptions with which to predict student behavior.

2. Mediating cognitive variables differ in their nature and effect from individual to individual. There is a good chance that everyone's response to the same stimulus will be different because everyone's experiences, in relation to which the stimulus will be processed, are different. The role of individual differences in learning and their relevance to the selection of instructional strategies has been a prominent theme in cognitive theory for 2 decades (Cronbach & Snow, 1977; Snow, 1992). Individual differences make it extremely difficult to predict learning outcomes for two reasons. First, to choose effective strategies for students, it would be necessary to know far more about the student than is easily discovered. The designer would need to know the student's aptitude for learning the given knowledge or skills, the student's prior knowledge, motivation, beliefs about the likelihood of success, learning style, level of anxiety, and stage of intellectual - development. Such a prospect would prove daunting even to the most committed determinist! Second, for prescriptive theory, it would be necessary to construct an instructional prescription for every possible permutation of, say, high, low, and average levels on every factor that determines an individual difference. This obviously would render instructional theory too complex to be useful for the designer. In both the case of the individual student and of theory, the interactions among many factors make it impossible in practice to predict what the outcomes of instruction will be. One way around this problem has been to let students decide strategies for themselves. Learner control (Merrill, 1988; Tennyson & Park, 1987) is a feature of many effective computer-based instructional programs (see 33.1). However, this does not attenuate the damage to assumption of predictability. If learners choose their course through a program, it is not possible to predict the outcome.

3. Some students know how they learn best and will not necessarily use the strategy the designer selected for them. Metacognition is another important theme in cognitive theory. It is generally considered to consist of two complementary processes (Brown, Campione & Day, 1981). The first is students' ability to monitor their own progress as they learn. The second is to change strategies if they realize they are not doing well. If students do not use the strategies that instructional theory suggests are optimal for them, then it becomes impossible to predict what their behavior will be. Instructional designers are now proposing that we develop ways to take instructional metacognition into account as we do instructional design (Lowyck & Elen, 1994).

4. People do not think rationally as instructional designers would like them to. Many years ago, Collins (1978) observed that people reason "plausibly." By this he meant that they make decisions and take actions on the basis of incomplete information, hunches, and intuition. Hunt (1982) has gone so far as to claim that plausible reasoning is necessary for the evolution of thinking in our species. If we were creatures who made decisions only when all the information needed for a logical choice was available, we would never make any decisions at all and would not have developed the degree of intelligence that we have! Schon's (1983, 1987) study of decision making in the professions comes to a conclusion that is similar to Collins's. More recently, research in situated learning (Brown, Collins & Duguid, 1989; Lave & Wenger, 1991; Suchman, 1987) has demonstrated that most everyday cognition is not "planful" and is most likely to depend on what is afforded by the particular situation in which it takes place. The situated nature of cognition has led Streibel (1991) to claim that standard cognitive theory can never act as the foundational theory for instructional design. Be that as it may, if people do not reason logically, and if the way they reason depends on specific and usually unknowable contexts, their behavior is certainly unpredictable.

These and other arguments (see Csiko, 1989) are successful in their challenge to the assumption that behavior is predictable. The bulk of this chapter has described the factors that come between a stimulus and a student's response that make the latter unpredictable. Scholars working in our field have for the most part shifted to a cognitive orientation when it comes to theory. However, they have not shifted to a new position on the procedures of instructional design. Since these procedures are based, like behavioral theory, on the assumption that behavior is predictable, and since the assumption is no longer valid, the procedures whereby educational technologists apply their theory to practical problems are without foundation.

5.5.4 Cognitive Theory and Educational Technology

The evidence that educational technologists have accepted cognitive theory is prominent in the literature of our field (Gagn6 & Glaser, 1987; Richey, 1986; Spencer, 1988; Winn, 1989a). Of particular relevance to this discussion are, those who have directly addressed the implications of cognitive theory for instructional design (Bonner, 1988; Champagne, Klopfer & Gunstone, 1982; DiVesta & Richer, 1987; Schott, 1992; Tennyson & Rasch, 1988). Collectively, scholars in our field have described cognitive equivalents for all stages in instructional design procedures. Here are some examples.

Twenty years ago, Resnick (1976) described "cognitive task analysis" for mathematics. Unlike behavioral task analysis, which produces task hierarchies or sequences (Gagné, Briggs & Wager, 1988), cognitive analysis produces either descriptions of knowledge schemata that students are expected to construct, or descriptions of the steps information must go through as the student processes it, or both. Greeno's (1976, 1980) analysis of mathematical tasks illustrates the knowledge representation approach and corresponds in large part to instructional designers' use of information mapping that we discussed in section 5.3. Resnick's (1976) analysis of the way children perform subtraction exemplifies the information-processing approach.

Cognitive task analysis gives rise to cognitive objectives, counterparts to behavioral objectives. In Greeno's (1976) case, these appear as diagrammatic representations of schemata, not written statements of what students are expected to be able to do, to what criterion, and under what conditions (Mager, 1962).

The cognitive approach to learner analysis aims to provide descriptions of, students' mental models (Bonner, 1988), not descriptions of their levels of performance prior to instruction. Indeed, the whole idea of "student model" that is so important in intelligent. computer-based tutoring (Van Lehn, 1988) very often revolves around ways of capturing the ways students represent information in memory and how that information changes, not on their ability to perform tasks.

With an emphasis on knowledge schemata and the premise that learning takes place as schemata change, cognitively oriented instructional strategies are selected on the basis of their likely ability to modify schemata rather than to shape behavior. If schemata change, DiVesta and Rieber (1987) claim, students can come truly to understand what they are learning, not simply modify their behavior.

These examples show that educational technologists concerned with the application of theory to instruction have carefully thought through the implications of the shift to cognitive theory for instructional design. Yet in almost all instances, no one has questioned the procedures that we follow. We do cognitive task analysis, describe students' schemata and mental models, write cognitive objectives, and prescribe cognitive instructional strategies. But the fact that we do task and learner analysis, write objectives, and prescribe strategies has not changed. The performance of these procedures still assumes that behavior is predictable, a cognitive approach to instructional theory notwithstanding. Clearly something is amiss.

5.5.5 Can Instructional Design Remain an Independent Activity?

We are at the point where our acceptance of the assumptions of cognitive theory forces us to rethink the procedures we use to apply it through instructional design. The key to what is necessary lies in a second assumption that follows from the assumption of the predictability of behavior. That assumption is that the design of instruction is an activity that can proceed independent of the implementation of instruction. If behavior is predictable and if instructional theory contains valid prescriptions, then it should be possible to perform analysis, select strategies, try them out, and revise them until a predetermined standard is reached, and then deliver the instructional package to those who will use it, with the safe expectation that it will work as intended. If, as we have demonstrated, that assumption is not tenable, we must also question the independence of design from the implementation of instruction (Winn, 1990).

There are a number of indications that educational technologists are thinking along these lines. All conform loosely with the idea that decision making about learning strategies must occur during instruction rather than ahead of time. In their details, these points of view range from the philosophical argument that thought and action cannot be separated, and therefore the conceptualization and doing of instruction must occur simultaneously (Nunan, 1983; Schon, 1987), to more practical considerations of how to construct learning environments that are adaptive, in real time, to student actions (Merrill, 1992). Another way of looking at this is to argue that, if learning is indeed situated in a context (for arguments on this issue, see McLellan, 1996), then instructional design must be situated in that context, too.

A key concept in this approach is the difference between learning environments and instructional programs. Other chapters in this volume address the matter of media research. Suffice it to say here that the most significant development in our field that occurred between Clark's (1983) argument that media do not make a difference to what and how students learn and Kozma's (1991) revision of this argument was the development of software that could create rich multimedia environments. Kozma (1994) makes the point that interactive and adaptive environments can be used by students to help them think, an idea that has a lot in common with Salomon's (1979) notion of media as "tools for thought." The kind of instructional program that drew much of Clark's (1985) disapproval was didactic--designed to do what teachers do when they teach towards a predefined goal. What interactive multimedia systems do is allow students a great deal of freedom to learn in their own way rather than in the way the designer prescribes. Zucchermaglio (1993) refers to them as "empty technologies" that, like shells, can be filled with anything the student or teacher wishes. By contrast, "full technologies" comprise programs whose content and strategy are predetermined, as is the case with computer-based instruction (see 12.2.3).

We believe that the implementation of cognitive principles in the procedures of educational technology requires a reintegration of the design and execution of instruction. This is best achieved when we develop stimulating learning environments whose function is not entirely prescribed but which can adapt in real time to student needs and proclivities. This does not necessarily require that the environments be "intelligent" (although at one time that seemed to be an attractive proposition [Winn, 19871). It requires, rather, that the system be responsive to the student's intelligence in such a way that the best ways for the student to learn are determined, as it were, "on the fly."

5.5.6 The Three "Ages" of Scholarship in Educational Technology

We summarize the main points in this section by describing the three ages of educational technology. We call these the age of instructional design, the age of message design, and the age of environment design.

The age of instructional design is dominated by behavioral theories of learning and instruction and by procedures for applying theory to practice that are based ultimately on the assumption that behavior is predictable. The decisions instructional designers make are driven almost exclusively by the nature of the content students are to master. Thus, task analysis, which directs itself to an analysis of content, dominates the sources of information from which strategy selection is made. The most important criterion for the success of the techniques used during the age of instructional design is whether or not they produce instruction that is as successful as a teacher. Clark's (1983) criticism of research in our field is leveled at instructional systems that attempt to meet this criterion.

In the age of message design, the emphasis shifts from instructional content to instructional formats. We believe that this is the immediate result of the concern among cognitive theorists with the way information is represented in memory, schemata, and mental models. There is an assumption (doubtless incorrect; see Salomon, 1979) that the format selected to present information to students in some way determines the way in which the information is encoded in memory. A less-restrictive form of this assumption has, however, produced a great deal of useful research about the relationship between message forms and cognition. Fleming and Levie (1993) provide an excellent summary of this work.

The age of environment design is likewise based on cognitive theory. However, its emphasis is on providing information from which students can construct understanding for themselves through interaction that is more or less constrained, depending on students' Deeds and wishes. The key to success in this third, current, age is in the interaction between student and environment rather than in content or information format. A good example of this orientation in instructional design is Merrill's (1992) transaction theory, where the instructional designer's main focus in prescribing instruction is the kind of transaction (interaction) that occurs between the student and the instructional program. Another example is the design of learning environments based in the technologies of virtual reality (Winn, 1993). In virtual environments, the interaction with the environment is potentially so intuitive as to be entirely transparent to the user (Bricken, 1991). However, just what the participant in a virtual environment is empowered to do and particularly the way in which the environment reacts to participant actions (Winn & Bricken, 1992) requires the utmost care and attention from the instructional designer.

5.5.7 Section Summary

In this section we have reviewed a number of important issues concerning the importance of cognitive theory to what educational technologists actually do, namely, design instruction. This has led us to consider the relations between theory and the procedures employed to apply it in practical ways. We observed that when behaviorism was the dominant paradigm in our field, both the theory and the procedures for its application adhered to the same basic assumption, namely, that human behavior is predictable. We then noted that our field was effective in subscribing to the tenets of cognitive theory, but that the procedures for applying that theory remained unchanged and continued to subscribe to the by-now discredited assumption that behavior is predictable. We concluded by suggesting that cognitive theory requires of our design procedures that we create learning environments in which learning strategies are not entirely predetermined, which requires that the environments be highly adaptive to student actions. Recent technologies that permit the development of virtual environments offer the best possibility for realizing this kind of learning environment.