The Pitfalls of Econometric Tests of the Uncertainty of Outcome Hypothesis: Interdependence of Variables, Imperfect Proxies, and Unstable Parabols
A substantial number of empirical analyses of attendance at team sports events have been devoted to testing the uncertainty of outcome hypothesis, according to which the interest of fans is positively influenced by the degree of uncertainty of an outcome. The results, however, have turned out to be inconclusive. This article examines a possible explanation based on the flaws of the testing method. I show that results consistent with the uncertainty of outcome hypothesis can be obtained even when attendance is solely determined by the quality of the competing teams. The reason for this is the direct relationship between team quality and outcome probabilities. Moreover, while controlling for the quality of teams would solve this problem, one can only use imperfect approximations of unobservable variables. This might lead to results consistent with either the uncertainty of outcome hypothesis or the loss aversion hypothesis, regardless of the true nature of attendance demand.