How to we make snap decisions that are more accurate than well thought out evidence-based ones.

It should have been a quite short book because the answer is that we don’t. At least more often than we would like to think. Gladwell says that the biggest problem is that these conscious decisions are locked away from our conscious minds. But the problem is that they are often inaccurate.

This is something that Daniel Kahneman addresses in Thinking, Fast and Slow. You can develop excellent intuitive thinking. But only only certain circumstances, those being where you get immediate feedback. For example if an anaesthesiologist “gets a hunch” that something is wrong you should probably listen to them; if a oncologist or psychiatrist gets one, you should probably ignore then. That is because the anaesthesiologist will find out if their hunch is correct almost immediately, the others will not find out for months, maybe ever, and so never get to refine this decision-making process.

This struck me in Gladwell’s discussion of a marriage analyst who claims he can tell if a couple will divorce just based on overhearing a conversation in a restaurant. Realistically, he doesn’t know if he can actually do that. How could he? He’ll never get feedback on whether that couple did eventually break up or not.

His discussion of biases and prejudices was interesting. He discussed what he called the Warren Harding error. Harding “looked like a president” but when he got to the Whitehouse, it turned out he was rubbish at it. Conversely, the way orchestras started recruiting more women was to do blind auditions (where the musician would be behind a screen) to prevent people from judging them based on what they looked like, for example, do they have a pair of tits.

He also discussed a heart attack decision-tree adopted by Cook County Hospital. This simple decision tree was actually more accurate than doctors were at diagnosing whether people were having a heart attack. Gladwell claimed that people can be overloaded by having too much information and this actually impairs their decision making.

This is something that struck a chord with me. It is one of the things that we have found in my work in predicting sporting outcomes is that too much information can cloud your models. In fact it is pretty much what Nate Silver’s book The Signal and The Noise is about. And the answer is the same – do some logistic regression.




Don't have time to check my blog? Get a weekly email with all the new posts. This is my personal blog, so obviously it is 100% spam free.



This entry was posted on Friday, December 5th, 2014 at 11:27 am and is filed under Books. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.