The Wiki Defense

What Floyd Landis taught the press about drug testing

It was a partisan crowd in Lancaster County, Pennsylvania, and they gave their local hero, Floyd Landis, a standing ovation that went on and on. The cyclist came home in March to raise money for his campaign to clear his name. Landis shot to fame in July 2006 as only the third American to win the Tour de France, but became infamous just four days later when it emerged that a urine sample he gave during the race had shown a testosterone ratio outside the allowed range.

To date, the campaign has involved hefty legal fees, a series of public fund-raisers such as that one, the first-ever public hearing of a doping arbitration (scheduled to take place in May), and an unprecedented Internet strategy known as the “Wiki Defense” that is forcing journalists to question the global antidoping operation that they too often treated as foolproof.

When the news of Landis’s test results first broke, the headlines screamed “Doping Scandal.” The next day the sprinter Justin Gatlin announced that he, too, had registered abnormal testosterone ratios in a test in April 2006. Both Landis and Gatlin denied doping. An army of columnists wondered if the public could ever trust sporting achievement again.

But recently Landis has begun generating more nuanced press coverage. The change is largely the result of his Wiki Defense, in which he posted 370 pages of his test documents online in the hope of unearthing experts and explanations for the suspicious result. “Wiki” refers to the open editing systems best embodied by Wikipedia, the online encyclopedia that relies on a wisdom-of-crowds approach to verifying the accuracy of its entries. Landis is the first athlete to use the Internet in this way, and the move sparked a series of debates on blogs and in chat rooms as scientists and other experts responded to the documents. “A scientist from NASA called me up and said, ‘I bet you never thought you’d be hearing from NASA,’ ” Dr. Arnie Baker, Landis’s longtime friend and coach, told the Pennsylvania crowd. “I told him he was the third one from NASA alone.”

Baker is running the science side of the cyclist’s defense and has built a case based on problems with the procedures followed by the laboratory that conducted Landis’s tests. Baker also notes apparent inconsistencies within the antidoping system run by the World Anti-Doping Agency (WADA). In Pennsylvania he presented a slide show of some of the biggest problems with Landis’s test. “Oh, my gosh,” the woman behind me kept exclaiming as Baker ran through what to the scientifically uninformed, myself included, seemed very basic procedural errors that could cast doubt on the ability of the lab that did the tests. For instance, the lab appears to have covered up changes to entries in its documents–a contravention of basic scientific protocol where mistakes are crossed out but left visible for record-keeping purposes. More intriguingly, it appears that different labs accredited by WADA use different criteria in deciding if a test is abnormal or not.

Those are the questions about the antidoping system that have interested the media, but it took Landis’s Wiki prod to get the journalists to challenge their own assumptions.

“There was an assumption that the process was bullet-proof, that you couldn’t fool the lab,” said Michael Hiltzik, a staff writer at the Los Angeles Times. In December he wrote a two-part series called “Presumed Guilty” that examined the internal workings of the anti-doping system. “Because of the Wiki, I’ve been able to take the documents to experts and get them to walk me through it,” said Hiltzik. “If he [Landis] hadn’t posted the documents, there wouldn’t be this emerging body of online discussion that questions the lab process.”

The effect of the Wiki Defense has been to circumvent the mainstream media by generating online debate. Cycling chat forums buzzed with scientific questions and Landis himself, unscripted and initially unbeknownst to his defense team, joined a forum. The debate rages on, but the chatter provided reporters with scientific questions–and expert sources–they might not have otherwise known to ask, or had access to.

Initially, Landis’s victory was a great human-interest story. His Pennsylvania Mennonite upbringing was credited with fostering a strong work ethic and a never-say-die spirit–especially when it emerged that he rode the entire three-week race in pain from a decaying hip that would soon need to be replaced.

It got even better. On Stage 16 of the race Landis had virtually collapsed, losing eight minutes and almost certainly any chance of winning. He attributed this to “bonking”– when an athlete doesn’t eat enough and runs out of gas. The next day, on Stage 17, he powered back into contention in what was universally hailed as one of the greatest performances in the Tour’s 103-year history. But it was the sample he gave on Stage 17 that produced the abnormal result.

Some common errors that turned up in early news reports included assertions that the test was “positive” for testosterone and that it showed abnormally high levels of testosterone. Neither was true. The initial test showed only a ratio of testosterone to epitestosterone (T/E ratio) outside the 4:1 limit set by the World Anti-Doping Agency (most people have roughly a 1:1 ratio). It later emerged that Landis’s testosterone levels were well within normal ranges but that low epitestosterone levels had skewed the ratio. If a test shows an abnormal ratio, further tests are done on the isotopic elements of the testosterone in the sample to determine whether the testosterone is natural or synthetic, the latter being a strong sign that it was deliberately administered.

“Even the isotope test is not as straightforward as it looks,” said Philip Hersh, a veteran sports writer for the Chicago Tribune who has covered doping stories for thirty years. “The Wiki has opened peoples’ eyes to the fact that the science is not infallible; these tests are open to interpretation.”

Landis’s results were released prematurely, which only added to the confusion. Samples are divided into A and B units, and only if an A sample shows something suspicious will the B sample be tested. The rules state that an adverse finding is announced only after the results of the B sample are known, but Landis’s test was made public after just the A sample had been tested. Caught in the media spotlight, Landis offered a list of possible explanations, ranging from the cortisone that he was allowed to take for his damaged hip to a naturally high testosterone level to the beer and whiskey he downed after his Stage 16 collapse. The “explanations” were derided as improbable, even absurd.

“We had no response initially because we didn’t have the information,” said Michael Henson, Landis’s spokesman. “We were learning in real time–we learned of the B sample result from the press. They were getting leaked information from governing bodies and the assumption was that it was all correct, and that it was all that there was.”

Landis himself is more forgiving. “I don’t blame the media,” he told me. “At first the authorities had very little information, and I had even less.

“Although they [the authorities] still made comments on how guilty I was,” he added pointedly.

It is part of the job of journalists to be skeptical, but in this case the risk appears to be that we applied our skepticism unevenly, giving the science (about which we know little) a pass and dismissing the protestations of innocence from a source backed into the proverbial corner (something reporters deal with all the time).

“There’s a presumption for some reason, and journalists have picked up the WADA attitude that every athlete is guilty,” said Hiltzik. “Maybe subconsciously journalists were afraid of looking prejudiced, of being taken. You never want to be seen as taking a guilty person’s word that they’re innocent, so the journalist’s credibility is at stake.”

Landis’s Wiki defense cannot overcome the inherent shortcomings of journalism, namely that we tend to be generalists and susceptible to a pack mentality. But it can, it seems, help us–even force us–to do our jobs better. There will be more doping scandals, of course, but maybe next time the reporters covering it will recall the Landis case (however it turns out) and be a little less quick to render a verdict.

Jennifer Hughes is a reporter at The Financial Times.