themediamademecrazy.com

Abstract: Cognitive load has been described extensively in educational literature. However, many multimedia learning tools fail to design programs that limit cognitive load to improve learning. This is a randomized, controlled, within-media, comparative study of interface design. An existing multimedia learning tool for plastic surgery served as a template to create two interfaces, which taught the pre procedural processes of carpal tunnel release. The experimental interface sought to limit audiovisual redundancy by altering the visual presentation of the text. Medical students that interacted with the experimental program demonstrated improved learning.

Background

Over the past decade, advances in computer-based education have led various medical schools to invest resources in new multimedia learning tools to train their students. There is a large body of information comparing multimedia-teaching methods to traditional teaching methods (e.g. lecture, lab sessions, etc.), which confirms that multimedia education is at least as effective as the standard approach. Unfortunately, there is no further discussion as to which design techniques are most effective. This experiment seeks to establish what design techniques are most successful in creating multimedia learning tools by combining educational strategies with theories of cognitive load.

When developing or evaluating a multimedia learning tool, it is necessary to make the distinction between the medium, which delivers the message, and the symbolic systems used to convey the message (Clark 2003), a point often neglected by program designers. Therefore, the efficacy of these design techniques depend both on the cognitive response to the interface, but also the method by which the interface constructs the material. Merely translating a preexisting curriculum to a multimedia format is insufficient; rather, new curriculums and lesson plans must be designed for multimedia environments.

An extensive review of literature pertaining to educational strategies, cognitive load and interface design revealed that the optimal interface for cognitive efficiency employed a constructivist lesson plan (gradually integrating discreet packets of information into the learner’s long term memory) and accounted for the learner’s cognitive capacities (Grunwald 2006). Existing multimedia learning tools fail to address cognitive load when anticipating the learner’s interaction with the program; they suffer from overloading the user (i.e. presenting more than nine items at a time) and redundancy (i.e. repeating items without reinforcing information).

Figure 1: Dual Channel Assumption (Mayer 2001)

Cognitive load theory is based on information processing assumptions, including Mayer’s ‘dual channel assumption,’ which states that humans possess separate information processing channels for visual and auditory material (text information is interpreted through the auditory pathway), and the ‘limited capacity assumption,’ which limits the amount of information can that be consumed by either channel at one time (See Figure 1) (Mayer 2001). Combined, the human working memory is capable of holding an average of seven items at once (Miller 1956). The number of elements that are consciously or unconsciously attended to by the learner is necessary in calculating the total cognitive load of an interface. The student must integrate the discreet bundles of information into a limitless long-term memory and the expanse of this long-term memory determines his or her proficiency with a given subject (Grunwald 2006). The following guidelines assist in creating an interface that reduces cognitive load.

1. Synchronize audio and visual information
2. Eliminate multi-tasking
3. Optimize representations/Approachable interface
4. Maintain stable learning environment
5. Eliminate redundant information
6. Manage navigational control
7. Maintain authentic context

Despite available studies, many designers and medical schools disregard multimedia learning strategies for flashy graphics and promises of infinite information. A literature review investigating computer aided instruction (CAI) revealed that 60% of CAI articles were demonstrations studies (studies which demonstrate the efficacy of a multimedia learning tool), 13% were analyses of the CAI field, and only 11% were media comparative studies that compared two multimedia interfaces for efficiency(Adler 2000). Of these, many did not address the intricate differences in design techniques, focusing instead on drastically different types of media intervention. Although copious resources have been allocated for interactive educational software, very few studies have addressed the efficacy of interface design. This preliminary study specifically addresses two of the abovementioned guidelines: (1) the synchronization of audio and visual information, (2) eliminating redundant information (Grunwald 2006). Some studies investigated the effect of an interactive supplement on standard teaching methods, but a comparison between different programs has yet to be conducted.

Tools and Methods

This randomized controlled trial quantitatively assesses the guidelines set forth in the aforementioned literature review and compares the efficiency of various interface designs. The experimental interface employs both the visual and audio pathways for optimal reception. Instead of repeating the text via the audio and visual pathways, we relegated the text to the audio pathway alone, leaving the visual pathway to receive the same information in a difference context. This reinforces the lesson instead of merely repeating it thereby reducing redundant aspects of the interface.

Two dynamic interfaces were designed specifically for this study that teach the preliminary processes of carpal tunnel release. They vary in presentation in order to cater to the students’ cognitive and educational needs. The control interface provides a simplistic multimedia interpretation of the material, similar to that of an existing multimedia textbook: the text is offered via visual and audio pathways supplemented by images and hyperlinks. The experimental interface rephrases the visual delivery of the text into an outline form that is revealed in synch with the audio reading (See Figure 2). Both interfaces offer a variety of optional learning resources to assist the student including a hyperlinked glossary and expandable images.

Figure 2: S.MA.R.T. Control and Experimental Interfaces – Control Interface (left) presents the total script and the audio reading of the script simultaneously. Experimental Interface (right) presents a script outline synchronized with the audio reading of the script. Red text links to a supplemental glossary with additional information and images.

Subjects included 40 medical students from the Keck Medical School at USC. Each subject interacted with the program for up to fifteen minutes, after which they completed a cognitive assessment. Subjects were observed during this time to discover how they interact with the interface and what learning resources they choose to utilize.

The assessment scores were compared to determine which interface was most “cognitively efficient” for the learner. Demographics were correlated with performance as well as other factors including time spent with the program and learning resources utilized. Students learning through advanced multimedia interfaces that account for theories of cognitive load should achieve greater comprehension of the lesson as compared to students learning through interfaces that do not.

Results and Conclusion

Forty students were enrolled in the study. Twenty-one were randomized to the control interface and 19 to the experimental interface. The mean score of the entire group was 68.5%. The control group had a mean score of 62.9% and the experimental group had a mean score of 74.7% (p value 0.007).

The assessment tested a series of items presented in the program; answers were divided according audio, visual, or integrated (both audio and visual) sources. Since the two interfaces differed according to visual presentation, not audio, the comparative data regarding visual questions further supports the thesis. The visual mean score of the entire group was 72.8%. The control group had a visual mean score of 67.9% and the experimental group had a visual mean score of 77.6% (p value 0.083). Alternatively, questions whose answers were provided via ado/text received a group mean of 75%; the control group had a textual mean score of 66.7% and the experimental group had a textual mean score of 83.3% (p value 0.016). This marked improvement in the visual oriented performance over audio is attributed to the decreased cognitive load of the interface design.

The experimental interface limited redundancy by altering the presentation of the audio information. In the control interface, the script was presented to the user concurrently with its audio reading via headphones, thereby overlapping cognitive items without contextual difference. In the experimental interface, the script was reconfigured into an outline format that focused the learner’s attention on key terms and ideas. The outline was also revealed in synch with the audio reading to improve learning. Students learn better when corresponding information is presented simultaneously in space and time (Mayer 2001).

Both versions of the program included rollover identification functions and a supplemental glossary that provided additional information and images. We observed subjects as they interacted with the program and documented the frequency with which they used the available resources. Each of these resources received a wide range of usage, but lacked any correlation to performance. In addition, we found no statistical correlation according to a Spearman Correlation Test between time spend with the program (which ranged from 147-603 seconds) and overall performance, regardless of group.

Discussion

In this study we studied one specific aspect of interface and program design associated with cognitive load. This type of “within media-comparative” study design reduces confounding factors associated with media-comparative studies. Many multimedia learning tools ignore educational theories when designing programs. The redundancy of seeing and hearing the same words increases the learner’s cognitive load, thereby decreasing the amount of information that may be assimilated into long term memory. The ultimate goal of the S.M.A.R.T. Project is to design a fully comprehensive multimedia educational environment that can reduce the time necessary to acquire and perfect a skill as well as the monetary and temporal demands on the teaching hospital and its faculty.

References

Adler, M., & Johnson, K. (2000). Quantifying the Literature of Computer-Aided Instruction in Medical Education. Academic Medicine, 75 (10), 1025-8

Clark R. (2003). Research on Web-Based Learning: A Half-full Glass. In: Bruning R, Horn, C. & PytlikZillig, eds. Web-Based Learning: What do we know? Where do we go?. Greenwich, CT: Information Age Publishing

Grunwald, T., Corsbie-Massay, C. (2006). Guidelines for Cognitively Efficient Multimedia Learning Tools: A review of literature relating to educational strategies, cognitive load theory, and interface design. Academic Medicine, 81 (3), 213-23

Mayer R. (2001). Multimedia Learning. New York, NY: Cambridge University Press

Miller, G.A. (1956). The Magical Number Seven, Plus or Minus Two: Some Limits on our Capacity for Processing Information. Psychology Review, 63, 81-97

Acknowledgements

This project was funded by the USC School for Cinema-Television and the Institute for Multimedia Literacy at USC.