Slava Kayuga Human cognitive architecture and its implications

Slava Kayuga Human cognitive architecture and its implications for the design of instruction: Introduction to cognitive load theory

Working memory (WM) o Information enters WM once it has been selected by allocating attention to it o We have limited attention because of limitations of WM o Corresponds to consciousness or awareness: we are conscious of everything that is in WM

Working Memory Repeat a telephone number What have you been doing just before this? 12 + 13 = ? 83468437 + 93849045 = ? Taking notes – extension of WM

Short-term or working memory? o Early models of memory referred to STM; it is still commonly used today o STM was thought of in terms of only storing information (temporarily remembering) o Baddeley and Hitch (1974): we not only store information for short periods of time but also process information - hence WM

WM capacity o Miller (1956) demonstrated that we have a short-term memory span of 7 ± 2 units of information – storage capacity o Reconsideration of WM capacity when processing is involved (Cowan, 2001) o In terms of processing information, 4 is a more likely number than 7

WM processing capacity Suppose 5 days after the day before yesterday is Friday. What day of the week is tomorrow?

WM duration Brown (1958); Peterson & Peterson (1959): When people are distracted from rehearsing, information is lost rapidly (e. g. , after 18 sec – everything was forgotten)

WM Structure Baddeley 1986, 2001 Executive Control System Controls the Operations of Working Memory Visual-spatial Sketch Pad Visual Rehearsal Phonological Loop Auditory Rehearsal n Allocates resources to other systems- governs what enters WM n Director of cognitive workselects strategies n Not a store or processor n Processes visual images n Spatial processing n Holds acoustic or speech-based information n Auditory rehearsal of verbal information

Working Memory Repeat an unfamiliar foreign word Close your eyes and pick up an object in front of you How many windows are in your house?

Long-term memory (LTM) o permanent repository of the lifetime of accumulated information o unconscious component of our memory: we are not conscious of LTM information until it is activated and brought into WM and LTM are two major components of Human cognitive architecture

Role of LTM CIABBCABCBHPAMP CIA BBC ABC BHP AMP

Effective WM capacity n Miller (1956): short-term memory span is 7 ± 2 chunks of information n What each chunk consists is dependent on our knowledge stored in LTM n What is in LTM would affect the way we process information in WM

Effective WM capacity n Information-“rich” chunks n Chunking information into meaningful parts has the effect of expanding the capacity of working memory n Examples: a Chinese character; a written English word; newspaper vs textbook

Chess studies de Groot (1966) o Compared performance of chess masters and weekend players o Question: Do chess masters n look ahead more moves? n Consider a greater number of alternative moves? o Answer: verbal protocols showed NO difference between chess masters and weekend players

Chess studies de Groot (1966); Chase & Simon (1973) o Investigated: players’ memory of chess boards o Tested: master’s vs. weekend player’s memory for real and random board configurations after brief (5 sec) exposure o Results: masters were superior in reconstructing real game configurations (80 -90% correct compared to weekenders’ 30 -40%) but NOT random configurations o Conclusion: Superiority was due to greater amount of real-game chunks in master’s LTM

Role of LTM o Grand masters have extensive and better organized LTM knowledge base o 50 -100 thousand configurations, at least 10 years of experience o This study radically changed our view on the role of LTM in human cognition o LTM is not just for memorizing things, but is the most critical component of our cognition (including learning), the source of our intellectual strength

LTM in human cognition o Grand masters read the chess board the same way you read words in a text o Similar mechanisms for all high-level cognitive skills (e. g. , text comprehension) o LTM - not a passive store of information; it is actively used in most of cognitive processes and is central to perception, learning, problem solving

Schemas (schemata) “Organized structures that capture knowledge and expectations of some aspect of the world” (Bartlett, 1932) Organized knowledge structures that represent generic concepts and categorize information according to the way in which we use it

What is this list about? table chair knife fork spoon cup plate toast butter jam cloth juice bowl tea

Schemas Examples: a tree schema a face schema reading a page of prose: schemas for letters, words, phrases, sentence structures Restaurant script (procedural schema)

Schemas as major building blocks of cognition o Schema theory is the most commonly used framework for understanding LTM o Memory is actively constructed using schemas o Pre-existing schemas determine what incoming material is relevant n Relevant material processed n Irrelevant material discarded

Schema automation is achieved by practicing skills until they do not require consciously controlled and effortful processing. When basic mental operations occur automatically, resources are available for more sophisticated cognitive operations (e. g. , reading, math operations, etc. )

Automation Explains why individuals can n conduct difficult tasks n simultaneously conduct several tasks n read for meaning rather than focus on the individual letters and words n be accomplished performers (e. g. , musicians) n Automation is slow to develop and requires significant practice

Schemas o Schemas affect not only what we memorize, but how we think, reason, solve problems o Intelligence – in number and complexity of acquired schemas o Nature of expertise

Expert characteristics: Domain-specific knowledge q Experts have a large store of domain-specific schemas for problem solving in the domain q Automated schemas reduce WM demands and allow higher order functions (monitoring, evaluating etc. ) q Experts deal with problems at a deeper level: categorize according to deep structures (principles) rather than surface structures

Expert characteristics: Treatment of problem Task: categorize the following into 3 groups Soldiers, 1492, discovery, kings & queens, 1914, revolution, sailors, war, 1789. n Surface structure grouping: 1492, 1914, 1789 Deep structure grouping: 1789, Kings and Queens, revolution (French Revolution) n Physics experts classified problems according to the laws of physics rather than surface structures (e. g. Chi, Glaser & Farr, 1988)

Implications for improving problem solving q Acquisition of extensive domain-specific knowledge (schemas) is essential: the only way to be good in problem solving § broken car: we call a mechanic (an expert), not a general “problem solver” q You can become expert problem solver in a specific area, not in every area q Studying expert solutions § emphasising higher-order skills, categorization of problems

Arithmetic word problems (Marshall, 1995) Analysis of the task domain to identify core schemas: q After 6 passengers had left the bus, 9 passengers remained. How many passengers were on the bus initially? (Change Schema) q Peter's book contains 50 pages. Peter read 15 pages in the morning. In the afternoon, he read the remaining pages and finished the book. How many pages did Peter read in the afternoon? (Group Schema) etc.

Go Solve Word Problems Tom Snyder Productions

Instructional implications Do not overload WM! If material is difficult to learn, learner WM is likely to be overloaded Manage information-processing “bottleneck” by chunking information into meaningful groups based on available knowledge Help students to link new information with prior knowledge

Instructional implications Enhance acquisition and automation of knowledge in LTM - a major goal Use dual modality (visual and auditory) Minimise interference /distractions Provide adequate time to enable processing Instruction that requires many inferences (things are not stated explicitly) overloads WM

Cognitive Load Theory o Instructional theory that takes into account limitations of learner working memory o Cognitive load (working memory load): working memory capacity required by a particular cognitive task o Cognitive load depends on the level of interactivity between elements of information Sweller 1999; Sweller, Ayres & Kalyuga, 2011

Element interactivity Low High List of variables: a, x, b Equation: ax=b Names of electrical symbols and what they represent Operation of an electrical circuit Learning vocabulary of a foreign language Learning grammar

Measurement of Cognitive Load Objective measures § Task and performance § Secondary task § Psychophysiological Subjective measures § Rating scales

Objective measures § Secondary task Slow RT Rapid RT Resources to secondary task Cognitive resources to simple primary task Fixed cognitive capacity Resources to secondary task Cognitive resources to complex primary task Fixed cognitive capacity

Subjective measures § Rating scales In solving or studying the preceding problem I invested: very, very low mental effort neither low nor high mental effort very, very high mental effort

Subjective measures: Rating scales (NASA-TLX)

Types of cognitive load o. Useful, productive load (intrinsic load) – relevant to achieving learning goals § § determined by the degree of element interactivity depends on specific instructional goals and prior knowledge of the learner (chunking!) o. Wasteful, unproductive load (extraneous load) irrelevant to learning, imposed by the manner in which information is presented to learners and the learning activities required of them § dependent on the design of instruction Intrinsic + Extraneous =Total cognitive load

Efficient learning q. Managing intrinsic (productive) load q. Reducing extraneous (wasteful) cognitive load q. General rule: Do not do anything that gets in the way of learning! q. If intrinsic load is low (simple tasks), there could be no need to reduce extraneous load

References §Sweller, J. , van Merriënboer, J. J. G. , & Paas, F. G. W. C. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10, 251 -296. §Van Merriënboer, J. J. G. , & Sweller, J. (2005). Cognitive load theory and complex learning: Recent developments and future directions. Educational Psychology Review, 17, 147 -177.