Examining Competitive Balance in North American Professional Sport Using Generalizability Theory: A Comparison of the Big Four

Abstract

ABSTRACT

The purpose of this study was to use Generalizability Theory to analyze levels of competitive balance in each of the four major professional sports leagues in North America (MLB, NBA, NFL, and NHL) to determine if Generalizability Theory has merit as a measure of competitive balance, if leagues are competitively balanced based on an absolute determination, and to what extent leagues are competitively balanced relative to the other leagues. The study analyzed a 10-year period for each of the four leagues from 2005-2014 and used game-by-game win/loss data to determine competitive balance. A single-facet, crossed design (Teams (T) x Games (G)) was applied for each of the leagues in each of the 10 observed seasons. G-Studies were performed to estimate the percentages of variance associated with each facet and their interaction. D-Studies were then performed to determine if leagues were competitively balanced based on an absolute decision, as well as how each league’s level of competitive balance ranked relative to the other three leagues. The results from the G-Studies showed a majority of the variance for each league came from the interaction term. The D-Studies showed that the NBA was the least competitively balanced of the four leagues and was consequently the only of the four to exhibit an absolute measure of competitive imbalance. MLB was the third most competitively balanced league, while the NHL and NFL were the most competitively balanced leagues. The D-Study results also indicate that Generalizability theory has merit as a method for measuring competitive balance, as the ranking of leagues on levels of competitive balance from the study were comparable to the findings of existing literature utilizing accepted methods for measuring competitive balance.