This blog post is a slightly altered essay that I recently submitted as part of my Master of Education (Educational Psychology) studies.

The delivery of learning via the use of multimedia (also known as eLearning) continues to be a popular choice in workplaces around the world. In my current role as an Instructional Designer, I am involved in the design and development of eLearning courses for corporate clients with the aim of closing knowledge gaps in employees and ultimately improving their performance. The challenge for myself, and indeed all instructional designers, is how to best present information in such a way that it does not overload the known limits of working memory and allows for transfer into long-term memory. This essay will discuss Mayer’s Cognitive Theory of Multimedia Learning and the resulting principles for designing multimedia instruction that when applied can improve learning delivered via eLearning. The term multimedia presentation is commonly defined as “any presentation containing words (such as narration or on-screen text) and graphics (such as illustrations, photos, animation or video)”.

Cognitive Theory of Multimedia Learning

Much research has been conducted into multimedia instruction (using multimedia presentations) with the goal of finding ways to foster meaningful learning or a “deep understanding of the material”. Drawing on the work of Paivio’s Dual Coding Theory, Baddeley’s Working Memory Model and Sweller’s Cognitive Load Theory, Mayer developed a theory specifically for multimedia learning. The Cognitive Theory of Multimedia Learning is based on three assumptions about how the human mind works during multimedia instruction:

1. Dual channel – humans possess separate information processing channels for verbal and visual material.
2. Limited capacity – the verbal and visual channels can only process a limited amount of information.
3. Active processing – learning requires substantial cognitive processing via the verbal and visual channels.

According to the theory, multimedia presentations containing words (text or spoken) and pictures are received by the learner via their sensory memory (ears and eyes). Selected words and images are then processed by the working memory at the shallow and potentially deep levels which is then integrated with prior knowledge held in long-term memory.
Multimedia presentations have the potential to overload the limited capacity of the learner’s working memory in many ways. Mayer and Moreno also highlight three types of cognitive demands placed on learners during multimedia instruction:

1. Essential processing – required to make sense of the presented words and images.
2. Incidental processing – required to process additional words and or images not related to processing the presented material.
3. Representational holding – the cognitive processing required to hold visual or verbal material in working memory for a period of time.

Whilst essential processing is linked to the material being learned, incidental processing and representational holding place additional demands on working memory capacity and can interfere with the learner’s ability to process and transfer the material to their long-term memory. In addition, Clark and Mayer use the term extraneous processing to describe the processing of material not related to the learning goal or caused by poor instructional layout.

The cognitive architecture underpinning Mayer’s theory has led to (via numerous studies) the creation of principles for the design on multimedia instruction. These principles create guidelines for instructional designers wanting to create multimedia instruction based on what is known about the processing capabilities of the human memory. Several of the principles will now be discussed and applied to a sample multimedia lesson that I created with the aim of teaching people about a coffee machine.

Multimedia Principle

According to the multimedia principle, learning is improved when words and pictures are used in multimedia presentations rather than words alone. Research into the multimedia principle by Clark and Mayer, examined lessons that taught scientific and mechanical processes, for example, how lightning is formed or how a bicycle pump works and found that “students who received multimedia lessons consisting of words and pictures performed better that those who received the same information via words alone”.

Figure 1

Figure 1 shows a screen from a sample eLearning module that contains only text. If learning is to be improved, applying the multimedia principle would require the addition of an image, an example of which can be seen in Figure 2:

Figure 2

The type of treatment of a screen of information in Figure 2 is often seen in eLearning courses, however the image is not directly related to the material being presented and as such does not support the instruction. Clark and Lyons refer to these types of images as decorative and are usually added to make content appear more ‘interesting’. However, they do not contribute to learning and actually require processing by cognitive resources that are already limited and therefore is an example of incidental processing. Another example of incidental processing would be the inclusion of background music playing as the content is delivered. In this example, an improvement would be via the inclusion of an image of the coffee machine that the text is referring to as seen in Figure 3 (additionally, a further improvement could be made by indicating which part of the machine each section of the text refers to).

Figure 3

Contiguity Principle

In multimedia presentations, words (either text or spoken) and pictures are sometimes used to describe how a process or piece of equipment works or to deliver content to the learner. The contiguity principle states that words and pictures should be integrated rather than separated. Two contiguity effects have been identified by Moreno and Mayer, firstly the spatial-contiguity effect where words and pictures are separated either on the same screen and need to be combined to understand the on-screen content (see Figure 4) or placed on different screens thereby requiring the learner to move between screens to integrate the material in order to make sense of it.

Figure 4

In Figure 4, the learner is required to move from the labelled image to the appropriate piece of information and integrate all the pieces with its corresponding label. To reduce the spatial-contiguity effect, the text should be integrated with the image or positioned close to the part of the image to which it refers to as seen in Figure 5. In this case, the learner can reveal the information about the parts of the machine by hovering their mouse over each number. This allows the learner to both explore the image and reveal only one piece of information at any one time thereby reducing the amount of cognitive processing required.

Secondly, there is the temporal-contiguity effect where spoken words are presented before or after on-screen visuals and therefore require learners to hold some information in their working memory and integrate it with other information contained in the presentation. This again requires the use of limited cognitive resources to process the information being presented and is an example of representational holding.

Figure 5

Clark, Nguyen and Sweller refer to the separation of on-screen content as leading to ‘split attention’ meaning that learners have to attend to both sources of information separately. Research by Moreno and Mayer found that “learning is impaired when on-screen text is spatially separated from the visual materials”. This was supported by Clark and Mayer who also identify other examples of where the contiguity effect is violated in multimedia presentations, for example, when information is presented in a scrolling panel and the learner must locate some information and integrate it with on-screen images or when a question is asked and the feedback given is placed on a different screen to the question or when links to references appear in a new browser window or where audio narration is followed by a video rather than being presented at the same time. Therefore, when designing eLearning, on-screen objects should be positioned in close proximity to descriptions and if audio narration is used, it should be synchronised with the objects being described.

Modality Principle

The nature of multimedia presentations allow for the inclusion of combinations of words, pictures, audio and video so another important design consideration is how and when to incorporate each (assuming that it is practicable to do so). According to the modality principle, learning is improved when on-screen text that describes an image is replaced by audio. As mentioned earlier, the Cognitive Theory of Multimedia Learning assumes that information is processed via dual channels i.e. visual and auditory. When only words and pictures are used on-screen, as seen in Figure 3, the visual channel can become overloaded. However, replacing on-screen text with audio spreads the cognitive load across both channels reducing the processing demand on a single channel. Mayer and Moreno refer to the use of audio in this way as ‘offloading’ as some of the cognitive processing of the visual channel is off-loaded to the verbal channel. It should be mentioned that the modality principle has the most research support of any of the design principles and many experiments have reached the same conclusion, that pictures combined with narration results in improved learning over pictures and on-screen text.

In Figure 6, much of the on-screen text has been replaced by audio describing parts of the machine with the on-screen labels (synchronised with the audio) used to guide the learner to the part of the machine that the audio is referring to. Incidentally, presenting the information in this way also supports temporal-contiguity and reduces the amount of information that must be held in working memory.

Figure 6

Redundancy Principle

Whilst using text and audio to present on-screen information spreads the cognitive processing across processing channels, presenting audio and identical on-screen text results in redundant information being processed by the learner i.e. the same information is being processed by both channels. Studies show that learners who viewed multimedia presentations containing animations and narration outperformed those who viewed the same presentations that contained animation, narration and on-screen text. Further research by Mayer and Johnson confirmed this but with the addition of a limitation “except when the on-screen text is short, highlights a key action described in the narration, and is placed next to the portion of the graphic that it describes”. Therefore, the example in Figure 6 would not be an example of redundancy and the text is short and is positioned near to the part of the image that the narration is describing.

Figure 7

The screen example shown by Figure 7 is an example of redundancy when the on-screen text is replicated by the audio narration being spoken. However, Figure 6 is an example that has had the redundant text removed except for a key piece of text that is timed to appear on-screen as the audio is being spoken.

Personalisation Principle

Another aspect of designing multimedia presentations is the style of writing used throughout. The opinions of designers is often divided when asked if it is better to write in a formal or conversational style. In studies by Mayer, Fennell, Farmer and Campbell, a multimedia lesson on how the human respiratory system works, the word ‘the’ was changed to ‘your’ in 12 places throughout the lesson. The hypothesis was that “using the self as a reference point increases learner interest and encourages the learner to use available cognitive capacity for active cognitive processing of the incoming information”. Results found that learning was improved on subsequent transfer tests but there was no significant improvement on retention tests. While the studies support the personalisation principle in multimedia instruction the authors caution the overuse of personalising as it may cause the instruction to contain interesting yet irrelevant details which in turn may distract the learner.

The use of on-screen coaches or ‘pedagogical agents’ has also been examined to determine if they assist or hinder learning in multimedia presentations. Research found that learning was improved in groups whose multimedia contained an agent over those whose did not. Furthermore, it was also found that the agent did not have to look ‘real’ as there was no significant difference in results from groups who had a cartoon character agent.

Summary

Delivering instruction via multimedia presentations is widely used in many organisations. In order to improve learning from multimedia, instructional designers should look to apply evidence based principles of multimedia instruction into their designs in order to improve learning outcomes. This essay has discussed the cognitive theory of multimedia instruction along with the assumptions of how the human memory system operates. Furthermore, principles of multimedia design that can improve learning such as the multimedia principle, contiguity principle, modality principle, redundancy principle and personalisation principle were also discussed and applied to a sample eLearning module.

References
Clark, R., & Lyons, C. (2004). Graphics for learning. San Francisco, Pfeiffer.
Clark, R., & Mayer, R. E. (2008). E-learning and the science of instruction: Proven  guidelines for consumers and designers of multimedia learning. John Wiley and Sons  Inc.
Clark, R., Nguyen, F., & Sweller, J. (2006). Efficiency in learning. San Francisco: John  Wiley & Sons Inc.
Mayer, R. E., Fennell, S., Farmer, L., & Campbell, C. (2004). A personalisation effect in  multimedia learning: Students learn better when words are in a conversational style rather than a formal style. Journal of Educational Psychology, 96(2), 389-395.
Mayer, R. E., Heiser, J., & Lonn, S. (2001). Cognitive constraints on multimedia learning:  When presenting more material results in less understanding. Journal of Educational  Psychology, 93(1), 187-198.
Mayer, R. E., & Johnson, C. I. (2008). Revising the redundancy principle in multimedia  learning. Journal of Educational Psychology, 100(2), 380-386.
Mayer, R. E., & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia  learning. Educational Psychologist, 38(1), 43-52.
Moreno, R., & Mayer, R. E. (1999). Cognitive principles of multimedia learning: The role of  modality and contiguity. Journal of Educational Psychology, 91(2), 358-368.
Reed, S. K. (2006). Cognitive architectures for multimedia learning. Educational  Psychologist, 41(2), 87-98.