Cognitive load theory


  • Cognitive load theory (CLT) is a theory of instructional design based on an experimentally derived model of cognitive architecture and focuses on the ease with which novel information is processed by working memory, which has limited capacity
  • The theory was developed by John Sweller in the 1980s and has clear implications for how instruction should be delivered, including differences between novice and expert learners


The human cognitive system has limited working memory capacity when receiving sensory information

  • holds ‘seven plus or minus two’ elements and actively processes only two to four elements at once
  • handles information for a few seconds and loses almost all information after 20 seconds unless refreshed by rehearsal
  • applies to completely novel, unorganised information

However, working memory has no known limitations when information is retrieved from long-term memory as cognitive schemas

  • schemas, even if highly complex, are handled as a single element in working memory
  • schemas organise knowledge and reduce working memory load
  • schemas vary in complexity
    • e.g. ‘illness scripts’ are schemas used to distinguish between different diseases, complexity and automation increases with expertise
    • e.g. distributive shock is recognised by an expert instantly but appears as unstructured set of symptoms and signs to a novice
  • automated schemas free working memory for other activities as they do not need to be processed in working memory

Schema construction and schema automation are learning processes

  • cognitive schemas are developed by
    • bringing new elements together (‘chunking’)
    • incorporating new elements into schemas already held in long-term memory
    • obtaining already schematized information from other people
  • automated schemas
    • are produced by repeated application with desired results through a great deal of practice
    • only develop for aspects of performance that are consistent across task situations

All novel information must be processed in working memory in order to construct schemas in long-term memory


Working memory is affected by 3 types of cognitive load:

  1. Intrinsic load — intrinsic to the learning task
  2. Extraneous load — characteristic of the instructional procedures used to present the learning task
  3. Germane load — created by the learning that occurs when dealing with the intrinsic load

Intrinsic load

  • depends on the number of elements that must be simultaneously processed in working memory
  • depends on the extent of element interactivity (the degree to which the elements of something can or cannot be understood in isolation)
    • low interactivity (e.g. learning words) yields low cognitive load
    • high interactivity (e.g. learning grammar) yields high cognitive load
    • load varies with expertise, i.e. an expert may have created schemas that incorporates interactive elements and thus reduces cognitive load

Extraneous load

  • increased if:
    • information must be searched for
    • information that is distributed in time or place needs to be integrated
    • information from multiple sources is presented via one sensory channel (e.g. text combined with a diagram)
    • information is provided by multiple sensory channels that compete (e.g. talking fast while showing a complex slide)

Germane load

  • results from:
    • induction or ‘mindful abstraction’ (combining new elements into a schema)
    • elaboration (connecting new elements to an existing schema)
    • incorporating another person’s schema

Cognitive load theory assumes that the different types of cognitive load are additive

  • e.g. decreasing extraneous load can compensate for increased intrinsic load
  • e.g. with complex learning tasks the combination of intrinsic and extrinsic load may exceed working memory capacity


Design principles and strategies aim to:

  1. decrease extraneous load
  2. manage intrinsic load
  3. optimise germane load

Decreasing extraneous load

  • Goal-free principle
    • replace conventional tasks that have a specific goal with goal-free tasks
    • e.g. ask “tell me some illnesses that match these symptoms?” instead of “which illness is indicated by these symptoms?” (avoids ‘back searching’ from the goal, which has high extraneous load)
  • Worked example principle
    • replace conventional tasks with worked examples that have solutions provided to be studied carefully
    • e.g. “what do you think of this treatment plan?” rather than “create a treatment plan” (avoids weak problem-solving methods)
  • Completion principle
    • replace conventional tasks with completion tasks that give a partial solution that must be finished by the learner
    • e.g. allow trainees to perform part of a procedure rather than attempting the whole procedure
  • Split attention principle
    • provide one integrated source of information rather than sources distributed in space and/or time
    • e.g. provide a ‘just in time’ video showing a procedure rather than providing them with reading and diagrams beforehand
  • Modality principle
    • use multimodal sources of information rather than unimodal sources with multiple elements
    • e.g. provide an animation with a voice-over (both visual) rather than an animation with written instructions (visual and auditory)
  • Redundancy principle
    • use one source of information rather than multiple self-contained sources
    • e.g. in a diagram of the flow of blood through an ECMO circuit do not provide a verbal description of blood flow through each component

Managing intrinsic load ( cannot be done without reducing learner understanding, so must involve a multi-step approach)

  • Simple to complex strategy
    • first present isolated elements (low element interactivity) then gradually work up to the tasks in their full complexity
    • e.g. identify the components of the ECMO circuit before explaining how blood flow and gas exchange is regulated
  • Low-to-high fidelity strategy
    • perform tasks in a low fidelity environment before moving to a high fidelity environment
    • e.g. start with case-based problems, then do simulation-based learning, before moving on to real patients in the workplace

Optimising germane load (sometimes desirable to increase intrinsic load to increase the associated germane load)

  • Variability principle
    • perform tasks in a low fidelity environment before moving to a high fidelity environment
    • e.g. start with case-based problems, then do simulation-based learning, before moving on to real patients in the workplace
  • Contextual interference principle
    • replace tasks that have similar surface features start with tasks that differ from one another in dimensions seen in the real world
    • e.g. use cases with different age, sex, past medical history for clinical teaching (helps construct schemas based on relevant information)
  • Self-explanation principle
    • replace separate worked examples or completion tasks with enriched ones containing prompts, asking learners to self-explain the information
    • e.g. to teach cardiovascular pathophysiology, present an animation of the circulatory system and provide prompts to self-explain the underlying mechanisms (increases interactive elements by including prior-knowledge items from long-term memory)


Expertise reversal effect occurs when instructional methods that are effective for novices are less effective or have adverse effects when used for experts

Cognitive load design principles that should be modified include:

  • Completion strategy
    • increase the amount of completion required by learners as they increase in expertise
    • e.g. start with worked examples, then completion tasks, then allow them to independently create their own complete solutions
  • Fading guidance strategy
    • provide a series of tasks that decrease in the amount of guidance (‘scaffolding’) provided as the level of expertise increases
    • e.g. initially provide step-by-step instructions and feedback for a procedure, then feedback only, then no guidance
  • Integrated to non-integrated strategy
    • replace a uniform series of integrated examples with a two-stage series of integrated examples, then non-integrated examples
    • e.g. provide novices with text and figures describing how to perform a procedure, but provide only the figures for learners with greater expertise
  • Dual-to-single mode strategy
    • replace a uniform series of dual-moded presentations with a two-stage series of dual-moded presentations, then uni-modal presentations
    • e.g. provide novices with an animation voice-over, but switch off the sound for those with more expertise


  • how to provide instruction to a group of learners of varying levels of expertise
  • how to measure cognitive load
  • implications for self-directed learning

Journal articles

  • Kalyuga S. Expertise Reversal Effect and Its Implications for Learner-Tailored Instruction. Educ Psychol Rev. 19(4):509-539. 2007. [article]
  • van Merriënboer JJ, Sweller J. Cognitive load theory in health professional education: design principles and strategies. Medical education. 44(1):85-93. 2010. [pubmed]

FOAM and web resources

Chris is an Intensivist and ECMO specialist at the Alfred ICU in Melbourne. He is also a Clinical Adjunct Associate Professor at Monash University. He is a co-founder of the Australia and New Zealand Clinician Educator Network (ANZCEN) and is the Lead for the ANZCEN Clinician Educator Incubator programme. He is on the Board of Directors for the Intensive Care Foundation and is a First Part Examiner for the College of Intensive Care Medicine. He is an internationally recognised Clinician Educator with a passion for helping clinicians learn and for improving the clinical performance of individuals and collectives.

After finishing his medical degree at the University of Auckland, he continued post-graduate training in New Zealand as well as Australia’s Northern Territory, Perth and Melbourne. He has completed fellowship training in both intensive care medicine and emergency medicine, as well as post-graduate training in biochemistry, clinical toxicology, clinical epidemiology, and health professional education.

He is actively involved in in using translational simulation to improve patient care and the design of processes and systems at Alfred Health. He coordinates the Alfred ICU’s education and simulation programmes and runs the unit’s education website, INTENSIVE.  He created the ‘Critically Ill Airway’ course and teaches on numerous courses around the world. He is one of the founders of the FOAM movement (Free Open-Access Medical education) and is co-creator of litfl.com, the RAGE podcast, the Resuscitology course, and the SMACC conference.

His one great achievement is being the father of three amazing children.

On Twitter, he is @precordialthump.

| INTENSIVE | RAGE | Resuscitology | SMACC

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.