Wednesday, February 13, 2008

Report Backing Clemens Chooses Its Facts Carefully

So, you didn't think that Roger Clemens and the alleged steroid use from the Mitchell Report would be on a physics blog, did you? Bear with me. This has everything to do with data analysis and statistics, something many of us physicists, especially experimentalists, have to do.

For those who don't know about this (especially those from outside North America or don't follow US baseball), there is a major scancal in the sports of US baseball. The recently released Mitchell report, commissioned by the Major League Baseball association has implicated many high-profile players of using steroids during their professional careers. One of them is Roger Clemens, a well-known baseball pitcher who was well on his way to the baseball Hall of Fame in a few years.

Clemens has vehemently denied such practice. He has gone on the "offensive" of trying to produce "evidence" for his innocence. One of the things he (more likely, his defense team) has done is to produce a set of statistics showing that his performance late in his career (during the period that he was accused of using human growth hormone) is not unusual. So his camp actually tried to produce some quantitative analysis to proclaim his innocence.

This is fine and dandy. Unfortunately, his "data analysis" is being disputed by no less than three Ivy League academicians. Three professors from the University of Pennsylvania has challenged the validity of the analysis and wrote an article in the New York Times to rebutt the conclusion from Clemens statistics. The most damaging conclusion they gather out of a more thorough analysis of the statistics was this statement:

Other measures suggest Clemens performed similarly to his contemporaries. But these comparisons do not provide evidence of his innocence; they simply fail to provide evidence of his guilt.

Our reading is that the available data on Clemens’s career strongly hint that some unusual factors may have been at play in producing his excellent late-career statistics.

In other words, if the Clemens camp was hoping that the statistics show his innocence, they are wrong. It certainly doesn't show that he has used any performance enhancing drugs, but it certainly also can't be used as evidence for his innocence like they had hoped.


No comments: