Nathan writes:

> The article is in the Atlantic Economic Journal (see
> http://springerlink.metapress.com/content/t971556g03w7/?p=a6b760807d0f45e8b1
> da99b5915aa717&pi=6), but it costs $30 to download the article, which is
> more than I am willing to pay. The concern would be what is and what is not
> held constant, to say nothing about the measures they are using to reach
> their conclusion ("a team accrued .07 POINTS and decreased their opponent's
> scoring by .24 GOALS"). Not every team has an enforcer, but those with the
> most prolific scorers (Wayne Gretzky) would have one (Marty McSorley), so
> since only these teams would show up in the data, the results can be skewed.
> Also, teams fight at only certain times - usually when a game is way out of
> hand. So to conclude fights have an influence on the outcome might be
> confusing cause and effect. But if I wanted to pay $30 I could probably find
> out if these things are accounted for in the research.

Why pay $30 to download it from the publisher?  Your library should be able to 
get it for you via interlibrary loan for free.

-- Erik