Quiz 2 - PSYC4191 Expertise Part. 2

What criticisms have been levelled at Ericsson’s proposal that deliberate practice can account for nearly all variation in skilled performance?

 [Note: you can get full marks if you can describe at least 4 out of the 6 options given in the presentation]

1. Causality issue: results are correlational so can’t infer causality. Natural ability may motivate people to do more deliberate practice (Sternberg, 1996). RCT would allow us to infer causality.

2. Other variables issue: Meta-analysis from Hambrick et al. (2014) argued that Ericsson overstates the importance of deliberate practice, saying it explains all variance in expertise. Hambrick proposes other factors, separate from deliberate practice, that also influence level of expertise achieved.

3. Other variables examples: These other factors include the age at which learning started (some ages better for learning than others) and general abilities like working memory capacity (which appear not to interact with amount of practice).

4. Falsifiability issue #1 (definitional): it’s hard to define deliberate practice (Sternberg, 1996), therefore it’s hard to falsify findings related to its efficacy. E.g. if in a certain field, more deliberate practice does not lead to higher skill level, Ericsson could argue the field needs to establish itself more to achieve expert level training in how best to practice. To make Ericsson’s theory more robust, he would need to define the conditions under which deliberate practice doesn’t work.

5. Falsifiability issue #2 (inadequate practice): If someone hasn’t improved in a skill even with significant deliberate practice, Ericsson could just say they need to practice more. Also a falsifiability problem.

6. Unknowability of expert practice method achieved: Ericsson assumes the experts have determined the most optimal way to practice, but he can’t know this. There could be a better way.

What are the problems associated with identifying a true expert?

Deliberate practice theory relies on genuine experts who have developed the best way to practice, but in some fields (e.g. the arts, clinicians, academics) this is hard to determine with adequate validity and reliability due to an absence of clearly measurable outcomes. In other fields, "expertise” has been brought into question entirely.

What happened in Hodgson’s (2008) study investigating wine expertise (methods, results, conclusions)?

Participants: panels of 4 wine judges in major wine competitions

Methods: Gave them 30 wine samples (some from the same bottle), to test consistency of judges’ ratings.

Results: Majority of judges gave different ratings to samples from the same bottle (though were more consistent if they were giving low ratings).

Conclusions: due to a lack of reliability in their ratings, they cannot claim to have validity in judging wine.

In terms of identifying experts, what are the problems with the dictionary and the legal definition of expertise?

Dictionary problem: defines it as “practiced, skilful, and well-informed,” but doesn’t stipulate to what degree an expert is each of these things.

Legal problem: defines it based on being experienced, but as established (e.g. medical example) experience does not equal expertise.

What’s the problem with defining expertise based on peer-review?

Peers don’t always apply evidence based criteria, e.g. police drivers think skid control is important in driving, but this is not supported by evidence on rate of crashes. Risk of circular definitions: people who value these things say they’re important because they value them.

How do scientific definitions of expertise make it easier to identify true experts?

Scientific definitions make it measurable (good at specific problems in specific domain; producing consistent performance across a range of tasks that are REPRESENTATIVE OF THE DOMAIN). Generally speaking, this makes it easier to measure and quantify, but not always (some domains are tricky).

What claims have been made in terms of how long it takes to become an expert?

Simon and Chase say 10 years intense practice to become chess expert.

Malcolm Gladwell in Outliers said 10,000 hours, based on Ericsson’s study of violinists.

What criticisms have been made of Malcolm Gladwell’s 10,000 hour rule?

The 10,000 hour number was just an average from Ericsson’s study. Some completed far less or far more. It also didn’t mean at 10,000 hours they were world-leading experts.

Ericsson estimates the number is more like 25,000 hours to become among the best violinists in the world.

Steve Faloon became global memory champion in 200 hours due to a lack of competition.

Ericsson makes competition argument: has more to do with how much you practice compared to your competitors. So this number is greater in more popular fields.

What is the capacity of working memory according to Miller, (1956) and according to Cowan (2001)?

Miller: 7 items (± 2)

Cowan: 4

Give your own example of how the same underlying information could be represented as different numbers of items in short term memory depending on how it is organized.

If a string of 28 jumbled letters were presented to me, it would be like memorising 28 individual items. If the same set of letters were re-arranged to reveal they spelled the first names of myself and the other three members of immediate family, I would only need to retain four items in my working memory.

In the context of memory, what is chunking?

Organising a larger number of items into a smaller number of “chunks” to reduce the strain on working memory capacity and be able to more easily recall them.

In the chunking demonstration, why did people remember more of the letters when the gaps between the letters was changed?

When the gaps were changed, what appeared to be random letters became recognisable initialisms or acronyms. This reduced the number of items in each row from 8 or 9 letters, down to the three familiar acronyms.

How does chunking improve performance in a skill?

Reduces cognitive load by decreasing number of items needed to remember. Effectively, this increases capacity of working memory.

Why does chunking become more effective, the more expert someone is in a skill?

Experts can draw-on long term memory to augment their working memory by using knowledge from long term memory to form and recall chunks. The more expert, the more frameworks for chunks are available, and the more readily accessabile these are.

What’s the link between chunking and deliberate practice?

Deliberate practice leads to forming mental representations that exist in long term memory. These representations allow for more effective chunking. More effective chunking generally leads to better performance.

Why is chess so often used in expertise research?

1. Easy and clear in how to define levels of expertise (win rates against each other)

2. Good for experimenting because it’s problem solving in a strict set of parameters

What key advantage do chess computers have over humans?

Can search for more moves more quickly (90 billion moves considered by Deep Blue each turn, 9 billion considered per second).

How can chess experts often beat chess computers, despite this advantage?

Human chess experts don’t need to search all 90 billion options (computers and novices search for more options).

What did de Groot (1946) discover about expert chess players thinking processes during a match?

Experts did not consider a greater number of moves, or consider their moves further turns ahead than novices.

What was the key difference between novice and expert chess players in terms of memory, according to De Groot’s work?

Experts could remember boards to a higher level of accuracy after a quick glance compared to novices.

What did De Groot argue was the key mechanism behind expert chess players’ greater skill?

“Pattern-based retrieval from memory.” Experts can recall more board positions and best next move to take from that position. These positions are familiar in long-term memory from all the games they’ve played.

What was Chase & Simon’s (1973) chess study involving two chess boards, one populated with a real game position (methods, results, conclusions)?

Participants: Expert and novice chess players

Methods: Observed novice and expert players reconstruct a board setup on an adjacent board. Board being copied was visible.

Results: Experts took fewer looks at original board to reconstruct new board.
Conclusions: Each look is a chunk. Fewer looks means fewer chunks needed, suggesting each chunk is bigger/has more info contained. Suggests they’re drawing on representations in long-term memory to augment working memory.

What was Chase & Simon’s (1973) chess study involving chess pieces randomly placed on a board (methods, results, conclusions)?

Participants: Expert and novice chess players.

Methods: Present randomly arranged chess boards for 5 seconds (impossible board states).

Results: Experts did not perform better in recreating these random board states than novices.
Conclusions: MEMORY ADVANTAGE from expertise is very specific to domain and context. Not just improving memory overall.

What activity is the main predictor of chess expertise (Charness et al., 2005)?

Studying games that chess masters play/have played, not time spent playing against others. Studying them involves trying to predict next move the master will make, and if wrong, figuring out why.

What did Simon & Gilmartin (1973) estimate that chess masters had 50,000 of?

Chunks of board states they can recognise.

How do novice and expert chess players differ in how they perceive a chess board, according to Ericsson?

Experts recognise chunks, rather than individual pieces. They see more sophisticated (and in larger chunks) elements, e.g. “lines of force.” Novices more likely to look at individual pieces.

What are mental representations and how do better mental representations give experts their advantage?

Reword this: •In general, mental representations are patterns of information (facts, rules, images, relationships, etc) held in long term memory.

•In experts, mental representations allow large amounts of information to be processed quickly, avoiding working memory limits.

•A high enough quality mental representation can lead to super human feats of skill. For example, in 2011, Marc Lang (Germany) broke a world record by playing 46 chess matches at once without being able to see the boards. He managed 25 wins, 2 losses, and 19 draws.