Table of Contents

Volume 4, Number 2/3 1995

Articles

Temporal Issues in the Design of Virtual Learning Environmentsby Byran Bergeron and Jihad Obeid

From Multimedia Instruction to Multimedia Evaluationby Martha E. Crosby and Jan Stelovsky

Configurational Validity: A Proposal for Analyzing Ethnographic Multimedia Narrativesby Ricki Goldman-Segall

Cognitive Media Types for Multimedia Information Accessby Mimi M. Recker, Ashwin Ram, Terry Shikano, George Li and JohnStasko

An Adaptive Presentation Model for Hypermedia Information Systemsby Amir Hekmatpour

Cognitive Mapping Architectures and Hypermedia Disorientation: An Empirical Studyby Robert Beasley and Michael L. Waugh

Connectionist Modeling as the Basis for Multimedia Clinical Patient Simulations With Diagnostic Capabilitiesby Bryan P. Bergeron, Luke Sato, Jihad Obeid and Ronald Rouse


Abstracts


Temporal Issues in the Design of Virtual LearningEnvironments

Bryan Bergeron
Research Affiliate, Harvard Medical School and MIT
Harvard Medical School and MIT, USA

Jihad Obeid
Decision Systems Group of Harvard Medical School and MIT
Brigham & Women's Hospital
75 Francis Street, Boston, MA 02115

Computer-based virtual environments, including multimedia clinicalpatient case simulations, are uniquely suited to modulating contentscope, difficulty level, and presentation rate. However, it isthe perceived manner and rate of temporal progression and eventduration that directly influences realism and educational effectiveness.Unlike other presentation modalities, computer-based educationalapplications provide a means of changing the perception of timethrough automatic manipulation of the user interface, contentdifficulty, technical complexity, and other areas that can bedefined by the application author. In addition, a complex setof subjective variables, including the user's attention span,interest in the subject matter, and a variety of personal factors,can distort the intended perception of time. Although subjectivevariability cannot be avoided, temporal cues can be used to provideusers with an idea of the present moment, the rate and mannerof time passage, the time remaining, whether time is consideredan absolute or relative quantity, a historical perspective onwhat has occurred. Authors should be aware of the user's perceptionof time and how this perception can be used to support the educationalprocess.

Return to Table of Contents

From Multimedia Instruction to Multimedia Evaluation

Martha E. Crosby and Jan Stelovsky
Department of Information and Computer Sciences
University of Hawaii at Manoa
Keller Hall 319, 2565 The Mall, Honolulu, HI, 96822
The advantages of multimedia courseware are often consideredto be self-evident. Nevertheless, we need to establish which mediaare most effective and, more importantly, what type of studentsbenefit most. To explore different aspects of multimedia in learningenvironments, presentations of typical computer science algorithms,such as a binary search and simple sorting methods (bubble, insertion,selection sort) were constructed. This courseware visualizes thedynamic behavior of the algorithms using multimedia techniquessuch as static and animated text, graphics and nonlinear structure.

To contrast the effectiveness of this courseware with traditionalinstruction, two lab sections were taught alternately, one sectionusing the courseware and the other section using static (viewgraph and blackboard) presentations. Comprehension gain scoreswere analyzed with respect to the subjects' cognitive style (sensingor intuitive) and media format of tests (textual or graphical).Subjects performed better when instructed using the multimediacourseware. More importantly, the subjects' cognitive style andthe media format of the test questions influenced their performance.In particular, this study indicates that multimedia dramaticallyimproves the performance of subjects categorized as "sensing"(concrete) and that graphical questions match multimedia instructionbetter than traditional textual tests.

Return to Table of Contents

Configurational Validity: A Proposalfor Analyzing Ethnographic Multimedia Narratives

Ricki Goldman-Segall
Faculty of Education,University of British Columbia
2125 Main Mall, Vancouver, BC V6T 124 Canada

Multimedia ethnographic researchers along with those persons theystudy have an opportunity to become a culture of inquiry whenusing emerging computational tools for data analysis. As partners,this community can build narratives which convey messages to andwith their readers. As readers contribute their own interpretations,the original story changes. Layers build; stories change; patternsemerge; and inquiry becomes a reflexive practice. The questionto be asked in this emerging world of telecommunications collaborativeinquiry is: What is the story and whose story is it? This articlewill present a theory of validity for building more robust analysesof multimedia stories that can be built by distributed communitiesof inquiry. I propose that stories of multiple "authors"can be layered in clusters, or "constellations," insuch a way that larger, more robust theories emerge. I term thisapproach a platform for multiloguing. To explain the theoreticalbasis of this approach and a software application that has beendesigned to support this approach, I first explain the need fora theory of configurational validity. Then, I describe a toolcalled ConstellationsTM that has been designed to address howrobust configurations could be used to build more valid accounts.

Return to Table of Contents

Cognitive Media Types for Multimedia InformationAccess

Mimi M. Recker
Victoria University of Wellington, P.O Box 600
Wellington, New Zealand

Ashwin Ram, Terry Shikano, George Li, John Stasko
College of Computing
Georgia Institute of Technology
Atlanta, GA, 30332-0280, USA

Multimedia repositories, libraries, and databases offer the potentialfor providing students with access to a wide variety of interconnectedinformation resources. However, in order to realize this potential,multimedia systems should provide access to information and activitiesthat support effective knowledge construction and learning bystudents. This article proposes a theoretical framework for organizinginformation and activities in educational hypermedia systems.We show that such systems should not be characterized primarilyin terms of the kinds of physical media types that can be accessed;instead, the important aspect is the content that can be representedwithin a physical media, rather than the physical media itself.We propose a theory of "cognitive media types" basedon the inferential and learning processes of human users. Thetheory highlights specific media characteristics that facilitatespecific problem solving actions, which in turn are enabled byspecific kinds of physical media. We present an implemented computersystem, called AlgoNet, that supports hypermedia information accessand constructive learning activities for self-paced learning incomputer and engineering disciplines. Extensive empirical evaluationswith undergraduate students suggest that self-paced interactivelearning environments, coupled with multimedia information accessand constructive activities organized into cognitive media types,can support and help students develop deep intuitions about importantconcepts in a given domain.

Return to Table of Contents

An Adaptive Presentation Model for HypermediaInformation Systems

Amir Hekmatpour
IBM Manufacturing Technology Center
B21/975-1, zip B84, Essex Junction, Vermont 05452

Authoring tools provide utilities for creating multimedia andhypermedia applications but they neither provide nor enforce efficientpresentation models or methodologies. The basic design issuesthat make the quality and usability of information possible areleft up to the developer. We present a three-facet adaptive modelfor online hypermedia information systems. The model providesspatial reconfiguration of the presentation objects based on activationfrequency, information content customization based on author specifiedcriteria, and hypergraphic network optimization based on usage.This paper discusses the theory underlying the model, its architectureand hypermedia structure, and finally provides a preliminary quantitativemeasure and evaluation of its impact on authoring time, as wellas study time. The results reported here, although limited inscope, show positive improvements in delivery and disseminationof technical subject matters and provide a better understandingof the required development efforts and resources.

Return to Table of Contents

Cognitive Mapping Architectures and HypermediaDisorientation: An Empirical Study

Robert E. Beasley
Departarment of Business
Union College, Barboursville, KY 40906, USA

Michael L. Waugh
Department of Curriculum and Instruction
University of Illinois at Urbana-Champaign, Urbana, IL 61801

An important hypermedia design issue is concerned with how theinformation in these types of systems should be organized andhow that information should be accessed in order to decrease disorientationin the learner. Decreasing disorientation is an important issuebecause, as disorientation increases, learning is likely to suffer(Beasley, 1994; Tripp & Roby, 1990). The purpose of this studywas to assess the relative effects of hotwords and two distinctlydifferent cognitive mapping architectures (spider maps and hierarchicalmaps) on learner disorientation in a hierarchically organizedhypermedia environment. To answer the research questions posedin this study, a Completely Randomized Factorial-322 experimentaldesign was utilized. Sixty-one subjects participated in the experiment.The results of this study suggest that the inclusion of a properlyconstructed cognitive map can indeed diminish feelings of disorientationin the learner. In this investigation, learners in the "hierarchicalmaps" treatment reported feeling significantly less disorientedthan learners in the "hotwords" treatment. No significantdifferences in perceived disorientation were found, however, betweenlearners in the "hotwords" treatment and learners inthe "spider maps" treatment.

Return to Table of Contents

Connectionist Modeling as the Basis for MultimediaClinical Patient Simulations With Diagnostic Capabilities

Bryan P. Bergeron, Luke Sato, Jihad Obeid, and Ronald L. Rouse
Decision Systems Group of Harvard Medical School and MIT.
Brigham & Women's Hospital
75 Francis Street, Boston, MA 02115

Multimedia patient simulations have become a popular vehicle forboth assessing and teaching clinical skills. However, in the creationof multimedia clinical patient simulations, the effort and timerequired of clinical experts for program development, validation,and maintenance are often major obstacles to timely and affordabledevelopment. To maximize the ease with which expert medical knowledgecan be captured for the purpose of creating realistic patientsimulations, we have developed a system that relies on neuralnetworks to control conventional simulations. In addition to simplifyingthe clinical expert's task of validating and maintaining patientsimulations, the underlying neural network can serve as the basisof a clinical decision support tools that complement the patientsimulations.

Return to Table of Contents