University of Birmingham > Talks@bham > Human Computer Interaction seminars > Understanding preference reversals as optimal under constraints

Understanding preference reversals as optimal under constraints

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Benjamin Cowan.

The attraction effect occurs when a choice between two alternatives is biased by the addition of a third irrelevant alternative. In a choice between a £500 flight with 3 connections and £600 flight with 1 connection, the addition of an alternative in which there is a £650 flight with 2 connections is likely to bias some people toward the £600 flight. This demonstration of context sensitivity in human decision-making is typically characterised as irrational. This is because value-maximising models of decision making show that the best alternative can only be guaranteed to be chosen if the available alternatives are evaluated independently of one another. However, with a mathematical analysis of computationally rational choice we have shown that when there is uncertainty about the value of each alternative, the third option is not, in fact, irrelevant – it provides information about the likelihood that either of the original options will be best. A consideration of the rank order of choice values leads to the attraction effect and higher expected value than would be achieved if the “irrelevant” alternative were ignored.

The talk will focus on reporting an experiment that tests this theory. It examines the relationship between the proportion of choice trials on which the attraction effect occurs and how easy it is for the participant to distinguish the expected value of the available alternatives. To achieve this, we offered participants choices between gambles in which we systematically manipulated the difference in expected value. As the difference increased so the attraction effect decreased, demonstrating that choice reversal is most likely to occur when expected value rank ordering is most uncertain, that is when it is most likely to be rational.

This talk is part of the Human Computer Interaction seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


Talks@bham, University of Birmingham. Contact Us | Help and Documentation | Privacy and Publicity.
talks@bham is based on from the University of Cambridge.