Skip to main content

STICERD Psychology and Economics Seminar

Channeled Attention and Stable Errors

Joshua Schwartzstein (Harvard Business School), joint with Tristan Gagnon-Bartsch and Matthew Rabin

Tuesday 21 February 2017 15:30 - 17:00

Due to the onging coronavirus outbreak, many of our seminars and public events this year will continue as online seminars. Please check our website listings and Twitter feed @STICERD_LSE for updates.

About this event

A frequent critique of recent models of ways people misunderstand the world is that people should figure out their mistakes after observing events they thought were extremely unlikely or impossible. This paper develops a framework to provide guidance in assessing when a particular error is likely to be noticed in a given environment, focusing on two criteria. First, we clarify that the notion of “unlikely” that should induce a person noticing inconsistent data to deem her mistaken theory implausible is unlikeliness relative to a compelling alternative theory. Second, and our primary premise, a person may ignore, disregard, or discard information her mistaken theory leads her to deem irrelevant. We propose solution concepts embedding such “channeled attention” that predict when a particular theory might persist indefinitely when a person encodes and analyzes data if and only if it is perceived as having positive value within that theory. For any erroneous theory, a person can selectively attend to information she deems sufficient without noticing anything she would find impossible. We investigate which combinations of errors and situations tend to provoke incidental learning, providing comparative statics on both preferences and information that make erroneous beliefs stable. The paper applies these principles to study the attentional stability of several common errors and psychological biases. We show, for example, how a person might remain naive about her self-control problems—and why full naivete can be more stable than partial naivete. Additionally, when certain errors lead a person to overvalue advice or listen to the wrong people, rich feedback on the quality of advice can in fact increase the stability of those errors relative to cases where feedback is sparse.

STICERD Psychology and Economics seminars are held on Tuesdays in term time at 15:30-17:00, ONLINE, unless specified otherwise.

Seminar organisers: Dr Matthew Levy and Professor Nava Ashraf

For further information please contact Lubala Chibwe, either by email:

Please use this link to subscribe or unsubscribe to STICERD Econometrics mailing list (pes).