Readers of the old blog on WordPress will recall that I enlisted BikeMonkey to cover the sports-doping beat a time or two on the old blog. Today's news pried him away from his political
ranting commenting for a guest appearance. -DM
Dave Stoller: "Everybody cheats. I just didn't know".
Professional cyclist Floyd Landis has lost his final appeal of his conviction for testosterone doping during the 2006 Tour de France. Most readers will be familiar with the backstory. If not, click the two prior links and then head on over to the trust but verify blog for the pro-Landis perspective.
I'm motivated to discuss this stuff not just because I follow professional cycling now and again; it has a lot of parallels with science misconduct.
In a brief fan's overview for those too lazy to Google, steadily improving journeyman* / domestique Floyd Landis started to show some real prospects as a Big Tour winner with some big performances as a super-domestique in 2004, and an initial foray as team captain in 2005. Landis was showing excellent signs of class in the early 2006 Tour but the usual Tour deal-breaker of a few great performances from competitors and one bad day had Floyd on the ropes. Stage 17 saw Floyd come out and just slay the competition with an all day break to put himself back in the race he would eventually win. It was a great stage to watch. A desperate attack in the early going which was surely doomed to failure. (This is a common rhythm for the bigger bike race stages- one man is usually unable to hold off the peloton until the finish if the teams are determined to catch him.) Yet Landis did. Despite the fact that the main General Classification teams knew he was riding back into and possibly away with the entire race. They couldn't catch him. Floyd just kept hammering away the kilometers, obviously suffering like a dog and continuing to pour on the power. It was amazing.
Mere days after the race the news broke that Landis had failed a dope test which indicated possible testosterone doping. In a word, cheating. Or, as the immortal Marvelous Mark Slackmeyer put it, "Guilty, guilty, GUILTY!"
I can't do the sports-doping beat anywhere near as well as Kevin Beck, now departed (one hopes not permanently) from the Chimp Refuge. And really, the science behind the likely performance enhancing effects of testosterone and the detection of a doped athlete are old news at this point. What I can do is point out the perhaps obvious parallels with science fraud. I'll leave it up to the reader as to whether solutions struck for sports doping can help with science fraud.
I was originally invited to discuss this sort of business on a science and science-careers blog because of a similarity that PhysioProf has recently identified quite eloquently in a recent post:
Academia, like many professional career enterprises in the United States, is structured in a winner-takes-all format. Many people enter the enterprise at the bottom level, trickle their way upwards, until a relative few make it to the top and enjoy the spoils ... This kind of system is common to academia, professional sports...We Americans have a real love-hate relationship with highly-competitive winner-take-all type systems like this...literally millions of youngsters work their tails off thinking that some day they will be a professional athlete in the "big leagues", and maybe a few thousand make it in one sport or another
Of course where professional cycling is concerned for the American audience, we're talking more like thousands of youngsters working for maybe a few dozen professional slots. Nevertheless, there are interesting similarities including the reality of cheating, the motivations for same, the probability of being detected and the impact of a high-profile cheating conviction.
Today's news is the fact that the long post-detection process of litigating and appealing the Landis doping conviction has ended with the court of last appeal denying Landis an acquittal. From the cyclingnews.org story:
The Court of Arbitration for Sport (CAS) in Switzerland announced today its decision regarding ex-Tour de France winner Floyd Landis vs. US Anti-Doping Agency (USADA). The court ruled against Landis, ending his fight to clear his name and regain his Tour victory. Landis will have to serve the full two-year suspension that is back-dated to January 30, 2007 as that is when he officially declared voluntary non-competition status. Additionally, Landis was ordered to pay $100,000 in costs to the USADA.
In a statement released by the CAS, it found that: "1. The LNDD is a WADA-accredited laboratory which benefits from the presumption that it conducted sample analysis in accordance with international laboratory standards. 2. The athlete has not rebutted this presumption by showing that a departure from the International Standard occurred."
The panel then went on to conclude from the evidence presented that the "presence of exogenous testosterone or its precursors or metabolites in Floyd Landis' sample proved that he violated the anti-doping rules of the UCI".
In short, "guilty". Landis was illegally doped with testosterone on the 17th stage of the 2006 Tour de France. The stage in which he produced a sporting performance and story that would easily have been in the top ten current cycling fan's list of heroic and amazing accomplishments. If he were competing fairly, that is. From the fans' perspective, "disappointed' doesn't even begin to cover it. Totally-and-completely-sucks is getting closer. and then there are the ... problems, shall we say.
Disbelief of the Evidence
In the cases of sports doping, the evidence is frequently indirect because full confessions are rare; Landis continues to assert to this day that he didn't dope with testosterone. The case revolved around a series of analyses suggesting that Landis' ratio of testosterone to epitestosterone in urine indicated an exogenous source and furthermore that a carbon isotope ratio test indicated a synthetic testosterone was in Landis' body. The defense is at least two pronged in suggesting that the actual test samples were handled improperly (thus one could not rule out tampering or accidental contamination) and that the threshold standards for a positive/negative test were applied somewhat arbitrarily.
Is this so different than cases of science fraud? True, in science fraud the results are supposed to be replicable. So the question of a limited resource, like a urine sample being irretrievably contaminated, is supposed to be moot. That's only from a certain perspective however, because in science fraud we recognize the possibility of being wrong, via error, in a way that cannot be evaluated for fault. Take the Ricaurte affair in which the drug vials were either mislabeled by the supplier, screwed up by the lab or used as a bogus excuse by the lab. Take the many instances of bench lab retractions or corrections in which the defensive argument seems to be "it happened once because of some weird confluence of events which we cannot replicate (and oh btw, the basic finding is still true and we'll prove it)". See the Buck/Zou retraction for a recent example. Leading one to then try to guess the probability that said defense is true versus craven excuse-making lying. In many cases our arguments are tarted up by completely bogus stuff from "I know Prof. Smith and she has a good reputation, surely it was an accident", to "I never believe a damn thing that comes from that lab", to "That's the cheater postdoc" to "Maybe that graduate student felt pressured by Dr. Sy C.O. Path to fudge a little" to.... And a lab that routinely admits to sloppy procedure, leaks results by press release, involves itself in the more political aspects of the business...well, I for one raise an eyebrow at this particular anti-dolping lab. Not that they aren't right about Landis. Just pointing out that their results do not always sit well with me because of such lapses. Just as with a scientific lab that shows all kinds of signs of sloppiness even if you don't have any specific reason to distrust one particular dataset.
Then there is the question of debatable standards for evaluating data. There is a level at which this is irrelevant in science- after all, one just puts the data out and makes one's own conclusion. Anyone else is free to disagree on what the data mean, whereas in the sports doping world we expect categorical decisions of cheated/not cheated. Yet we come to find out that in some of these doping cases the decision may rest on a disputed scientific background on just how definitive some tests or decision criteria for doping test results may be. In science, sometimes we may have a little disagreement over the degree to which Photoshopping an image represents fraud. Putting someone's career on the line [h/t: bayblab] makes it really important to get the assignment of culpability right.
Everyone Does It
The concept that everyone is doping thus it is not really cheating against each other is a constant feature of the cycling doping discussions. This is the NASCAR approach in which the ethos of "It's only cheating if you get caught" applies. I must say that as a fan of pro cycling, the past couple of decades, from the mysterious young Dutch cyclist deaths, to the Festina affair, to Operacion Puerto, to the Telekom confessions, to the Bjarne Riis 'fess up, to..... well, there's a good chance the people saying everyone was doped were right. But we don't know. Not for sure. When some guy pops off a fantastic result...he might be clean. Maybe.
This ties back to science fraud and some people's perception that "they all cheat". Where "they" is most usually the scientific elite which I'll just go ahead and shorthand as those who publish regularly in the Glamor Magazines. Deal. I have heard this from more than one person who is around such elite environs, almost exclusively in the context of excoriating their labs' reputed cheaters, the willing blindness on the part of the PI and a hefty suspicion of similar behavior in the competition labs. There are some around here who suspect that cheating and fraud is equivalent all up and down the talent ranks. Bollocks. The motivation to slap a testosterone patch on your nether regions is just plain different between a person challenging for the TdF win in their late career and one who is just happy to be completing their first Tour in their second or third pro year. Different, for that matter, between a solid domestique rider and one who is constantly on the edge of losing his contract if results don't improve. Different for a college athlete who is going to med school next year and one who wants to make the pro cycling ranks (and has no other obvious talents or prospects in hand). Context and contingencies matter!
The trouble is that we have next to zero good quality data available to evaluate the belief that everybody does it in either cycling or academic science. This is a big problem.
Perception is Everything
You can call it framing if you like. I won't care. Another commonality is that failing to address the problems on the ethical and cheating side runs the risk of killing the golden goose. In the case of cycling of losing fans and therefore advertisers and team sponsors. In the case of science, of losing public financial support, yes. But even more pernicious, losing the public respect for the value of what science is telling them. If everyone is faking it and they just haven't been caught yet, why should the public shape personal or public behavior based on what science has to say? Wouldn't that be just as crazy as assuming our latest Tour de France hero raced 6 hours per day for a month clean?
*This is a relative concept. Landis was a professional cyclist, all of whom have demonstrated in the past that they are in a very select group of athletes indeed. For those readers that are unaware, cycling is a team sport in which there are one or two individuals per team in each race who are designated to be the likely challenger for the win. The other half-dozen or so teammates fulfill specific roles in helping that designated individual to win.