Experts Don’t Give the Best Advice – Just More of It

Quick Hits
Daily brief research updates from the cognitive sciences

Ok, that is a massive generalisation, and the research didn’t actually look at experts in the sense that we understand it. It is nevertheless insightful and does indeed match some of my personal observations in many fields – from sports to business (I am an expert in a few areas). It also matches other research into the area of how experts are often not as expert as they think themselves.

So, what did these researchers do – and find?

First off, Levari et al. conducted a study with 1’100 participants playing a game called Word Scramble. In this participants are given a board of letters and have 60 seconds to form as many words as possible. Participants played three rounds with different boards. They then asked the participants who they would like to get advice from to learn how to perform better. Unsurprisingly most participants responded that they would like to get advice from the top performers.

That is a logical and natural instinct. If you want to perform better, ask those that are better. But then it gets really interesting.

The researchers then asked 100 of the participants to play a further 6 rounds, then write advice for future players, and rate the quality of their advice. Unsurprisingly the best performers rated their advice as better.

From here another 2’085 participants were then randomly assigned to an advice or no-advice group. Both groups played rounds in total with the advice group receiving advice after their first round. And what happened? The expected. Those that received advice improved quicker and improved with each subsequent piece of advice.

So far so good but when we look at who was giving the advice this is where we get a surprise. Those that were better performers did not give better advice than those who were lower performers. Or this was not reflected in the results. Those participants improved with any advice from anybody. Remember those that had performed better had rated their advice as better. Not so according to the data.

They then got a group of undergraduates who were uninvolved with the study to categorise the advice. For example, it could be ranked as more articulate, or it could be “should” advice, or “should not” advice. And what did they find. There was only one property that was significantly different to the advice that was ranked as best. There was more of it!

This shows that those who were ranked as experts and whose advice was seen as best simply gave the most advice. However, this advice did not correlate with any better performance in contrast to advice from “non-experts”. But interestingly people receiving advice perceive quantity as a sign of quality. Ummm!

There are multiple reasons for both of these effects one is that being an expert does not necessarily mean you can put yourself in the position of a non-expert and formulate more effective advice for them.

And as they showed at the start of the experiment, we do want to receive advice from the best performers “the experts” – it just that this may not be any better than from non-experts. In Word Scramble at least.

Reference

David E. Levari, Daniel T. Gilbert, Timothy D. Wilson. 
Tips From the Top: Do the Best Performers Really Give the Best Advice? 
Psychological Science, 2022; 33 (5): 685
DOI: 10.1177/09567976211054089

More Quick Hits