Skip to content

Dopey anti-doping tests?

Jim points me to this article by Don Berry, which argues that studies of doping in sports often don’t correctly perform probability calculations.


  1. Kaiser says:

    I saw this one and had several issues with the article. First, Don isn't exactly the neutral observer as he has previously advocated on behalf of an accused athlete, plus in this article, there was passing reference to data supplied by Floyd Landis' defense team.

    The entire article focused solely on the rate of false positives without mention of false negatives. We can't discuss one without the other. My suspicion is that just like HIV screeners and so on, tests were designed to minimize false negatives, and when there are positive findings, those are subject to further scrutiny to minimize false positives. For instance, they use A and B samples, with additional more precise test methodologies applied after the initial positive. These were ignored in the first-order analysis.

    Also, the multiple comparisons issue is important but needs to be addressed with something more than a first-order analysis. Independence between tests is clearly violated.

    While Don's effort to make sure drug testing is valid is applauded, I don't see how more openness can help the cause. If the testers disclose everything they do, then the cheaters can tailor their methods to beat the system. The Balco book gives an inside look at how sophisticated the cheaters are.

    This is clearly an important topic and very hard because the actual accuracy cannot be measured. Marion Jones and others have proven the lengths to which some athletes are willing to lie about their drug use.

  2. ZBicyclist says:

    While following the Landis arbitration I felt that it was a shame there was no way to rule against both parties.

    Clearly, cycling is dirty overall, regardless of what one may think of Landis on a particular day.

    But the lab procedures seem oddly designed. For example, both A and B are tested in the same lab (and in Landis's case, the testers were not entirely different people). Why not have B tested in a different lab — a lab which doesn't have a vested interest in confirming its own findings? For an event like the TdF, there's plenty of money available to do this.

    For another, the rules for determining a positive vary (in terms of how many tests must be failed). In other labs, Landis's A result would not have been flagged as a positive.

    Given this (and a lot of other curiosities) it's easy to be cynical / suspicious about what the real purpose of drug testing is.

    Is it to clean up the sport, or is it to pretend to clean up while really protecting commercial interests?

    For example, what about all those gigantic NFL linemen? Did they all get that way on Campbell's Chunky Soup?

  3. Kaiser says:

    I had some nasty thoughts along those lines while watching the Olympics. Would the Olympic Committee or whatever their name is really want to see loads of positives leading to medalwinners being kicked out every day?

    As for Landis and other accused athletes, they want me to believe that the testers are cheaters rather than the athletes. I don't know why I should, as I haven't seen any evidence of foul play, only innuendo.