Brent G. Wilson

University of Colorado at Denver

April 2004

For inclusion in J. M. Spector, D. Wiley, C. Ohrazda, & A. Van Schaack (Eds.), Innovations in Instructional Technology: Essays in Honor of M. David Merrill. Mahwah NJ: Erlbaum, 2004.

Three or four years ago, David Merrill told me he wrote his now-classic manifesto Reclaiming Instructional Design (Merrill, Drake, Lacy, Pratt, & the ID₂ Research Group, 1996) in direct response to a lunchtime conversation we had enjoyed together at an AERA meeting. By his report, I had been looking for another name to call myself—perhaps the "instructional" part of instructional design felt too limiting at the time. Merrill and colleagues describe a science of instruction and a technology of instructional design (ID).

While I had earlier written a paper partly in response (Wilson, 1997), I had no idea that Merrill’s spirited charge had been largely targeted, at least originally, at me. I do not consider myself anti-science, but I believe ID need not be narrowly framed in strictly scientific terms. In this chapter, I continue the conversation about ID’s foundations by exploring some ways to think more liberally about its practice. This contribution is not a point-by-point response to Merrill’s paper, but covers much of the same ground from a different perspective. My hope is to work toward a conception of ID practice that is inclusive of different perspectives while preserving a coherent identity through shared concerns and purposes.

The practice of instructional design has been defined as a science, a technology, a craft, and even an art (e.g., Davies, 1991; Glaser, 1965; Lumsdaine, 1964; Melton, 1959; Merrill et al., 1996; Reigeluth, Bunderson, & Merrill, 1978; Skinner, 1954; for additional definitions see Ryder, 2003a). What instructional designers do, in essence, is design instruction and related resources that meet learning needs for defined audiences and settings. This includes tasks of management, implementation, and evaluation—all in the service of designing and delivering good instruction.

From the earliest days of Thorndike (1913) and Skinner (1954) to Gagné (1962) and Glaser (1964), instructional design has been offered in direct response to challenging problems of practice facing the education and training professions. Thus, instructional design can be best understood within a context of reform, that is, working to improve how instruction gets designed and delivered, applying the best available principles of learning and instruction so that learners do not have to suffer through the excruciatingly bad practices that define the period. While to some extent ID theories reflect that period and its practices, they are also intended to be a stimulus to reform and improved practice.

Instructional design’s origins, even present practices, relate closely to the fields of instructional technology and educational psychology (Dick, 1978, 1987; Reiser, 1987, 2001). ID has grown up, however, without a single enduring professional home. The only professional organization I am aware of that includes the specific words ‘instructional design’ is PIDT (Professors of Instructional Design and Technology), a group of about 100 professors who have met annually with their doctoral students for nearly two decades. Other American-based organizations for instructional designers include AECT (Association for Educational Communications and Technology) and ISPI (International Society for Performance Improvement). In 1977, AECT and ISPI formed a joint task force to establish professional competencies for instructional designers, leading in 1984 to the independent incorporation of the International Board of Standards for Training, Performance and Instruction (IBSTPI for short; see Richey, Fields, & Foxon, 2001 for current ID competencies). Both AECT and ISPI have played important roles in the historical development of instructional design (Reiser & Ely, 1997; Rosenberg, 1982; Rosenberg, Coscarelli, & Hutchison, 1992; see also Dean & Ripley, 1998).

Through professional organizations, conferences, journals, books, Web sites, and various other tools, the current ID community helps designers be more effective in their work by:

• promoting practice that meets a professional standard, including the education of new designers in graduate programs;
• developing promising and useful models, theories, and procedures, based on experience, direct research, or often derived from related disciplines;
• advancing usable knowledge through the sharing of professional experience, codified into stories, cases, listserv exchanges, and conference presentations; and
• further advancing knowledge by refereed reports of systematic inquiry around models, theories, procedures, and practices, to test their validity, effectiveness, and range of use.

Two core ideas permeate thinking and theorizing in instructional design: conditions-of-learning theories of instruction, and a systems approach to instruction and instructional development. These central ideas affect how ID gets done in real life. Each is discussed, in the following sections.

Conditions-of-Learning Theories

In the days of programmed instruction, researchers held to a few general principles of learning, based on behavioral psychology, which were thought to apply universally to all settings and organisms. Results of programmed-instruction research, however, showed that some strategies worked better than others, depending on conditions. This led Lumsdaine (1964) and others to articulate a vision for an emerging science of instruction: through factorial experiments, instructional scientists would develop a sophisticated series of rules, sub-rules, and meta-rules for employing instructional strategies to teach different kinds of content in different settings. This idea of a rule set which links conditions, instructional methods, and learning outcomes was promoted and refined by other theorists such as Gagné (1965), Bruner (1966), and Reigeluth (1983) as a defining feature of instructional theories. These prescriptive theories could be considered technologies in their function as tools for designers of lessons and courses.

The publication in 1965 of Robert Gagné’s Conditions of Learning was a seminal event in the history of instructional design. Since then, ID theorists have taken for granted the conditional nature of design prescriptions offered for various kinds of learning outcomes. The conditions-of-learning framework continues to guide theory development and professional practice (Ragan & Smith, 1996). However, a few questions, presented below, still remain concerning the nature of prescriptive theories of instruction.

What is their ontological status? Typically, scientific theories include principles describing how the world is—the way things are. These are formulated in a particular way to allow explanations or precise understandings about mechanisms, dynamics, processes, and so on. Theories of instruction contain a similar descriptive element. For example, Gagné (1965) presents a typology of different learning outcomes, ordered from simple to complex. Attainment of these learning outcomes may be observed in the world. However, the heart of instructional theory is not in the description of outcomes, but rather in the prescriptive linking between outcomes and related conditions. Here the ontological status of the rules becomes less clear. We are moving beyond descriptions to guidelines, or rules for action. The link to traditional science is often quite indirect (Simon, 1996), but the link to observed practices is equally obscure (Wedman & Tessmer, 1993). In other words, prescriptive theories of instruction lie in a space between descriptions of the world and direct guidelines for practitioners.

How are they validated? Descriptive claims about the world can be tested by systematic observation. Prescriptive rules of design are not so easily or directly validated. Design prescriptions may be grounded directly in a scientific theory, as is the case with Ausubel’s (1963) strategy of advance organizers and its link to his theory of meaningful learning; or the design rules may be tested themselves (providing advance organizers and experimentally observing their effects). Challenges occur when testing prescriptions due to the many forms and contexts of practice. If an advance organizer fails to teach, it may be that the particular organizer was poorly written or that it was not appropriate for the particular situation in which it was used. Maybe it was not a proper advance organizer after all. Many contingencies exist in practice, making direct empirical validation of general instructional design prescriptions very difficult.

Because of this fuzzy link between the idealized design prescription and what really happens in practice, we may need to look beyond the traditional method-conditions-outcomes model (Reigeluth, 1983) to account for learning outcomes. That is to say, it is what happens during instruction that mediates learning, not what theory was applied or what design heuristic was supposedly followed. The details of interaction and activity—the experienced instruction rather than the designed instruction—are so often what make the difference between good and bad instruction (see Feyerabend , 1975; Palmer, 1999; Ryder, 2003b).

Are they for people or machines? People routinely look for guides to action in the form of conceptual models, rules of thumb, mnemonics, recipes, and so on. Often, guidelines for action can be directly designed into a tool or automated machine, as in the affordances of a saw or hammer, a tax form, or an online interactive program. Some instructional prescriptions seem clearly intended for designers to keep in mind as they approach design problems, such as using the rule-example-practice model for tutorial design or following Gagné’s (1985) nine events of instruction in lesson design. The prescription is simple enough to hold in working memory and serves as a heuristic or template for developing a lesson. Other formulations of instructional theory, however, are so technically defined and presented that intentions for use are unclear. Diagrams can be overburdened with hard-to-read detail; jargon can be far removed from normal discourse. While such technical models may serve as blueprints or specifications that could be programmed into automated instruction, their use by practicing designers on everyday projects is problematic.

In spite of these concerns, heuristics, or guidelines for design, are essential. The challenge is how to formulate them in a way that is useful in practice. This requires that design prescriptions be sensitive to the complexity of real-life conditions of use.

Systems Thinking

The second core idea, thinking systemically about instruction—seeing learners, teachers, content, and so on as components in a larger system—has a surprisingly long history in education (Banathy, 1968; Finn, 1956; Lumsdaine, 1960; Merrill, 1968). By combining behavioral principles of learning, information-processing principles of message and content, and systems principles of interactivity and interdependence, early instructional theorists established a frame for viewing instruction as a system that could be designed, measured, and optimized. Instruction was seen as more than what the teacher did, but rather as the complex interaction between participants within a larger systemic context. The discourse could also be quite progressive, with advocacy for learner control, conversation, and open systems (e.g., Merrill, 1975; Pask, 1975; Silber, 1972; Winn, 1975). In the last decade, systems theory has enjoyed a resurgence due largely to advances in complexity theory and self-organizing systems (Kelly, 1994; Senge, 1990), with continuing implications for instructional design (e.g., Carr, 1997; Land, 2000; Milrad, Spector, & Davidsen, 2002; Spector, 2000; Spector & Anderson, 2000; Wilson & Ryder, 1996).

A systems model of instructional development (ISD) has been used and taught for more than forty years among instructional designers (Gustafson & Branch, 1997). Many think of ISD as being a core model defining the field. However, I am not aware of a solid body of research empirically demonstrating its advantage over other curriculum-development models (see Hannafin, 1983a, 1983b, and Sullivan, Ice, & Niedermeyer, 2000, as examples of work in this direction). Of course, scientific testing of comprehensive procedures is very difficult, but I also take the lack of research as a sign of the model’s axiomatic status within the field. ISD’s enduring value lies, I believe, in its embodiment of rational-planning principles and in its cohering role in defining the instructional technology community. The ISD model provides practical value partly by providing management controls, such as decomposing a complex process into parts with deadlines, reviews, and signoffs. Another major strength is that ISD models ensure a logical consistency between learning goals, activities, and assessments. This is hard to argue with, unless somehow in the process, an intangible part of instruction—a valued goal, experience, or outcome—is lost as things are documented and codified (Bunderson, Gibbons, Olsen, & Kearsley, 1981). This problem seems avoidable if designers are careful to leave room for unanalyzed or unidentified elements in holistic learning experiences. Instruction that is highly compartmentalized and controlled may be more vulnerable to reductive loss because it relies more on analyzed and predefined content.

Alternatives to ISD

After many years of relying almost exclusively on ISD models, more theorists are acknowledging alternatives to the objectives-based efficiency model of curriculum (cf. Kliebard, 1987, for alternatives to Tyler’s (1949) objectives-driven curriculum model). One team of theorists (Nieveen & Gustafson, 1999; van den Akker, Branch, Gustafson, Nieveen & Plomp, 1999; Visscher-Voerman, Gustafson, & Plomp, 1999) has called attention to four different "paradigms" (or perhaps aspects) of instructional development:

• Instrumental - linear procedures based on needs and objectives established at the outset; proceeding rationally through development and implementation.
• Pragmatic - continual evaluation and feedback mechanisms within the process, ensuring useful and effective instructional outcomes.
• Communicative - including stakeholders in establishing needs and goals; consensus-building and inclusion of diverse perspectives throughout the development process.
• Artistic – such as Eisner’s (1979) connoisseurship model of curriculum, relying on the judgment of professionals (teachers) through a process of criticism and reflection.

In Table 1, I have collapsed these four approaches into two, while trying to maintain the gist of the original work.

Table 1

Of course, the two paradigms need not be positions in strict opposition to one another. Most practicing instructional designers identify professionally with the ISD paradigm above, but I would guess many also employ consensus-building strategies in much of their work. Curriculum developers outside the ID community often start from consensus-building assumptions, then apply ISD principles to really get the work done once a level of consensus has been achieved.

The emphasis of ISD is focused on efficient and reliable development of products. Consensual approaches do a better job acknowledging the multiple interests, values, and goals in establishing curriculum and on their inclusion in the process. It seems understandable, then, that an ISD emphasis would tend to flourish in training environments where learning goals are often defined in technical terms, whereas a consensual emphasis may work better in non-technical forms of education, including management education and much K-12 education, where experts and constituencies often disagree. As public education adopts a more technology-centered stance toward curriculum and assessment, with high-stakes testing aligned with standards-based instruction, ISD might be expected to gain in influence (Wilson, 2002).

When seen juxtaposed with other paradigms, the implicit values attending ISD become more apparent. Drawing on earlier work of Nunan (1983), Rose (2002) challenges ISD’s pretense as an objective, value-neutral process:

Unannounced values often accompanying ID practice are carried through an array of questionable assumptions, including:

• Experts—if they really are experts—should agree with each other.
• Getting content out of subject experts is an unproblematic extraction procedure.
• Designers can design excellent instruction in content areas they really don’t know well.
• Knowledge can be explicitly documented, codified, and transferred through explicit lessons and materials, in the form of clearly defined rules, concepts, and procedures.
• People process information similarly across situations and content areas.
• Novices need high levels of structure and guidance; more experienced learners can be trusted with self-directed activities.
• People of all skill levels benefit from essentially the same well-designed instruction.
• Learners’ special needs and cultural backgrounds are nice to know but not essential in designing instruction that works.
• Instructional materials mean what they say; there is no hidden curriculum.
• Instruction not developed systematically is not to be trusted.
• Teachers and instructors are not to be trusted to adapt or modify lessons and strategies.

These implicit values and assumptions are expressed in stark terms that thoughtful ISD followers would disavow. Even so, I believe that daily practice often proceeds with some combination of these values operational—all the more reason to adopt a more critical stance, to avoid negative values and assumptions that might accrue to the process.

Of course, rather than choosing paradigms, designers may choose to keep in mind the values and concerns of both technical and consensual approaches, and even modify procedures to include both. I participate in a research team that has developed a four-level scheme for doing instructional design that combines ideas from social science and the humanities, whose approaches to design can be very different (IDEAL Lab, 2003; see Table 2). The four-level scheme outlined below could fit within either ISD or consensual approaches to instructional development.

Individual cognition and behavior. The first level of concern is the individual learner’s thinking and acquisition of knowledge and skill. The designer considers how lessons and strategies fit with existing theories of learning and cognition. Questions for design include: Do learning activities help the learner acquire new conceptual understanding and procedural skills in an efficient and effective way? Is the cognitive load kept manageable?

Social and cultural learning. The next level turns to issues of cultural context and social support, including peer-to-peer interactions; group identity and motivation; and participation within communities of learning and practice. Questions for design include: Are learners given opportunities for social interaction, collegial support, and inclusion in meaningful practices? Are learners supported as they come to see themselves (and relate to others) in new ways?

Values. Critical theorists focus on questions of justice, asking questions of instruction like: In addition to the stated objectives, what is this lesson really saying? What is not said, that reveals something about the values of the people involved? Where is the power? Who is the lesson designed for, and who is left out? How, and upon whom, is status granted? What kinds of practices does the lesson encourage or perpetuate? What alignment is there between stated institutional values and observed instructional materials? Questions like these are not usually built into ISD models, although they should come up in consensual development processes.

Aesthetics. Finally, through an aesthetic layer of analysis, designers ask questions about the shape and form of the learning experience, as well as the design of messages within the experience. Aesthetic considerations might lead to offering learners an adventure by adopting a dramatic journey metaphor to a curriculum, or to a certain pattern of introducing, heightening, and resolving tensions within a lesson. Questions for design: Is the learning experience satisfying? Is it cathartic? Is there sufficient tension and movement to hold learners’ interest? Is the timing and pacing designed to help learners identify with and participate in the experience?

Table 2

The first two levels come from the social sciences—psychology and anthropology in particular. Indeed, most research in instructional design and technology relies heavily on this social science disciplinary base. The latter two levels are less common; in critical-theory terms, they are less privileged. Like so many other fields, ID places more value on science than on other perspectives, even though, ironically, science is not in a position to define value. Our research team believes the issues raised in the bottom two levels should be integrated into established design practices; in other words, we believe the technical ISD procedures can be revised to include a broader range of concerns when establishing needs, standards, and criteria for success. The four levels reflect more than an expanded view of learning outcomes (for example, Anderson & Krathwohl, 2001). We are suggesting an expanded view of design processes themselves and criteria for evaluating designed products.

Today most academic programs in instructional technology rely, at least implicitly, on a common knowledge base of instructional theories as a curriculum foundation, along with a systems metaphor for instruction and a systems model for developing instructional materials. True to these general models, most departments serve multiple settings; that is, graduates may work in school settings, higher education, business, industry, government, and so on. They may apply ID principles in the classroom, workplace, or online. Regardless of medium or setting, the principles of instructional design are still thought to be generally applicable.

The forces pulling against this general approach to ID practice are formidable, however. For a generation, instructional design has been more positively received in training settings than K-12 schools, leading to higher status for adult-learning settings such as universities and business workplaces. Many academic programs divide students into different tracks according to work setting—e.g., K-12 schools versus adult learning settings. While a challenge to coherence, this seems more consistent with theories of situated learning, which would favor setting-specific practices over general theories.

The changing landscape of ideas is another threat to coherence. The number of competing educational paradigms continues to grow. Most of these paradigms claim to be scientific, yet their notions of science differ dramatically. Designers must choose now from among an increasing array of theories of learning and instruction. Foundations of ID practice, conditions-of-learning and systems thinking, have been examined, critiqued, and deconstructed. While these foundations have proven impressively resilient, the coherence afforded to an earlier generation seems unavailable to current practitioners and theorists.

Current theorists tend to think about professional knowledge in new ways. Rather than a vast rule base of contingent generalizations, professional knowledge is now seen as more pluralistic: explicit textbook knowledge complements the tacit knowledge held by practitioners, comprised of skills and understandings developed in the context of everyday practice and shared informally within workgroups and professional meetings. Moreover, the clear emergence of competing paradigms, co-existing together in time, argues for an eclectic, opportunistic stance toward the various theories and models available in the literature. Instead of a single huge rule set, researchers and practitioners alike prefer to think of a somewhat messy toolbox from which a particular model or theory or technique may be chosen, according to the demands of the situation. As Rorty (2000) observes, academic discourse is often a matter of re-description rather than logical argument from shared premises of fixed meanings. Designers can try out lenses of different models or paradigms, and re-describe a given situation many times. This ability to see things through multiple perspectives can serve designers well as they try to fit a number of elements together into a working system, but it comes at a cost to coherence.

We observe, then, some contradiction in our professional identities. Hopes for an integrative foundation—in learning theories, systems theories, learning technologies, and prescriptive principles of instruction—seem at odds with the proliferation of competing perspectives, each carrying a different set of tools, terms, models. Each perspective serves a somewhat distinct professional sub-community, discernible by the cliquish clustering of reference lists at the ends of articles. We want to identify with a coherent professional community, but finding the ties that bind can be an elusive task.

As a response to the threat of fragmentation, I return to the enduring aims of instructional design presented at the outset. The field of instructional design is largely defined by the challenges we choose to tackle, which at the most general level are: how to design and deliver good learning experiences for learners in a variety of contexts and, secondarily, how to best use various technologies in the service of that learning. The practical problem is the mediocre quality of instruction. The response is instructional design. As Richey (1998) notes, agreeing on the details of formulating a problem requires some degree of shared ideology, but that is precisely where we agree—on the general nature and importance of these problems. Then and from that base, competing theories and perspectives enter the dialogue. As researchers and practitioners grapple with problems of practice, they are led to countering explanations and theories, leading to re-descriptions of problems and proposed solutions. As so many have argued, the interplay between theory and practice is a dialogue, which is the healthiest possible condition for a field, even in the face of proliferating perspectives. Cutting short that dialogue would be a mistake. Keeping our eye on the end goal, improving instruction, should be enough to hold us together as a community of professionals.

Ausubel, D. (1963). The psychology of meaningful verbal learning. New York: Grune & Stratton.