Monday, January 28, 2013

Choices over Time

Ariel Rubinstein sent me an email with a new paper of his, that is related to a topic I blogged about earlier, and which I think is understudied. His paper "Response Time and Decision Making: A “Free” Experimental Study" has the abstract

Response time is used to interpret results in decision problems. A correlation is found between low response time and choices which are clearly a mistake. The lack of correlation between low response time and behavior which is inconsistent with some familiar theories is interpreted as evidence that such phenomena are not anomalies and cannot simply be labeled as mistakes. Furthermore, the behavior of slow versus fast responders is dramatically different in some problems. It is suggested that a typology of slow and fast responders may be more useful in some cases than standard typologies."

Personally, I think that making the connection between fast and slow responses in a style as championed by Marina Agranov, Andrew Caplin and Chloe Tergiman in "Naive Play and the Process of Choice in Guessing
Games" may be even more illuminating. There players play against a group of people ho already made their choice, and then have 3 min to make a choice themselves. The important feature is that within those three minutes players can adjust their choices, and eventually, one second is randomly chosen, and the choice made at that second is implemented. This also allows players to take more time, but does not free them to leave faster just because they decided faster. And it allows to observe how choices change as players think longer.

For more on that topic, see my former post: Fast and Behavioral versus Slow and Rational

If anyone has more information on that topic, or is running experiments on this, I'd be glad to hear of it.