Sunday, October 19, 2008

NYT on referee home-field bias in soccer

Carl Bialik points to this "Keeping Score" column from last Sunday's New York Times (registration required), talking about bias among soccer referees.

The article doesn't link to the actual academic articles it cites, but does mention one author, Peter Dawson. I've found two papers on the subject by Dawson and others; this one, which I can't download (if you can, can you send it to me?), and another one, which I can.

The one I was able to view is called "The Influence of Social Pressure and Nationality on Individual Decisions: Evidence from the Behaviour of Referees" (pdf), by Dawson and Stephen Dobson, both British economists. It tries to look for evidence of home-team favoritism among referees, but it's not really structured in such a way that it could find it, whether or not any bias exists.

That's because the paper just tries to figure out how many penalties (yellow and red cards) were given to home and visiting teams. But even if it turns out that visiting teams were called for a lot more infractions than home teams, that wouldn't tell you anything about the referees – because it could be that visiting teams just commit more infractions!

There could be many reasons for this. Perhaps the cheers of the home crowd frustrate the visitors, and they become more aggressive. Perhaps visiting teams trail more often, due to home field advantage, and have to become more physical in an attempt to come from behind. Maybe the visiting players just miss their wives. Who knows?

But it seems reasonable to assume that whatever makes players better at home could easily lead them to commit fewer fouls at home, too. So just counting penalties, I think, doesn't tell you much about the refs.

Still, one of the more interesting findings in the paper is this: if there's a running track in between the football pitch and the crowd, the referee calls more penalties against the home team (statistically significant at 3 SDs). The authors interpret that as meaning that, the farther the officials are from the screaming, potentially hostile fans, the more willing they are to incur their wrath by punishing their team.

It sounds reasonable, except that it's possible that teams that play in those stadiums just happen to be teams that are more aggressive in general. It would have been better to control for that, perhaps by including a variable for penalties taken on the road.

Labels: ,


At Sunday, October 19, 2008 4:41:00 PM, Anonymous Anonymous said...

Back during my grad school days, a colleague and I published an article using a very similar design with NBA basketball (Lehman & Reifman, 1987, Journal of Social Psychology).

We compared the average number of fouls called on players at home and away. However, in an attempt to get around the problem Phil identified -- namely that teams may just play worse on the road -- we divided players into stars and non-stars, based on all-star game appearances and other criteria.

We figured that intense booing from the crowd would result primarily in association with star players. In Chicago, for example, a foul call on Michael Jordan or a non-call on Magic Johnson would likely generate more spectator agitation than calls or non-calls on the Bulls' or Lakers' last guys off the bench. Although a foul on any Bull (or Laker) could affect the outcome of the game, Chicago fans would be particularly concerned about Jordan fouling out.

Indeed, stars were called for fewer fouls at home than away, but there was no difference for non-stars. That doesn't prove officiating bias, but it seems to advance the case beyond a pure home-away comparison.

I have e-mailed a copy of the article to you, Phil.

At Monday, October 20, 2008 3:57:00 PM, Blogger Phil Birnbaum said...

Just to fill in a detail from Alan's paper ...

42 out of 213 players were designated stars. The stars committed 3.05 fouls on the road, 2.43 at home (2.01 SDs difference). The non-stars committed 2.25 fouls on the road, 2.23 at home (obviously not significant).

Leaf fans will find the result very interesting, even though Kerry Fraser's missed call on Wayne Gretzky came while Wayne was on the road.

At Saturday, October 25, 2008 6:37:00 PM, Blogger Julius Q. Vernon said...

Could you update the link to Charlie Pavitt's Sabermetric Research? It is no longer valid. Thx.

At Sunday, October 26, 2008 12:37:00 AM, Blogger Phil Birnbaum said...

Not sure what the new link is. I've e-mailed Charlie ... will update when I hear from him. Thanks!


Post a Comment

<< Home