Numbers Lie: How the Media Failed Floyd Landis

Everyone loves a winner, and no one loves a cheater. When Floyd Landis won the Tour de France, he briefly became the golden boy of the sports world. Unfortunately Landis didn’t get to bask in the glow of his achievement very long. Four days after the tour finished, the UCI announced that abnormal results of Landis’ urine test immediately after an epic comeback on the 17th stage of the Tour de France suggested Landis cheated.

While everyone who is interested in bicycle racing, or merely the Tour de France, now knows that the results of Landis’ follow-up test confirm the results of the original analysis, other than reporting Landis’ testosterone to epitestosterone (T/E) ratio of 11:1, few in the media have reported on the numbers behind the ratio. Mostly, the media’s reaction has been, “Holy Toledo, Batman! Floyd Landis had 11 times the level of testosterone he’s allowed.”

Not so fast, Robin. When taken literally, the test results looks really bad for Landis. Given that athletes must have a T/E ratio of 4:1 or less, it suggests he’s been caught cheating. The thing about numbers and statistics is that when taken out of context they can be used to “prove” just about anything.

Since only a few reporters have written about the numbers behind this result, let’s take a look the information available on floydlandis.com.

On his own web site, Floyd Landis states that the testosterone level in his tests came back in the normal range. However, his epitestosterone level was so low that it skewed the results of the T/E ratio.

Generally speaking, someone who’s cheating would be expected to have a testosterone level that comes back above the normal levels. Cheaters, however, know that the numbers aren’t reported. Only the ratio is reported, so they also take enough epitestosterone to make the ratio appear normal.

It’s now up to Landis and his medical/legal team to explain why his epitestosterone level was so far below normal. The media who’ve reported on this story should certainly be able to find sources who could give possible explanations for how this could occur. As responsible journalists, they owe the public the opportunity to hear the full story, including Landis’ side.

A few, at least, are stepping up to the plate. Lynne Kiesling, on knowledgeproblem.com, addresses whether the doping labs and the media understand statistics.

Carl Bialik, who writes The Numbers Guy column for The Wall Street Journal Online, adds a thoughtful voice to the topic, but says the numbers don’t bode well for Floyd Landis. Still, he gives Landis the benefit of the doubt, and seems to believe that Landis is innocent.

ESPN commentator John Eustice, quoted in an article on lancasteronline.com, stands up for Landis.

Even Phil Liggett, the renowned narrator of professional cycling, believes that Floyd Landis may clear his name.

Sadly, many in the media — pundits, editorial writers and sports commentators, and even hard news reporters — ridicule Floyd Landis as he tries to answer how this result came about. Without examining the details of the case, many in the media have branded Landis a cheater. To my eyes, at least, Landis appears to be an innocent man grasping at straws trying to explain what’s happened.

The results that have been made public are far from conclusive. In fact, the “carbon isotope ratio” test, which is widely reported as proof positive, isn’t necessarily so rock-solid. These results may have come from poor lab technique, improperly cleaned equipment or some other contamination of the sample.

Tom Fine, who’s added a comment to this post, has written a highly informative piece on the carbon isotope ratio test. If you don’t have time to read it, here’s a Cliff Notes-style summary: The test is not quite as black and white as the UCI and the testing lab would have you believe.

Bialik’s article mentions that the carbon isotope ratio test is a very difficult test to perform, and very easy to mess up. Who’s to say the lab didn’t mess up either the test or the interpretation of it?

The announcement by the UCI of Landis’ B results was vague, making only statements but offering no data to back up their “findings.” This, at least to me, is hardly proof positive.

The UCI and the Tour de France organization, in their rush to root out all the cheaters appear willing to discredit and disgrace any and all who have a even the hint of an allegation of cheating about them. They’re certainly willing to drag a man’s name through the mud before he’s even given an opportunity to learn the charges against him and defend himself.

In the media’s rush to get the story out, regardless of whether the full story is presented, a man who may indeed be innocent has been left with his character destroyed, his reputation shattered, his job taken away, and will be known to many as a cheat – regardless of the outcome of his appeals.

When Floyd Landis clears his name, will it be trumpeted with the same fanfare as the stories up to now? I doubt it. Who is going to be able to give him back his good name and reputation? Or the opportunities he’s lost? Or the money he’s going to have to spend on his defense? Or even the ability to make a living?

Back when I was in journalism school, we were taught to understand and verify the facts before we wrote our stories. That means all the facts, not just the ones that an organization such as the UCI finds convenient. It seems that many media outlets have only done half the job in this regard. A ratio is just a couple of numbers, unless it’s placed in the proper context.