A Cognitive Perpective on How People Learn: Implications for Teaching Geoff Norman, Ph.D.

McMaster University

GOALS

 To explore theories of cognitive psychology related to learning, transfer and problem-solving  To examine implications of these theories for teaching

The Cognitive Perspective

thinking “ The essence of intelligence is less a matter of reasoning and more a matter of knowing a lot about the world ” H.A.Simon, 1989

Teaching MUTES



Memory and



Understanding



Transfer



Exercises



Skills

Some assertions about learning and recall

     Learning and remembering results from assimilation of new knowledge into existing knowledge, and meaning is critical to learning Transfer (applying old knowledge to new situations) doesn ’ t happen easily Structured, planned, practice with multiple examples is key to transfer Experience is critical in everyday and expert performance General skills don ’ t exist – it ’ s all imbedded in knowledge

Learning and Understanding

 Learning is strongly influenced by the meaning .

 If we can understand what we are learning in terms of pre-existing knowledge, better learning and retention results  Meaning is a consequence of the interaction between learner and ‘ to be learned ’

The Jeopardy Challenge

 (Picture removed) of two Jeopardy contestants

Watson

 (Picture removed) of Watson the computer

Watson definition available at Wikipedia http://en.wikipedia.org/wiki/Watson_(computer)#Hardware

Watson can process 500 gigabytes, the equivalent of a million books, per second. Watson has much faster reaction time.

The humans were notified by a light, which took them tenths of a second to perceive. Watson could activate the buzzer within about 8 millisecond.

But when he ’ s wrong he ’ s spectacularly wrong….

 Trebeck:  This U.S. city has two airports named after a World War 2 pilot and a WW2 battle.

 Watson  What is Toronto?

(it ’ s Chicago – Midway and O’Hare)

So Watson, reading a million books a second, and button-pushing in 8 msec., can beat Ken….just.

How can Ken be so gosh-darn good?

CPU speed

Me

and my

iBook

1/5 sec. 1/2000,000,000 sec.

RAM ROM 3 bytes 4,000,000,000 bytes ?inf

250 Gb

We should be less impressed that computers can do about as well as humans than that humans can do as well as computers, given the large architectural disadvantages they suffer from.

Paul Johnson , Medinfo 1977

The Secret Ingredient

………meaning….

Tomorrow, and tomorrow, and tomorrow, Creeps in this petty pace from day to day, To the last syllable of recorded time. And all our yesterdays have lighted fools The way to dusty death. Out, out brief candle! Life ’ s but a walking shadow, a poor player That struts and frets his hour upon the stage, And then is heard no more. It is a tale Told by an idiot: Full of sound and fury Signifying nothing W. Shakespeare, Macbeth, V , v

Sound is walking, stage struts and a tale is heard. No more a poor candle, frets life. A brief idiot, fury and shadow, is in a dusty fool.

drswa gtrus hdrkl opono rluta sflta dnaro lensa bfdoa radit sogfv sonap vfhoe qpofs cpoas

Γικνξ δΦγε βγ ηγδφπιψ ιαβθξπ

Meaning is imposed by the learner and involves an interaction between existing knowledge and new information

The procedure is quite simple. First you arrange things into different groups. Of course, one pile may be sufficient. If you have to go somewhere else due to lack of facilities, this is the next step. It is better to do too few things at once than too many. At first it seems complicated, but soon it just becomes a fact of life. After it ’ s over, you arrange the materials in groups again, then put them in the right place.

Washing Clothes

The procedure is quite simple. First you arrange things into different groups. Of course, one pile may be sufficient. If you have to go somewhere else due to lack of facilities, this is the next step. It is better to do too few things at once than too many. At first it seems complicated, but soon it just becomes a fact of life. After it ’ s over, you arrange the materials in groups again, then put them in the right place.

Evidence of the Role of Meaning

 Chess  Nephrology

How do you get to be a chess master?

Is it: - learning the rules?

- learning to think of more moves and deeper strategy? (process) - learning to think

better

moves? (knowledge)

Recall of Chess Positions

 4 levels of chess player  mid-game positions  5-7 sec exposure

Recall after 5 sec. Exposure (real positions)

25 20 /24 15 10 5 0 <1600 16-2000 20-2350 Skill level >2350

Recall after 5 sec. exposure

25 20 /24 15 10 5 0 <1600 16-2000 20-2350 Skill level >2350 Random Real

It

’

s not just Visual Patterns

 Lab data, nephrology problems  5 research associates  6 students  5 experts

Recall of Nephrology Data

14 12 /20 10 8 6 4 2 0 NOVICE Expertise EXPERT Random Real

Summary

 Remembering for meaningful material is enhanced because there are more links or pathways to the memory trace

Implications for Teachers

How can we, as teachers, help students impose meaning on what they ’ re learning?

Implications for Curriculum

 What are we doing now?

 “ Traditional ”  PBL  Does PBL enhance learning ”  MACRO -- no or maybe  MICRO:    Active Learning Imbedding problem Everyday analogy

Effect of active, problem oriented processing

(Needham & Begg, 1991)  Intro psychology students, 5 classic problems  “ Try to solve these difficult problems ” ( 27% successful) vs.

“ Remember the problem and solution so you can solve some additional problems ” (21% successful)

Effect of Active Problem solving

90 80 30 20 10 0 70 60 50 40 Problem-oriented

Needham & Begg, 1991

M emory oriented

Imbedding Principle in Problem

(Ross & Kilbane, 1997) Practice and Test problems with:  SEQUENTIAL  Principle explanation, then problem example  IMBEDDED  Principle imbedded in problem, explanation as part of problem “ Reversal ” = using original principle incorrectly

Reversal Errors

60 50 40 30 20 10 0 Sequential Imbedded

Analogy in Learning Science

(Donnelly & McDaniel, 1993)  48 students, 12 concepts  Literal description of concept vs. description + analogy in familiar domain  (e.g. pulsar star and lighthouse)  24 MCQs; 4/concept, 12 basic +12 inference

78 76 74 72 86 84 82 80 No analogy Analogy Recall Inference

An application in Medical Education

Pressure and Tension on a Membrane r P T Law of Laplace T = P * r

The

“

weight and string

”

problem

T a T = W / 2 sin(a) W

T = W / 2 sin(alpha) T W T a W

T T t t

Dual Explanations

(Krebs, Dore, Norman, 2006)  Three “ Laws ”  Laplace , Right Heart Strain, Starling  Intervention   Mechanical + Biological Active Comparison vs.

Biological explanation only   Test 9 diagnostic cases Sample -- undergrad psych students

50 45 40 35 30 25 20 15 10 5 0

Percent Correct

Dual Biological Explanation

Implications for Teaching/ Curriculum

 Arrange learning to integrate with prior knowledge    Active learning Problem – based learning Imbed principle in problem  Everyday analogy  Sequencing of concepts

Transfer

using old knowledge to solve new problems

 As teachers, we act as if all the knowledge we impart to students will be available to them to solve problems in the future

 As teachers, we act as if all the knowledge we impart to students will be available to them to solve problems in the future Unfortunately….. it won ’ t

Views of Transfer

General Transfer (1900 1915….) Subjects like Latin, algebra teach general “ habits of mind ”

(disproved by Thorndike, 1913)

Specific transfer (Behaviorism,1910--> Now) Learned concepts can only be transferred if new behavior = old behavior

(disproved by Judd, 1908, Wertheimer, 1959, Pressley 1990)

Intermediate / hybrid transfer Learned concepts can be applied (with difficulty) to new, dissimilar problem situations

 A general wishes to capture a fortress located in the centre of a country. There are many roads radiating from the fortress. All have been mined so that, while small groups of men can pass over the roads safely, a large force will detonate the mines. A full-scale direct attack is therefore impossible. The general ’ s solution is to divide the army into small groups, send each down a different road, and have the groups converge simultaneously on the fortress.

You are a doctor faced with a patient who has a malignant tumour in his stomach. It is impossible to operate on the tumour. X-rays can be used to destroy the tumour. If sufficient rays reach the tumour all at once, the cancer cells will be killed, but surrounding tissue will be damaged as well. How can you arrange the procedure to destroy the tumour cells without severely damaging the surrounding tissue.

Gick & Holyoak, 1980

Transfer and Context Specificity    The initial solution (multiple simultaneous paths) was learned in, and stored with the problem context (fortress and army).

To solve the new problem, must recognize that the old problem was analogous to the new, despite different contexts To recognize analogy, we must recognize similarity in deep structure this rarely happens…..

 Why not just teach them the principle?

 Teach the principle, then give them an example of the principle

“ …during early learning, the principle is only understood in terms of the earlier example… the principle and example are bound together. Even if learners are given the principle or formula, they would use the details of the earlier problem in figuring out how to apply that principle to the current problem ” Brian Ross

Effective Use of Practice Examples

 Multiple examples vs. “ Principle + Example ”  Active Compare and Contrast vs. Separate (Gentner, 2003, Holyoak,1989)

Multiple Examples vs. Principle + Example

  MBA Students , negotiation problem Factor 1 Two cases, implicit principle vs.Principle + Case  Factor 2 Read case and principle (on successive pages) vs. Compare Case and Principle Loewenstein& Gentner, 2003

Effect of Examples and Comparisons

70 60 50 40 30 20 10 0

Gentner, 2003

Compare No Compare Two Cases Case + Principle

Implications for Teaching

 Transfer can be facilitated by use of examples during initial learning  multiple examples > principle + example  compare and contrast  Active search for deep structure

Transfer, examples and practice

 Critical to learning, transfer is the opportunity to see the concept arise in multiple contexts  This can only arise with multiple practical exercises What can we do to enhance the value of practice?

Strategies to Optimize Practice  Mixed vs. Blocked Practice (Hatala, 2002)  Distributed vs. Blocked Practice (Schmidt &Bjork,1992)

What do you need to do stats?

An Observation: With the availability of sophisticated statistical software, the central issue facing the statistics student is “ What test do I use?

” To learn this, students have to see data sets, think of possible strategies, and get feedback

What do you get in stats courses?

 Instructional time occupied by equation proving, formula remembering  Practice at end of chapter of the form: “ Do a t test on these data ”

So when do you do a t test?

At the end of the t test chapter

The solution

Mixed practice

Mixed vs. Blocked Practice

In the face of ambiguous features (which are subject to reinterpretation), and multiple categories, students must learn the features which

discriminate

one category from another, not those which

support

a particular category

Mixed vs. Blocked Practice

Hatala, 2000  ECG Diagnosis -- 3 categories  6 examples / category Blocked Review, then 6 examples/category Mixed Review, 2/category, 12 (4 x 3) practice TEST6 new ECGs

Accuracy -- %

50 45 40 35 30 25 20 15 10 5 0 M ixed Blocked

Timing and Sequence of Learning  Would you rather learn to skate (type, play violin, speak Spanish):  1 hour/day, biweekly, for 60 weeks = 30  1 hour / day for 3 days/wk for 10 wks = 30  3 hours/day, 1 day/week, 10 weeks = 30  6 hours/day, 5 days, 1 week = 30

Massed vs. Distributed Practice  Massed  All learning takes place at one time  Distributed  Learning takes place over multiple occasions

Massed vs. Distributed

(Raman, McLaughlin, 2010) 20 GI residents Nutrition course - 4 hr, one 1/2 day vs. 1 hr. 4 1/2 day Multiple choice test, 0, + 1 wk., + 3 mo.

Massed vs. Distributed 35 30 25 No of items recalled 20 15 10 5 0 Change 0-1 wk Change 0 - 3 mo Condition Distributed Massed

Implications for Teaching

 Practice is critical for learning and transfer    to impose meaning on concepts to overcome “ context specificity ” to enhance transfer  Some practice works better than others  Mixed >> blocked  Distributed >> Blocked

Exercises, Experience and Expertise

The critical role of deliberate practice in acquisition of expertise  Is practice just a matter of learning to apply the rules?

 remember the chess master!!!!

 How long does it take to learn chess?

 To learn the rules --- 10 hr.?

 To become an expert ---10,000 hr. / 10 yr.

 Experts know about 50,000 strategies (Ericsson, 2004)

Age and Skilled Chess Performance  (pic removed) FIDE 1995 ratings graph from

Ericsson and Charness, 1998

How long does it take to learn to play: Violin Field Hockey

 (Pic of Graph removed) Estimated Accumulated Practice Hours compared with Age of Musicians in Years  (Pic of Graph removed) Accumulated Practice compared with Years into Career and Chronological Age

Who do you choose?

 Dr. JS. finished residency last year and was in top 5 on cardiology exam?

 Dr. KT finished residency 10 years ago and was in top 1/3 on cardiology exam?

What does the clinician gain from years of experience?

Years of experience

s

 A challenging diagnostic task…..

 A much easier diagnostic task

 (Picture removed) Cat and Dog comparison sketches  Cat and Dog concepts combined  Picture of Dalmatian dog with black spotted cats

Similarity and recognition of everyday objects

 When we recognize everyday objects, the process is effortless, seemingly unconscious.

 We are not aware that we are eliciting or weighting individual features  The process appears to occur all at once (Gestalt)

Familiar Categories  Rapid, effortless, accurate recognition - despite massive within – category variation - despite no overt understanding of rules Unfamiliar Categories  Slow, effortful, inaccurate recognition  Despite NO within – category variation  Despite an explicit and simple additive rule

Exemplar Theory Medin, Brooks

 Categories consist of a collection of prior instances   identification of category membership based on availability of similar instances Similarity is “ non-analytic ” (not conscious), hence can result from objectively irrelevant features  Ratings of typicality, identification of features, etc. done “ on the fly ” at retrieval

Effect of Similarity

(Allen, Brooks, Norman, 1992)  24 medical students, 6 conditions Learn Rules Practice rules Train Set A (6 x 4) x 5 Train Set B (6 x 4) x 5 Test (9 / 30)

 (Picture removed) different pictures of various areas of same skin condition  Picture of most commonly diagnosed area on the arm

Accuracy by Bias Condition

90 80 30 20 10 0 70 60 50 40 Correct Incorrect Other Bias Corr Bias Incorr

Hatala et al, ECG Interpretation

   Medical students/ Fam Med residents

PRACTICE

(4/4 + 7 filler)   middle aged banker with chest pain OR elderly woman with chest pain 

Anterior M I TEST

( 4 critical + 3 filler)  Middle aged banker 

Left Bundle Branch Block

RESULTS Percent of Diagnoses by Condition

50 45 40 35 30 25 20 15 10 5 0 Bias No bias Correct Diagnosis Prior

CONCLUSIONS The Role of Examples  Categories and concepts are based on our

specific

experience with the world as well as application of rules

Implications for Teaching

 Practice with examples is critical in ambiguous domains  Practice results in a collection of exemplars as a problem-solving resource

What happened to Skills?

 Any measure of “ problem-solving ” , “ reasoning ” , “ critical thinking ” , “ clinical judgment ” , etc. correlates across problems at about 0.1-- 0.3.

 Process measures of the above show no gradient with expertise

Recurring Themes

Learning

Human learning and remembering is critically sensitive to the meaning the learner imposes on the “ to be learned ”

Recurring Themes

Transfer of concepts to new, dissimilar problem situations does not occur effortlessly or frequently  Enhanced by active learning, search for principles, multiple practice examples  Impeded by learning for memory, passive learning, single example

Recurring Themes

Formal conceptual knowledge is insufficient for expertise Experience provides an array of prior examples to draw from and reduce memory load

Recurring Themes

Kinds of Knowledge

Expertise is more a matter of having the right knowledge (both formal and experiential) and being able to mobilize it, than of any general skills

 Thinking depends on specific, context-bound skills and units of knowledge that have little application to other domains….. The case for generalizable, context-independent skills that can be trained in one context and transferred to other domains has proven to be more a case of wishful thinking than hard, empirical evidence. Perkins & Salomon, 1989

Conclusion

“ The problem-solving difficulties of novices can be attributed largely to the inadequacies of their knowledge base and not to limitations in their problem-solving capabilities ” R. Glaser, 1984 We have discussed a number of strategies to improve the knowledge base

The End

Thanks