UC San Diego

This dissertation consists of four separate but related papers. The overarching theme is how decision makers process information, form beliefs and make decisions. Chapter 1 examines how individuals' beliefs respond to objective information about their ranking on a neutral quality - a meaningless number on a card - or on a quality that has a significant self-image component - intelligence or beauty. For favorable news in the image tasks, subjects respected signal strength and update as "optimistic Bayesians," but they heavily discounted and largely ignored signal strength in processing unfavorable news, leading to noisy posterior beliefs nearly uncorrelated with Bayesian inference. None of these patterns were observed in the control. Chapter 2 uses standard theoretical concepts to examine the optimality of the shooting decisions of National Basketball Association (NBA) players using a rich data set of shots, collected by watching 60 games and recording the shot conditions and outcomes. Shot timing is modeled as an optimal stopping problem and the prediction of a monotonically declining reservation expected shot value" with time remaining on the shot clock is confirmed in the data. For shot allocation, the mixed strategy Nash equilibrium also finds strong support in the data. Chapter 3 studies shot selection in professional basketball to see if players succumbed to the "hot hand fallacy." I find that a majority of the players in the sample significantly changed their behavior in response to hit-streaks by taking more difficult shots, while no player responded to miss-streaks. However, controlling for difficulty, shooting ability is not related to past outcomes. A quarter of the sample suffered a significant loss in shooting efficiency as a result of their mistaken responses. Consistent performance is linked to heightened individual economic incentives. Chapter 4 uses a subjective probability sequence (professional basketball shooting) to see if the gambler's fallacy (GF) and the hot hand fallacy (HHF) can both exist within a single agent, as predicted by theoretical models. Across subjects, there is statistically significant GF after short streaks and HHF after long streaks. On the individual level, a transition from GF to HHF is most common, consistent with the theoretical prediction