Design decisions are not made based solely on a given foundation, but upon presumed processing requirements, the strategies and methods deemed reasonable in supporting those processes, and the manner in which technology options support or hinder combinations of learning strategies and cognitive processes (Park & Hannafin, 1993, p. 67).

Among important variables are teacher-student interactions, methods, learner traits, and motivation. Based on our review of the literature, finding a multiple-channel research article that addressed more than one of these variables was an exception. At the beginning of this chapter, we highlighted the information-processing model, its impact on research, and the implications research results have on instructional design.

To briefly recap, the information-processing model hypothesizes several information storage areas governed by processes that convert stimuli to information. The goal for instructional designers is to take advantage of suggestions from multiplechannel research in order to facilitate cognitive processes particularly in the development of multimedia presentations.

Our review has focused on the effectiveness that multiplechannel communications, cue summation, and related areas such as multi-image and subliminal perception research may play in learning situations. Unfortunately, most literature addressing these issues is conflicting and/or dated. Not once did we encounter research that thoroughly investigated these theories in the context of hypermedia or multimedia. In addition, much of the research reported is based on the well-documented limitations of media comparison studies. We also feel that the literature dealing with multiple-channel communications and cue summation should provide a portion of the foundation from which to design learning environments in the multimedia arena. Based on the review of pertinent research that are the antecedents of the concept of inultimedia--e.g., multiple-channel, cue summation, multi-image, and subliminal perception-what did we find? We feel that instructional designers, looking for simple rationale methods or guidelines for effective multimedia (multiple-channel) presentations will be disappointed in the relevant research. While much of the evidence from the research studies appears to support multiple-channel design, the overall evidence on the effectiveness of single-channel versus multiplechannel presentations is confusing at best. The human information-processing system appears to function as a multiple-channel system until the system capacity overloads. When the system capacity is reached, the processing system seems to revert to a single-channel system. In other words, a fixed cognitive capacity limits the absolute amount of information that the individual can "handle." Adding information channels does not enlarge the system rather it distributes the capacity across the two input channels. Conflicting research results are also present concerning the use of redundant information presented across two or more channels. People apparently view highly redundant information presented over two or more channels as components of a single message. Research on the cue summation and stimulus generalization theories have produced opposing results (no surprise). However, there appears to be some evidence to suggest that multiple-channel presentations are superior to single-channel presentations when cues are summed across channels, but when neither channel is superior or when content is redundant or irrelevant across channels. Redundancy may cause information processing to fluctuate and become less efficient. There also may be failure to take into account the theory about processing capacity in human beings. It is suggested that designers sometime do not understand the possibility that in multiple-channel communication irrelevant cues in either channel can cause interference. Research on multi-image presentations suggest that the mere presentation of simultaneous images does not necessarily lead to simultaneous mental processing. Like the other research in this area, multi-image research has revealed few usable results. The familiar problem of how much information an individual processes at any one time is also raised by multi-image presentations and with studies on subliminal perception. Inconclusive results leaves us with no definite evidence as to subliminal perception's effectiveness or ineffectiveness. However, there appears to be evidence that there is human Perception below the threshold of awareness. Where does this leave us in relationship to multimedia? For one thing, educators appear unable to determine a universal definition for the concept of multimedia. Secondly, there is little research concerning the design and value of multimedia systems. Certainly, the use of the research and theoretical antecedents of multimedia reviewed in this paper (e.g., multiple-channel communication and cue summation theory) have not for the most part made it into the research literature on multimedia. Most of the literature appears to deal with their adoption and implementation or the visions of their Potential use. Some of the evaluative studies available, however, tend to support the use of such presentations.

There is a rather obvious lesson to be learned in reviewing the literature in this area and, we suspect, many of the areas that this handbook is meant to deal with. Theory based research, such as that grounded in dual-coding theory, cue summation theory, etc., add up,, over time. Research comparing media against media (see 4.3.4.2), which we have characterized as evaluations, does not. As Clark (1983) readily acknowledges, such studies were criticized long before he put forth his delivery truck metaphor. This metaphor does not seem counterintuitive or, for that matter, controversial.. We invite you to look up the term media in any dictionary. It will say vehicle, as in television or radio, or words to that effect. The concept, though blindingly simple, is still misunderstood. Evaluating media against media in terms of learning outcomes (as in film versus television,) has not helped us. Testing media attributes (e.g., text and audio) against another doesn't help us either (see 4.4.4.6). What the argument doesn't have is a theory that explains what happens from a human learning/memory point of view. Clark and others suggest that there are "deeper processes" at work in learning and that the various media attributes employed are surrogates for those processes that can be cued or accessed in many ways. Simply put, that learning may be unaffected by a particular media and that learning of any type can be achieved through a variety of paths (media) if the methods of providing information are well designed, have a theoretical base and are well executed. If research in multimedia does not move quickly from evaluation to theory-based research, we will not only repeat the mistakes of the past, we also, as a discipline, will be made redundant by those working in human computer interface (HCI) and industrial systems engineering (ISE), who are grounding their work in theory.