Browsing M.A. Applied Health Sciences by Subject "Decision-making"
Now showing items 1-1 of 1
An Examination of Decision-Making Biases on Fourth Down in The National Football LeagueThe recent developments in the ﬁeld of sport analytics have given researchers the tools to examine an increasingly diverse set of topics within the world of sport in ways not previously possible (Alamar, 2013; Fry and Ohlmann, 2012). This study analyzes the decision-making processes of high level coaches under diﬀerent contexts and then determines whether or not a speciﬁc subconscious psychological bias, known as the representativeness heuristic, caused the individual to make the choice they did. Past empirical research has examined people’s decisions in diﬀerent contexts and, from those con- texts, made inferences about how those individuals made their decisions and what errors in their decision-making processes could have led to their suboptimal choices (Kahneman and Tversky, 1979; Kobberling and Wakker, 2005; Tom et al, 2007; Tversky and Kahneman, 1992). The representativeness heuristic explains that errors in people’s judgment occur because their mind places too much emphasis on the current situation (new information) and not enough on the original odds (prior information). Previous researchers have been unable to separate the new and prior components of people’s decision-making when studying real-world scenarios in a sport context (Carter and Machol, 1978; Carroll, Palmer, and Thorn, 1989; Carroll et al, 1989; Patel, 2012; Romer, 2006). This research is diﬀerent than the previous related research in that we utilize statistical models to gauge how people weight diﬀerent information when making high-pressure decisions in sport. We hypothesize that coaches are disproportionately weighting new information against prior information when making decisions, and thus, yielding to the representativeness heuristic. To test our hypothesis, we construct numerous Bayesian updating models to represent the impact of National Football League (NFL) coaches’ decision-making on the likelihood of winning games. Utilizing a Bayesian approach enables us to keep the new and prior odds of winning the game separate, and thus, keep the two components of the representativeness heuristic separate. Regression analysis is then used with both of the components to directly test for the representativeness heuristic in NFL coaches’ decision-making by estimating the eﬀect each component has on the coaches’ decisions. These estimates form the basis of our hypothesis tests.