CYBERPSYCHOLOGY & BEHAVIOR

Volume 3, Number 1, 2000Mary Ann Liebert, Inc.

Designing a Better Learning Environment with the Web:Problems and Prospects

ANDREW DILLON, Ph.D.

ABSTRACT

The virtual university looms inevitably on the horizon, and Web-based education requireshypermedia technologies to be designed and implemented in a learner-centered fashion. Todate, the technology has not been well designed and the evidence for learning advantages isvirtually nonexistent. In the present article, two claims about web-based educational tech-nologies are made and defended: that the technology is poorly designed from a user’s per-spective, and that the theories of instructional design that we require to build better learningenvironments do not exist. The field of educational technology is thus presented as both theculprit and the victim of these shortcomings. Suggestions for improving this scenario are pre-sented.

97

INTRODUCTION

IN A RECENT REVIEW of the empirical findingson hypermedia and learning outcomes, Dil-

lon and Gabbard1 concluded that contrary tomany people’s assumptions, the use of hyper-media-based instructional systems in educa-tion had not produced significant learninggains. Indeed, their review concluded that suchinstructional technologies rarely showed anybenefit for learners over existing paper- or lec-ture-based instructions. Although it is com-monplace these days to dismiss as irrelevantany media comparison study, the Dillon andGabbard review went further, also examiningcomparisons made between alternative hy-permedia implementations (a within-mediacomparison) and between single and grouplearners employing this technology. Becausehypermedia is the underlying technology of theWorld Wide Web, their findings made de-pressing reading for those of us who believe

that this technology is important and could beput to powerful instructional use.

The present issue contains papers from manyleading theorists who advocate the use and ex-ploitation of information technologies such ashypermedia and the World-Wide Web in ourclassrooms, and I am not completely in dis-agreement with them. However, I wish to ques-tion the very assumptions on which the use ofthe Web and standalone hypermedia applica-tions are based. What I aim to provide in thisarticle is a sense of the gaps in our knowledge,and to speculate on why education is so poorlyserved by the wonderful technologies that arewithin our grasp.

Two major claims are made: (1) the applica-tion of educational technology is poorly servedby the field’s weak analysis and limited un-derstanding of the basic human engineering is-sues of usability and user-centered design; and(2) the theoretical analysis of learning is un-necessarily partitioned into competing grand

HIC Laboratory, SLIS, Indiana University, Bloomington, IN.

views, or paradigms, that place less emphasison empirical analysis of learning and more onreifying the field of education.

This is no Luddite treatise. It is a call to arms;a request for proponents to challenge assump-tions about learning technologies, to test theclaims made for them, to design instructionaltechnologies in a learner-centered manner.

WHY THE WORLD WIDE WEB?

On one hand it is easy to get caught up inthe hype of the Web. After all, it is a marvel ofinformation access, worldwide connectivity,and even, in some rare cases, ease-of-use. Therehas never been an information technologyquite like it, and it is tempting to dismiss thenay-sayers as ignorant stick-in-the-muds,clinging to outdated views of computers, teach-ing, and learning. But a closer look at the Weband variant hypermedia-based technologiesthat aim to support education through cyber-space throws up some interesting questionsthat require more serious consideration.

One such question is why the Web shouldoffer us anything special for instruction? Afterall, since when did information access and presentation equate to instruction? We can in-corporate more information for learners to ex-plore, navigable by link structures, and this cer-tainly can enable greater and faster access tomore information than most educational li-braries provide, but is this enough to ensurelearning? We can communicate over a distance,allowing isolated individuals to form groups,to contact experts, and to share information inreal-time and off-line. Again, this could all bedone by older technologies but that is not mymajor concern because the Web is surely a moreattractive, powerful, and responsive environ-ment for these activities. My concern is the rel-evance of such capabilities to learning.

In my view, while these features are certainlymajor drivers toward acceptance of Web tech-nology by a large user population, none of theabove advantages are in themselves major dri-vers of learning outcome. This is crucial be-cause, if true, it seems that the lure of the Webhas little or anything to do with learning. If wetry to understand why educationalists are so

enamoured of links and nodes then we need tolook elsewhere. Consider, for example, the typeof claim made for Web technology in the learn-ing context. It is seen as a breakthrough in cognitive compatibility, more natural, morelearner-centric, more adaptable, and more flex-ible than any previous technology (see Dillon2

for a review of the claims). Each of these claimsought to be examined critically by education-alists but too often they are assumed, taken forgranted as obvious or self-evident truths. Evi-dence in their favor is normally anecdotal,rarely, if ever, derived from well-controlledstudies. After more than 50 years of existenceas a field of enquiry, this is a poor state of af-fairs for educational technology.

Unintended consequences of technologies al-ways follow any implementation (e.g., the easeof communication afforded by E-mail has theeffect of swamping some experts with innu-merable requests, frequently trivial ones thatcould be answered with existing resources).Technologies are frequently designed with lit-tle consideration of usability, yet it is good hu-man factors design that at least ensures the pri-mary advantages of technology (mentionedabove) can be exploited. Perhaps even moreworrying, there is little consideration of themeasurable learning advantages that might ac-crue from Web use, greater emphasis beingplaced on preference than performance, de-spite evidence that these two variables maycontradict each other in interactive tasks.3

Does the Web have a place in education? Ofcourse it does but to gain that place requires ef-fort to be spent on design, testing, and refine-ment of resources that are targeted appropri-ately at the learning contexts most likely tobenefit. You cannot develop a reasoned case orargument by merely stringing words together,yet it appears that just offering information viaweb links is all that is needed to make an on-line learning environment. Just as the thesisneeds to refine, build, and sequence ideas, us-ing words and figures as bricks in a planned,meaningful fashion to construct a case, so doesa Web environment require planning, organi-zation, and revision in order to convey struc-ture and offer instruction. Web-based instruc-tion has a long way to go before it can compete.Any headlong rush to embrace the technology

DILLON98

as a panacea for instructional design is flawedand fanatical.

BUILDING FOR USE—THE HUMANFACTORS OF INTERACTION

The Web is here and it will not go away. Sohow can we best exploit it? In the first instance,any interactive technology must be designedfor usability by which I mean we must considerthe effectiveness, efficiency, and satisfactionwith which the intended users can employ thetechnology. To do this we must design forlearning contexts, meaning we must analyzeour learners, their instructors, the tasks to besupported and the physical and social envi-ronment in which the learning is to occur.Twenty years of research into human–com-puter interaction (HCI) has shown that withoutsuch analyses, the chances of ever developinga technology that is accepted by users andyields the desired benefits are significantly re-duced. While I suspect few instructional tech-nologists would disagree with the paragraphabove, I contend that too few of them will havestudied the human factors findings on interac-tion deeply enough to learn the lessons it mightimpart. It surprises many to find that there isan extensive empirical literature on usabilityand usefulness that explains, for example, whypeople read up to 30% slower on screen, whyusers experience significantly greater disorien-tation with hypermedia than with paper, thatusers tend to imbue human characteristics tosoftware, or what determines long-term useracceptance of new technologies. This literatureis not largely the work of instructional tech-nologists, but its relevance to their goals is un-deniable. Too often it seems, learning tech-nologies are designed without any properaddressing of the factors that are important tousable and useful design and implementation.Some times these technologies are even de-signed in contradiction to such factors. This ar-ticle is not the appropriate vehicle for a reviewof that literature, but the interested readerwould do well to consult such resources as He-lander et al.4 or Dillon5 for basic reviews.

An interaction analysis makes us focus firstand foremost on the activities a learner will per-

form and I believe this is a key reason for thefailure of HCI findings to break into the in-structional domain where current conceptionsof learning are attempting to move away fromtraditional instructional models predicated onnarrow tasks or prescribed task sequences. In-deed, there is at least a superficial similarity be-tween the HCI approach to design analysis andthe behavioral objectives tradition of 1960s in-structional design. However, I would suggestthere is at least one major difference. HCI doesnot equate performance in using the technol-ogy with learning. Rather it is a necessary butinsufficient component of the learning sce-nario. Learning is a complex, multiply deter-mined activity or process that cannot just beequated with a single task such as informationretrieval, target location, navigation, or memo-rization alone. Yet part of learning involves theactive engagement of those tasks—and if thetechnology interferes, slows down or otherwiseraises usability obstacles in the path of learn-ers, then this will negatively affect education.

In this sense, learning, as a goal, needs to beaddressed at a level where aspects of informa-tion location, summarization of ideas, linking,and memory are components. Such a compo-nential analysis is essential to the appropriatetechnological support of the learner. In so designing the technological support (hypert-extual, Web-based, or otherwise), empiricalmethods can be employed to determine us-ability, to rectify design problems and to assessthe extent to which the technology increaseslearner performance on that component (al-though this is usually a different issue fromwhether or not learning was being supported).The complete instructional analysis then takesaccount of the necessary performance compo-nents or tasks to be included and sequenced inthe learning environment so as to ensure learn-ing goals are attained. In short, interactionanalysis becomes a part of instructional analy-sis.

This is by no means a complete recipe forsuccess, but it is a tight form of design ratio-nale that yields demonstrable benefits and en-sures that re-design can be maximally targetedto produce the greatest effects. Many instruc-tional technologies would be improved byadopting this style of design at some point in

WEB DESIGN 99

their development to avoid the situation wherea fine instructional system is significantly ham-pered by poor implementation of the user in-terface. What is required to put this approachto use is not only an understanding of the interaction literature and findings, but also astrong theoretical analysis of learning thatwould guide the earliest conceptions of designand encourage new ideas for the technologicalsupport of education—which leads me to mysecond claim.

THE COMPETITION OF EDUCATION THEORIES

Clearly the problem of designing successfulweb-based learning environments rests onmore than simply applying HCI principles.Theoretical insights into learning are crucial toimproving our designs and understandinghow we can advance education. Indeed, as Ihave argued elsewhere, there can never be apurely empirical approach to design—any ar-tifact must be its nature embody the theoreti-cal assumptions of its designers, whether thesebe consciously and clearly articulated or not.5

What is required of any theory to be helpful inthis domain is its applicability to design in amanner that can drive development.

Space precludes a detailed exploration of theapplication of educational theory to design, butit is necessary to point out that the major con-cerns of educational theorists, as evinced in thetheoretical literature of the field, are less withthe problems of explaining the findings oflearning outcome with various media andmore with offering broad conceptual frame-works of learning. Much educational theoriz-ing is given over to proposing one or otherbroad model of learning or countering the cri-tiques of other broad model proponents. To anoutsider, the rather heated debates betweenconstructivists and behaviorists are bewilder-ing, seeming more concerned with attractingadherents to a cause rather than to enlighten-ing our analyses of education. Furthermore,both points of view seem to attack stereotypednotions of learning and cognition that are out-dated and long past interest to most cognitiveand social scientists.

As mentioned above, the emphasis on un-derstanding the context within which learningoccurs is crucial to good technology design.Such an emphasis precludes any reliance on anoverarching theoretical position that seeks torepresent the workings of the human mind asa mass of stimulus-response arcs, a semanticnetwork, or an attuned rhizome. Invoking alearning scenario serves to ground the interac-tion literature. It also ensures that generaliza-tion of findings is bounded to the context of useexamined.

What is really lacking in our work is a meansof plugging the gaps in our knowledge, aframework for inference, and potential gener-alization of findings beyond the exact contextof measurement. This is the type of role theoryplays in a mature science. Even if we refuse toacknowledge education as a science, there cansurely be no argument that useful theorywould advance our understanding of the tech-nologies we create for instruction. Yet manycurrent theoretical positions in education seemfar removed from these practical concerns,leaving technology designers to guess or intuitthe appropriate means of implementing alearning tool. Worse than that, certain theoret-ical positions are so loosely articulated thatthey enable designers to invoke theoretical sup-port independent of its demonstrable applica-tion.

It is not just the debate over constructivism,cognitivism or behaviorism that limits us. Thelatest edition of the award-winning Handbookof Research for Educational Communications andTechnology6 provides separate treatments of behaviorism, systems theory, communicationtheory, cognition, sociology, constructivism,ecological psychology, critical theory and post-modernism. Each makes a case for the value ofits theoretical position for education, outliningits origins, basic assumptions and constructs,while only limited attempts are made to linkthese to specific technological issues. Evenwhere technology is invoked, the examples arefrequently narrative in nature, not empirical orprescriptive. Little effort is made to bridgethese theoretical divides either, to show howthe importance of context is shared by manytheories or to make explicit the idea that alearner must be both a cognitive and social be-

DILLON100

ing (how could a human ever be otherwise?).In a nutshell, the theories of the field place em-phasis on too many high level claims about theultimate nature of learning and the humanmind, and provide too little analysis of howtheory might guide instructional technologydesign. We need to ground ourselves as a fieldand to emphasize data over opinion, and to usetheory as a strong adjunct to our investigationsand technology designs (an excellent attemptat this in the work of Carroll7). Only then willthe field advance.

THE WAY FORWARD?

Will we one day have universities withoutwalls, virtual classrooms, and distributed net-works of scholars and students? Almost cer-tainly. To paraphrase Ted Nelson’s argumentsfor hypermedia—the question is not if or when,but how can we make them great. I’ve pointedout that while we seem to be on an inevitablecourse toward greater use of technology, weare limping along, not progressing in leaps andbounds. While I believe we can take greaterstrides if we address some basic interaction is-sues immediately, there remain worrying the-oretical shortcomings to our designs that war-rant our serious consideration.

If we are content as a field to follow tech-nologists and their inventions, hoping to even-tually refine our new environments into shape,then progress will be slow, and generations oflearners may have to suffer our mistakes enroute. Even then, I fear we will always look to-ward new technologies to solve old problems.There has to be a better way—and that way de-mands theoretical analysis at the contextual

and performance levels. Without this, the vir-tual university will be the actual business thatmany of us currently in education fear physi-cal universities are in danger of becoming: thehome of industry-dictated courses and pur-chasable degrees. There is a better way.

REFERENCES

1. Dillon, A. & Gabbard, R. (1998). Hypermedia as an ed-ucational technology: A review of the quantitative re-search literature on learner comprehension, control,and style. Review of Educational Research, 68(3):322–349.

2. Dillon, A. (1996) Myths, misconceptions and an alter-native perspective on information usage and the elec-tronic medium. In: Rouet, J.-F. et al. (eds.) Hypertext andCognition. Mahaw, NJ: Lawrence Earlbaum Associates.

3. Bailey, G. (1993). Performance vs. Preference. In Pro-ceedings of the Human Factors Society Annual Conference.New York: HFES, pp. 316–320.

4. Helander, M., Landauer, T., Prabhu, P. (eds.). (1997).Handbook of human-computer interaction, 2nd edition.Amsterdam: Elsevier Science Ltd.

5. Dillon, A. (1995). Artifacts as theories: Convergencethrough user-centered design. In: Proceedings of theAmerican Society for Information Science Annual Confer-ence. Online document: http://www.slis.indiana.edu/adillon/adillon-artifacts.html

6. Jonassen, D. (ed.) (1996). Handbook of research on educa-tional communications and technology. New York:Macmillan.

7. Carroll, J. (ed.). (1998). Minimalism beyond the NurnbergFunnel. Cambridge: MIT Press.

Address reprint requests to:Andrew DillonHCI Laboratory

SLIS10th and Jordan

Bloomington, IN 47405

E-mail: adillon@indiana.edu

WEB DESIGN 101