In a recent Slate column — with a nod to the upcoming Vancouver Olympics — I describe a study by Dartmouth economist Eric Zitzewitz, who analyzed the sometimes shady world of Olympic figure skating. The figure skating community was shaken by a scandal in the 2002 Salt Lake City Games when a Russian judge, together with a bloc of other judges, allegedly colluded to hand victory to a Russian ice dancing pair over their Canadian competitors.

The International Skating Union responded by changing the way competitions are judged. Today, the scores are reported anonymously and only a subset of those are used in the final judging process. It may seem odd, at first, to expect that by removing direct public scrutiny of individual judges and concealing their identities, it would curtail vote trading. But the idea is that anonymity makes it hard to verify that corrupt judges have actually delivered the scores that they’ve promised — no one can tie any individual judge to a score. It’s hard to collude if you can’t tell whether your partner in crime is keeping up his end of the bargain.

Yet the study finds that these attempts at reducing collusion actually had the opposite effect. While it may have become a bit more difficult for corrupt judges to maintain a collusive arrangement, this effect was overwhelmed by the negative impact of reduced public scrutiny. Zitzewitz found that in the new system, having a home-country judge on the anonymous panel boosted a skater’s score by even more than it did under the earlier regime. (His data did not reveal, however, whether the home-country bias came just from the skater’s countryman, or a cabal of colluding judges — they were anonymous, remember.)

These findings are interesting even if you are not a fan of figure skating or sports in general, as they reveal some of the challenges in designing rules to fight corruption. With the right data we could run a similar type of study for, say, highway procurement contracts in California to analyze whether making bids public decreased corruption — because of greater public scrutiny — or increased it by making collusion among bidders that much easier.

The findings also highlight the unexpected effect of well-intentioned efforts to combat corruption. Anonymity could, in theory, have helped to make figure skating a cleaner sport. In reality, things played out very differently. The lesson, then, is that we should be open to experimenting with different means of keeping competition honest, whether there are highway contracts or gold medals at stake. And we need forensic statisticians with the data to see whether our experiments are successful.