We prove that a subtle but substantial bias exists in a common measure of the conditional dependence of present outcomes on streaks of past outcomes in sequential data. The magnitude of this streak selection bias generally decreases as the sequence gets longer, but increases in streak length, and remains substantial for a range of sequence lengths often used in empirical work. We observe that the canonical study in the influential hot hand fallacy literature, along with replications, are vulnerable to the bias. Upon correcting for the bias, we find that the longstanding conclusions of the canonical study are reversed.
The NBA Three-Point Contest has been considered an ideal setting to study the hot hand, as it showcases the elite professional shooters that hot hand beliefs are typically directed towards, but in an environment that eliminates many of the confounds present in game action. We collect 29 years of NBA Three-Point Contest television broadcast data , apply a statistical approach that improves on those of previous studies, and find considerable evidence of hot hand shooting in and across individuals. Our results support fans' and experts' widely held belief in the hot hand among NBA shooters.
The hot hand fallacy has long been considered a massive and widespread cognitive illusion with important implications in economics and finance. We develop a novel empirical strategy to correct for several fundamental limitations in the canonical study and replications, conduct an improved field experiment to test for the hot hand in its original domain (basketball shooting), and gather all extant controlled shooting data. We find strong evidence of hot hand shooting in every dataset, including on the individual level. Also, in a novel study of beliefs, we find that expert observers can predict (out-of-sample) which shooters are hotter.
Multiple attribute search is a central feature of economic life: we consider much more than price when purchasing a home, and more than wage when choosing a job. Nevertheless, while single attribute search problems have been studied extensively, little is known about optimal search in multiple attribute environments. I introduce a partial characterization of optimal sequential search in a problem with multiple searchable attributes and alternatives, no order restrictions on search, and full recall. Upon applying the partial rational benchmark to a rich dataset I find that subjects systematically deviate from optimal sequential search by (1) searching too deeply within alternatives and (2) switching too adjacently between alternatives. Finally, I explore how these deviations affect payoffs, and explain why they may constitute a form of boundedly rational search behavior in which subjects re-optimize only occassionally, while "smoothing" search order so as to make costly memory failures less likely.JEL Classification Numbers: D11; D12; D81; D83.
The NBA Three-Point Contest has been considered an ideal setting to study the hot hand, as it showcases the elite professional shooters that hot hand beliefs are typically directed towards, but in an environment that eliminates many of the confounds present in game action. We collect 29 years of NBA Three-Point Contest television broadcast data , apply a statistical approach that improves on those of previous studies, and find considerable evidence of hot hand shooting in and across individuals. Our results support fans' and experts' widely held belief in the hot hand among NBA shooters.
We show how classic conditional probability puzzles, such as the Monty Hall problem, are intimately related to the recently discovered hot hand selection bias. We explain the connection by way of the principle of restricted choice, an intuitive inferential rule from the card game bridge, which we show is naturally quantified as the updating factor in the odds form of Bayes’s rule. We illustrate how, just as the experimental subject fails to use available information to update correctly when choosing a door in the Monty Hall problem, researchers may neglect analogous information when designing experiments, analyzing data, and interpreting results.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.