Cognitive Load Theory getting stuff into long term memory.
Cognitive load theory theory explores how to optimize use of our working memory. The essential bottleneck is that working memory is extremely limited. Some research indicates that if one does not refresh the info in STM every seven seconds then it starts to dissipate. One implication is that it’s quite easy to overwhelm a learner to the extent they can really learn new material, or they can’t learn very efficiently.
Cognitive load theory was started by John Sweller in 1989. In the original theory he proposed there are three sources of load on short-term memory. In the 2000s Richard Mayer proposed some adjustments to the theory, and renamed it the cognitive processing theory. Today researchers use either theory. Although not strictly equivalent, they are very similar.
Three Sources of Processing
Most multimedia research either uses Sweller’s framework of cognitive load or Mayer’s complementary framework of cognitive processing. Both frameworks have been used to test and develop evidence-based instructional practices that better meet the needs of students learning conceptually difficult material. We employed Mayer’s cognitive processing framework to describe why a new application of multimedia learning seems to hold promise as an effective method for helping new doctoral students in education become efficient and thoughtful readers of original research.
Mayer’s cognitive processing theory (2008) explains how to optimize the constraints of working memory. Specifically Mayer identified three sources of processing that can occur in working memory: extraneous, essential, and generative. The key concepts are that working memory capacity is extremely limited and the three processing sources are additive. This means that if one source of processing is too high, then there will not be enough capacity left in working memory to engage in the other forms of processing. This can lead to inefficient and ineffective instructional experiences.
Sweller’s cognitive load theory uses the terms extraneous load, intrinsic load, and germane load. It’s crucial to understand that both variations of the theory see the three “loads” or “processing types” as being both independent and additive. The major consequence of this thinking is that if extraneous processing is too high, then that leaves little space for essential and generative processing work to be done in working memory. Most of the research to date has looked at various methods to eliminate extraneous processing.
Extraneous processing occurs when the learner is overwhelmed or distracted by irrelevant information or poor instructional design. Mayer (2008) describes extraneous processing as a type of cognitive processing that wastes cognitive capacity without helping the learner.
Very simple and seemingly small things can increase extraneous processing. For example, if text and a relevant graphic organizer in a book are not well integrated, then the learner may end up wasting cognitive energy going back and forth between the image and the relevant text.
Reducing Extraneous Processing
The best examples I’ve seen about how to reduce cognitive load on the web are some of the examples developed by Joey Cherdarchuk at Darkhorse Analytics. His motto is: Remove to Improve.
He has developed three great examples. When you go to his blog page you’ll see each example in a self-running animated gif. Underneath you’ll find a slideshow widget that allows you to go through each stake of the remove-to-improve process at your own pace.
Check out Joey Cherdarchuk’s complete set of examples by using the links below.
One episode of Brain Slides was devoted to designing slides for cognitive load. This short less-then–4-minutes video provides a very nice example of how to make a table in a presentation simpler, more intuitive, and thus more compelling and understandable.
Essential processing is the amount of processing needed to work with the instructional material. This type of processing is largely explained by the complexity of the material. Sweller (2005) defined complexity as the number of interactive elements impacting concept mastery. The greater the number of interactive elements, the more difficult the material will be for learners.
Mayer proposed that there are at least three principles one can employ to reduce the demands on essential processing: segmenting, pre-training, and modality. Synthesizing 25 different tests from his own research, Mayer (2008) reports an average Cohen’s d effect sizes of 0.98 for segmenting, 0.85 for pre-training, and 1.02 for modality. These are all large effect size results supporting the contention that essential processing demands can be reduced.
Other researchers have provided evidence that additional instructional approaches may help reduce essential processing. For example, van Merrienboer, Kester, & Paas (2006) made a compelling argument that when learning complex tasks, instructors should limit the natural element interactivity inherent in the task. They specifically suggested sequencing using a part-whole approach to scaffold learning. Other researchers (e.g. Ayres, 2006 and Gerjets, Scheiter & Catrambone, 2004) have explored alternative techniques for reducing intrinsic load by using modular examples or part/whole tasks.
Overall current research provides optimistic initial results regarding the ability to reduce the essential processing needed to engage in complex learning challenges. Based upon the current research base it would seem reasonable to infer that instructional techniques which scaffold the learning of difficult conceptual material succeed because they reduce the essential processing demands on working memory.
Generative processing is the cognitive processing used to make sense of new learning materials. The result of good generative processing is that the learner constructs new schemas. If extraneous and essential processing are too high, then there will be no capacity left in working memory to engage in generative processing. However, assuming that extraneous and essential processing demands are managed well in the learning environment, then the level of generative processing seems to be mainly determined by student motivation and the structure of the learning challenge (Moreno & Mayer, 2010).
Generative processing mainly involves organizing and integrating. As such, generative processing is an active process. Relatively little research has been conducted regarding optimizing generative processing. However, based on the research conducted so far, Moreno and Mayer (2010) identified five principles for enhancing generative processing:
- multimedia: asking students to build referential connections between verbal and pictorial representations
- personalization: inducing the feeling that students are participants rather than observers of a learning environment
- guided activities: scaffolding the exploration of interactive games
- feedback: providing principle-based feedback to students’ responses
- reflection: prompting students to explain and evaluate their understanding
Each of these principles is worthy of consideration when using various instructional activities. In many cases all five principles can be used with an activity.
This is a special mini-website. This site includes examples and downloads. The presentation covers a wide spectrum of research and practical applications. While this mini-website is not in-depth, it does point to several very helpful resources.