Intrinsic cognitive load is the mental effort demanded by the inherent complexity of the
learning task and material (Cooper, 1998). For example, the level of difficulty of quantum physics has a relatively high inherent difficulty; the level of difficulty of a simple addition task (e.g., 1+1=2) has relatively low inherent difficulty. Extraneous cognitive load is cognitive work that does not have a payoff in learning; for example, distractions such as background noise in lecture videos or unreadable handwriting increase the difficulty of the learning task but have not been shown to increase learning (De Jong, 2010). Germane cognitive load is the cognitive work necessary to connect incoming information to existing knowledge; the term also refers to the product of the work. For example, when a learner is able to understand a new concept such as multiplication through associating it with the familiar context of addition, the learner is making the new concept germane and assimilating it to existing knowledge.
Exploring applications of Atkinson and Shiffrin's model in conjunction with cognitive load theory, Mayer and Moreno (2003) proposed a cognitive theory of multimedia learning based on the dual-channel assumption and limited-capacity assumption. Their theory states that a limited capacity of the learner’s brain selects and obtains a multimedia presentation (e.g., text, pictures, and auditory information) through the learner’s dual-channel and selects and organizes the presentation dynamically to produce logical mental constructs rather than interpreting them mutually exclusively.
Positing limitations on the cognitive load that can be processed and the high selectivity of the brain, Mayer, Moreno, and many other cognitivist learning theorists have studied ways to create effective learning designs by setting the objective of a learning outcome to minimize extraneous cognitive load, increase germane cognitive load, and manage intrinsic cognitive load
(Brame, 2016). They found that a wide variety of ways to achieve this objective had been
developed in recent years, including segmentation, signaling, matching modalities, and weeding, all of which I discuss in the following paragraphs.
Segmentation is simply the division of a body of information into small sections to ease the management of intrinsic load and increase the germane load. Studies have shown that maintaining students’ attention is difficult after 13 minutes and that presenting material in 3-6 minute videos leads to engagement of students, as measured by their willingness to continue watching up to 100% of the time (Guo et al., 2014). When learners are engaged, they can process the incoming information so that it becomes germane. D. Zhang, Zhou, Briggs, and Nunamaker Jr (2006) examined the influence of interactive videos on learning outcomes and learner satisfaction in online learning environments. Their results showed that segmenting a video is critical for students’ engagement. Ibrahim, Antonenko, Greenwood, and Wheeler (2012) conducted a study showing that segmenting a video into smaller units enabled students to transfer knowledge better and reported lower levels of learning difficulty.
Signaling is a strategy used to reduce extraneous load. Mayer and Johnson (2008) conducted experiments on college students in which they had two groups: one group’s lecture slides contained 2–3 signaling words (i.e., short, redundant words) that were identical to the words in the lectures’ speech and the other group’s did not. Results showed that the students whose presentation included short redundant words outperformed the non-redundant group on a subsequent test of retention, on the basis of which they concluded that signaling and redundancy reduced extraneous load. Moreno and Mayer (2007) analyzed the effect of directing attention to relevant information with signaling and segmentation in dynamic instructional videos by creating one with signaling and segmentation and one without. The findings showed that, while the
control group outperformed the signaling and segmentation group on the retention of theoretical information, the signaling and segmentation group performed better when asked to evaluate what they learned and to apply teaching skills in a classroom scenario. The signaling and segmentation group appears to have had lower levels of cognitive load. Within the framework of cognitive load theory, signaling is understood to reduce time spent identifying key ideas in lecture slides and thus to reduce extraneous cognitive load and enhance learning.
Matching modality is a strategy predicated on the assumption that learners can better manage the cognitive load if the proper channels are activated (Mayer & Moreno, 2003). Research suggests that using audiovisual materials selectively to activate visual and auditory channels appropriately would increase student engagement with videos to provide flexibility in learning experiences (Thomson, Bridgstock, & Willems, 2014). One additional common practice is weeding (i.e., reducing background noise or eliminating extra animation that does not add value). Weeding minimizes the extraneous load so that more of the learners’ cognitive capacity can be used for the germane load; (Ibrahim et al., 2012). Mayer and Johnson (2008) has also conducted research exploring the redundancy effect; the study found that it tends to show reduced extraneous load processing. These applications of cognitive load theory to multimedia learning have spurred numerous advances designed to ease the cognitive load placed on students.
A typical MOOC lesson demonstrates how these concepts underlie the design: it takes approximately 30 minutes and the lesson is composed of 4-9 minute of modules, tests and quizzes, and various tasks (Abbakumov et al., 2020). The features and components named above reflect the research-informed choices that the online learning designers must make and take into consideration how the length, difficulty, order, etc. of the materials affect students’ learning in various ways. These have been based primarily on quantitative analysis. However, there is a lack
of research looking at quantitative and qualitative data in combination. Furthermore, at the same time that the research was showing positive effects of cognitive load minimizing strategies, a seemingly contradictory idea called desirable difficulty began to be investigated, in which making things harder rather than easier is posited to have long-term benefits.