Sports Law Blog
All things legal relating
to the sports world...
Saturday, January 15, 2011
Home-field advantage and the umpire analogy

An article by Tobias Moskowitz and L. Jon Wertheim in the new Sports Illustrated (I cannot find it on-line for some reason) examines the cause of home-field advantage in sports. If the study is empirically sound (and I want to down with some empiricists to help me figure out if it is), the results are groundbreaking. Moskowitz and Wertheim argue that home-field advantage is mostly explained by official bias, influenced by a combination of the closeness of the game and the game situation; the home crowd (size, loudness, proximity, and intensity); and limited attention to, or accountabiltiy for, particular decisions. Read the whole thing if you can get it (or it eventually comes on-line).

Briefly, officials conform their calls to social pressure created by the home crowd. Officials use crowd noise  to help them resolve uncertainty in making a call, resulting in more calls going the way the home crowd wants them to go. Studies done for or discussed in the the story showed a range of calls in a range of sports that systematically favor home teams--extra time, fouls, and yellow and red cards in soccer; called (non-swinging) balls and strikes in baseball; close plays on the bases in baseball; traveling in basketball; and penalties and fumbles in football. The psychological effect is more pronounced in well-attended games (according to the story, in 2007, the Italian government ordered teams with deficient security to play games without spectators; 21 games were played in empty stadiums and a study by two economists found dramatic decreases in home-team benefits in fouls, yellow cards, and red cards).

The bias is revealed, in part, by the rise of technology, particularly in football. Visiting teams are more successful in overturning calls favoring the home team, especially where the home team is trailing. In other words officials make mistakes in the home teams' favor more often than they do in visiting teams' favor (although the difference is small). Replay thus has resulted in the narrowing or elimination of the home-team advantage, at least as to turnovers, because some of those erroneous calls are corrected (so maybe I need to rethink my opposition to replay in football). Technology also reveals  that officials get it right most of the time (about 85 % on balls and strikes). But the mistakes they make are not random--they tend to favor the home teams. And, of course, most mistakes are not discoverable or reversible--thus the home-field advantage continues.

I am not sure what to do with the story, which I find fascinating. For starters, I wonder what this tells us about the much-despised umpire analogy. One of my objections has been that the analogy, as used, misrepresents what umpires do. This study supports that thought. Umpires clearly do not just call balls and strikes as a simple, clear, robotic exercise--umpires (and other officials) are human and they and their decisions are subject to outside pressures and influences, such as, essentially, public opinion.
Similarly, critics of the umpire analogy have focused on the outside influences that (everyone who is being honest recognizes) affect judicial decisionmaking--life experience, ideology, politics, empathy, public opinion and pressure--just as outside influences affect umpires. But is there a still more-precise comparison between judicial decisionmaking and officiating, given what this new study shows? Is there a litigation "home team" that systematically gets the benefit of judicial decisions? Perhaps the government (especially in criminal cases) or any other repeat player in litigation? Are judges affected by the (unconscious) need/desire to make the populace happy, just as umpires are similarly affected, and does that affect decisions?

What else can this study tell us about judicial decisionmaking?


I'm pretty sure the study is not only accurate, but is potentially low-balling it. That is, home field advantage is at LEAST mostly officiating. There is a secondary effect in that umpires also help the team or player that is trailing. That is one of the reasons there are so many close basketball games.

I truly look forward to the day heartless machines call balls and strikes, as well as all shots in tennis in or out.

Blogger Wheell -- 1/16/2011 2:07 AM  

I thought the article was very interesting as well. The only section of the piece that I thought was unconvincing was the attempt to debunk the theory that crowd support boosts home team player performance. Admittedly, I'm no empiricist, but I thought that the authors begged the question a bit in that section. For example, when considering the NBA, the authors attempted to isolate crowd support from the myriad of other factors that could influence performance -- such as defense and referee bias -- by focusing solely on home and visitor free throw percentage. However, just because crowd support is difficult to isolate does not mean that it does not play a role at other times during a game. For instance, it is plausible that strong crowd support could motivate the home team to play better defense, even if free throw shooting percentage is unaffected.

In any event, that is a minor criticism, as overall the article was quite persuasive in my mind.

Blogger Nathaniel Grow -- 1/16/2011 11:26 AM  

I'm but a novice at empirical studies, having taken just a few graduate courses in statistical analysis, but the study is completely and utterly bogus. The methodology behind it is questionable from the start, and and article essentially claims that because we know one psychological phenomenon does exist, and our other hypotheses are difficult to examine, we must only accept the hypothesis that relates to the psychological phenomena (the susceptibility of officials to vary calls based on social pressure). There are hundreds of interesting empirical approaches that came to mind when I read the article, none of which the authors employed. It does not appear that they even ran a regression at any point, a truly laughable and fatal omission from any article that seeks to examine a correlation. This is one of those terrible uses of statistics that will mislead people, and nobody will read the scholarly articles that debunk it with ease, as they won't be published in SI. A shame that it was published.

Blogger Dantheman -- 1/20/2011 1:30 PM

Anonymous Anonymous -- 1/20/2011 3:51 PM  

Post a Comment