COGNITIVE LOAD THEORY: RESOURCE DOCUMENT

Daniel A. Taylor

Overview

        When we study learning and how it occurs, we begin to see that which can be learned is shaped and limited by how it is presented, in what context, and in what environment.  Neurology teaches us that the mind is reliant upon the brain and its complex network of synapsing neurons.  Though numerous, the quantity of neurons in a single brain is finite, as, as such, can only transmit or compute a certain amount of data at any given time.  The amount of data that is perceived at any given time is the perceptual load, while the amount that is being understood and considered thoughtfully is the cognitive load (Chen & Epps, 2014; Choi, van Merriënboer & Paas, 2014; Sweller, van Merriënboer & Paas, 1998).

        Cognitive load theory is concerned with the limitations of the cognitive pathway, which connects working and long-term memory.  When a learner is actively attempting to learn, the theory states that extraneous information will increase the cognitive load.  Since information transfer between these centers of the brain is limited, extraneous information will reduce the amount of learning that occurs.  To combat this, cognitive load theory directs instructional designers to create simple instruction that emphasizes the content that is to be learned, to assure that transfer of this important material occurs (Sweller, et al., 1998).

        There are several foundational concepts that are essential for understanding cognitive load theory: working memory, long-term memory, schema construction, schema automation, and the types of cognitive load.

Working Memory

        The perceptions that we are aware of at any given time are considered to be within working memory (also known as short-term memory).  Perceptions are within working memory when they are being attended to, and, once there, can be further considered, processed, and organized.  According to the famous paper by Miller (1956)^[1]--and others who have reinforced his findings--working memory can hold seven elements of information (plus or minus two) at any given time.  However, the very process of evaluation of the material in working memory takes up working memory space itself.  Thus, when we are actively thinking, we can hold fewer discrete facts readily at hand, and can only perform a limited amount of processing.  Additionally, different types of sensory information (e.g., visual information versus touch information) appear to be processed by different mechanisms (Sweller, et al., 1996).  

Long-Term Memory and Schema

        Those facts and concepts that are stored in our minds are said to be in the long-term memory.  While in long-term memory, facts are not consciously perceived, of course, but are available to be retrieved from long-term memory and transferred to working memory for consideration.  The data that can be stored in long-term memory appear to be practically unlimited.  These data are organized into schema, which sort the data by their perceived relationships, determined based upon experience and reasoning.  Thus, when retrieving information from long-term memory, a thinker does not have to reconsider the many relationships of the information in question, so long as they had been understood in the past and stored in a schema.  Schemas can be conceptualized as a cabinet with many slots; the cabinet is a concept, and the slots are filled with other concepts that are perceived as being related to the main cabinet.  As a perceived experiences reveal new relationships, new information enters some of the slots and incorrect or anachronistic information is replaced (Sweller, et al., 1996).  

        Schemas, therefore, help reduce the cognitive load for familiar tasks.  By accessing the relational information found within a schema, the working memory is freed from considering those relationships again.  They are obvious, and part of the fact that is being considered.  Thus, the understanding of a relationship, which might account for, say, three units in the working memory when unfamiliar, may only account for one unit once it has been learned and placed within a schema.  This frees two units of working memory for the consideration of other unfamiliar facts or relationships (Sweller, et al., 1998).

Schema automation.  An important distinction must be made here.  Not all data that are stored in long-term memory go through the process described above;  only those which are consciously perceived do so.  Perceptual data go through automatic processing when they are elements in a cognitive procedure that has been previously practiced.  This practicing develops schemas that act as sophisticated rules in future encounters.  When our perceptions match one of these schemas, we automatically process the perceptions without a concurrent demand on working memory.  Thus, it is important that schemas not simply be built, but be automated, in order to reduce cognitive load and allow for learning (Sweller, et al., 1996).

Types of Cognitive Load

        We have already defined cognitive load as the amount of data that is being consciously analysed at any given time, and identified that it occurs within the limitations of the working memory.  This analysis process will eventually build a schema that can be recalled in future situations, thus reducing the cognitive load.

        With that being said, research has indicated that there are different types of cognitive loads, which are processed by different mechanisms within the working memory, and require varying amounts of working memory for their functions.  Not all cognitive loads, therefore, are created equal (Sweller, et al., 1996).

        Intrinsic cognitive load.  Material that is being percieved has a certain amount of cognitive load that is simply part of it.  A sentence, for example, is likely perceived by most readers in units of individual words that are efficiently (perhaps in some cases, automatically) transformed by existing schema into concepts.  The words themselves, however, cannot be reduced.  There will always be the same number of words that must be read to understand a sentence and, thus, some cognitive load that will be intrinsic to the material (Sweller, et al., 1996).

        Intrinsic cognitive load is related to the interactivity of the content.  Specifically, interactivity refers to how self-sufficient the elements of the material are.  An element that does not require the simultaneous cognition of other elements has low interactivity, and low intrinsic cognitive load.  High-interactivity material requires many concepts to be held in working memory, and is thus difficult to understand.  Students can fail to understand when the interactivity is too high, but--since students with more expertise will understand content according to their existing schema, the interactivity will be lower for them than for a novice (Sweller, et al., 1996).

        Extraneous cognitive load.  Elements of instruction that are not requisite parts of the data to be learned, but are rather arbitrary presentation techniques or peripheral activities, comprise extrinsic cognitive load.  In other words, the learner has some of his working memory occupied by perceptions that are not necessary to the learning of the material.  This is under the control of the instructional designer, and should be minimized as much as possible so that the limited resources of working memory can focus upon intrinsic and germane cognitive loads.  Techniques for reducing extraneous cognitive load will be discussed in a later section (Sweller, et al., 1996).

        Germane cognitive load. Similar to extraneous cognitive load is germane cognitive load, which focuses upon the cognitive demand made from the construction of schemas.  Since this type of cognitive load reflects an effort that leads to the development of sophisticated and useful schemas, it should be encouraged as instruction is designed.  Methods for accomplishing this will be discussed in the next section.

Implications for Instructional Design

        Knowledge of the components of cognitive load theory supports several design principles that help instructional designers create learning experiences that maximize the efficiency and quantity of learning.  It is essential that the amount of material presented at one time not overload the cognitive limit of working memory, lest the learner be unable to consider the many facts without forgetting them.  Therefore, complex reasoning should not be presented along with many unfamiliar terms.  It is better to teach the terms first, so they can be learned and stored in long-term memory, and presenting the reasoning exercise later, when the requisite facts are well-known to the learner.  By doing this, the process of problem-solving is simplified, reducing the number of extraneous elements that the learner must keep in his working memory to complete the task.  Similarly, mental comparison and integration of text and pictures in a resource may cause high demands on cognitive load.  Instead, the wise designer might use either text or pictures.  These design decisions reduce the extraneous cognitive load (Sweller, et al., 1996).

When increasing germane cognitive load to encourage schema formation, it is important that overall cognitive load not overload the working memory.  Thus, efforts to increase germane load should be coupled with decreases in extraneous load.  This redirects the attention of the learner from irrelevant (e.g., resource media, lecture hall environment) to relevant processes (Sweller, et al., 1996).

Clark and Mayer (2011) have compiled many instructional design principles that reduce extraneous cognitive load and increase germane load when followed.  According to the multimedia principle, words and graphics should be used together, but only graphics that support the learning (thus increasing germane load).  The contiguity principle hold that words should be aligned with their graphics (which reduces the extraneous load required to compare the words and graphics).  The modality principle states that words should be presented as narration rather than text, which is based upon the idea (described in Working Memory, above) that visual and auditory sensations are processed separately by working memory, so that visual and auditory information together are less likely to overload it than two different types of visual information.  By no means should text and audio both be used, which would unnecessarily increase extraneous cognitive load, violating the redundancy principle.  Clarity and lack of clutter in a lesson is summarized as the coherence principle, which, when followed, keeps extraneous cognitive load low, so that focus may be put on germane and intrinsic cognitive load instead. Unnecessary sounds, text, and graphics should be eliminated.  The personalization principle holds that informal language is easier to learn than formal language, and that it is easier to learn from a perceived person than an impersonal source.  Since most of us have developed our cognitive schemas mainly from personal interactions with others, these schemas are more likely to be used in the understanding of new material if it appears to be coming from a person.  Finally, the segmenting and pretraining principles divide instruction into smaller parts to avoid overloading the working memory (as discussed at the beginning of this section).

History

        Cognitive load theory emerged from schema theory, which states that information is organized in the human memory according to different schemas, or perceptual organizational centers.  We rely upon our schemas to help make sense of the world, by applying certain concepts to our experience when it  contains some of the elements found in a schema (Kennedy,  Mason,  Mazurik, & Meadows, 2012).

        Sweller (1988) identified an experimental oddity in schema theory.  Subjects who were engaged in solving physics problems performed differently depending upon whether they were experts in such activities--and thus had an existing problem-solving schema to use--or they were novices, who lacked a pre-existing schema for the mental activity.  So, one can understand that it is in the best interest of a novice to acquire a robust problem-solving schema, to make his solving easier and more automatic.  However, when engaged in working physics problems, subjects did not acquire this schema.  Sweller hypothesized that this lack of schema acquisition was due to cognitive interference; that is, limited cognitive processing capacity meant that one could not work problems and develop a problem-working schema simultaneously.  Thus, cognitive load theory was born.

Key Persons

        John Swiller.  Research performed by John Swiller led to the development of Cognitive Load Theory from the midst of Schema Theory.

        Fred G. W. C. Paas and Jeroen J. G. van Merriënboer.  These men used Swiller’s work to further develop Cognitive Load Theory.

Differentiation

        Cognitive load theory helps instructional designers create instruction that efficiently directs the learner’s attention to the material that is to be transferred, without adding to his cognitive load with extraneous material.  Thus, it focuses on efficiency in instructional design.  This differs from other instructional design theories (Driscoll, 2005)^[2].

        Radical behaviorism focuses upon actions and behaviors and does not consider cognitive processes.  It has a different emphasis than cognitive load theory.  However, one might expect that, eventually, a behaviorist would stumble upon the same methods as a cognitive load theorist, since they will lead to more efficient learning and subsequent changes in behavior (Driscoll, 2005)2.

        The variety of theories that hypothesize on the nature of knowledge have little in common with cognitive load theory, mainly because of varying emphases.  Though founded in schema theory, cognitive load theory goes beyond it, positing that there is a limited amount of schema that be accessed or created simultaneously, while schema theory merely describes the nature of knowledge organization.  Similarly, situated cognition theory posits that all human thought is based on the environment of that human; and constructivist theories, that knowledge is self-generated and not reflective of any objective reality.  These theory are not primarily concerned with efficiency of learning, but where knowledge comes from.  Another example is Piaget’s genetic epistemology, which  holds that children are genetically-designed to develop mentally in certain stages.  Again, though cognitive load theory simply has little to disagree nor agree with in this theory, the cognitive load theorist would likely emphasize instructional design over genetics when determining the vital factor in learning and mental development.  In this way, cognitive load theorists would likely have more in common with Bruner and Vygotsky, with their non-genetic emphases on the origins of thought, though cognitive load theory is more concerned with eliminating the causes of problems (Driscoll, 2005)^[3].

Bibliography

Chen, S., & Epps, J. (2014). Using task-induced pupil diameter and blink rate to infer cognitive load. Human-Computer Interaction, 29(4), 390-413. doi: 10.1080/07370024.2014.892428

                All commonly-used measures of cognitive load are subjective, making research on the theory imperfect.  Some have proposed the use of biometric indicators, like blink rate and pupil diameter, as objective measures of cognitive load.  The authors find that these measures only were useful in measuring cognitive load when perceptual load was low, indicating that other objective readings may be necessary (Available at http://www.researchgate.net/profile/Julien_Epps/publication/262577448_Using_Task-Induced_Pupil_Diameter_and_Blink_Rate_to_Infer_Cognitive_Load/links/53f491ee0cf2fceacc6e8e0c).

Choi, H. H., van Merriënboer, J. J. G., & Paas, F. G. W. C. (2014). Effects of the physical environment on cognitive load and learning: Towards a new model of cognitive load. Educ Psychol Rev, 26(2), 225-244. doi: 10.1007/s10648-014-9262-6

In this revision of cognitive load theory, the authors add the physical environment        to the list of factors that cause cognitive load to increase.  This represents a shift from the original exclusive consideration of factors that are internal to the learner.

Clark, R. C., & Mayer, R. E. (2011). E-Learning and the Science of Instruction: Proven Guidelines for Consumers and Designers of Multimedia Learning (3rd ed.). San Francisco, CA: Pfeiffer.

                Though specifically concerned with eLearning design, Meyer and Clark present guidelines, based on their research, on the best ways to reduce cognitive load when developing instruction for maximum retention by students.  This book is a masterpiece of evidence-based instructional design.

Paas, F. G. W. C., & van Merriënboer, J. J. G. (1994). Instructional control of cognitive load in the training of complex cognitive tasks. Education Psychology Review, 6(4), 351-371.

                This article takes Sweller’s cognitive load theory and directs it towards instructional design, identifying elements of instructional design that may increase cognitive load, and strategies for reducing cognitive load.  It is an important article for both historical and practical reasons (available at http://anitacrawley.net/Articles/Paas%20Instructional%20control%20of%20cognitive%20load%20in%20the%20training%20of%20complex%20cognitive%20tasks.pdf).

Seery, M. K., & Donnelly, R. (2012). The implementation of pre-lecture resources to reduce in-class cognitive load: A case study for higher education chemistry. British Journal of Educational Technology, 43(4), 667-677. doi:10.1111/j.1467-8535.2011.01237.x

                This case study demonstrates the benefits of preparatory work prior to coming into the classroom, from a cognitive load theory perspective.  Preparation decreases cognitive load that would otherwise be wasted on learning new nomenclature and concepts, allowing more cognitive load to be used in true understanding.  This can lead to improved grades in courses.

Smith, A., & Ayres, P. (2014). The impact of persistent pain on working memory and learning. Educ Psychol Rev, 26(2), 245-264. doi: 10.1007/s10648-013-9247-x

                An article reporting research results that show cognitive capacity may be reduced by physiological phenomena, such as pain.  This has important implications to the field, as stimuli need not be cognitive in nature to affect cognitive load.

Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12, 257-284.

                Arguing that problem-solving activities may prevent the acquisition of mental schema, thus interfering with learning, Sweller opened the door to the development of cognitive load theory by positing a limited cognitive processing capacity.  This is the seminal article for the theory, and shows its foundations in schema theory quite well (available at http://csjarchive.cogsci.rpi.edu/1988v12/i02/p0257p0285/main.pdf).

Sweller, J., van Merriënboer, J. J. G., & Paas, F. G. W. C. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10(3), 251-296.

A fine review of the first ten years of research in Cognitive Load Theory, Sweller, van Merriënboer and Paas describe the concepts and ideas that are foundational to the theory, particularly as they relate to instructional design.  This is a fine, in-depth overview, and worth referring to (available at http://www.davidlewisphd.com/courses/EDD8121/readings/1998-Sweller_et_al.pdf)

Web Resources

Angelo, T. (2008, June 14). Lecturing for learning: Cognitive load. [Video file]. Retrieved from http://youtu.be/OtXtYNOiEIU

                This short video lecture describes and illustrates cognitive load use in a classroom setting, emphasizing the problems Cognitive Load Theory identifies when lecturing to students for whom English is not the primary language, or who come from another culture.

Bozarth, J. (2010, August 3). Nuts and bolts: Brain bandwidth - cognitive load theory and instructional design. Learning Solutions. Retrieved from http://www.learningsolutionsmag.com/articles/498/nuts-and-bolts-brain-bandwidth---cognitive-load-theory-and-instructional-design

                For a more user-friendly introduction to the theory, this article does a good job walking the reader through the basics.

Chabris, C., & Simons, D. (1999). Selective attention test. [Video file]. Retrieved from http://youtu.be/vJG698U2Mvo

                Though not an instructional design video per se, this demonstrates how our cognitive load can influence the details we perceive.  It is worth the minute or so it takes to watch, and a great demonstration for the classroom.

Cognitive load. (2014, August 15). Retrieved September 28, 2014 from Wikipedia: http://en.wikipedia.org/wiki/Cognitive_load

                The Wikipedia article for cognitive load has a nice overview of the theory, its component ideas, history and use.  This is a quite comprehensive and effective review.

Cognitive load theory. [Video file]. (2011, July 29). Retrieved from http://youtu.be/kaGl2u7J2wY

                Illustrates a practical implication of cognitive load theory in marketing.  Though not specifically related to instructional design, it does show that cognitive load theory has many possible considerations.

Kennedy, C., Mason, J., Mazurik, S., & Meadows, K. (2012, September 6). Schema theory & cognitive load theory. [Prezi presentation]. Retrieved from Prezi: http://prezi.com/1apjupioiqts/schema-theory-and-cognitive-load-theory/

                This elegant presentation describes cognitive load theory in relation the schema theory, from which it emerged.  It gives a history of its development, and its main points, as well as implications for instructional design.

meagancamp. (2010, November 24). The cognitive load theory. [SlideShare slides]. Retrieved from SlideShare: http://www.slideshare.net/megancamp/the-cognitive-load-theory.

                This SlideShare deck goes through the basics of Cognitive Load Theory in some depth, and provides nice visuals to help improve understanding of the theory.

Soloman, H. (n. d.). Cognitive load theory. Retrieved from http://www.instructionaldesign.org/theories/cognitive-load.html

                In addition to providing a good overview of the theory, this webpage also includes several videos that help describe and illustrate it.