Road Bike Racing - Doping dilemma: how do you prove a negative?

Humor me here for a moment. If Lance never doped, how could he ever PROVE that? You can't prove a negative. So even if Lance is as pure as the driven snow, a large number of observers will presume him guilty by association.

And that's as big a potential problem as anything else in the entire mess.

Logically you can't. At least universally.

But you can use Occam's Razor (http://en.wikipedia.org/wiki/Occam%27s_razor).

Simplest explanation: Lance didn't dope

Complex explanation: Lance used a highly sophisticated technology to avoid all of those doping test and paid everyone that involved in his doping scheme huge amount of money for them to keep quiet.

Can't be done.

It's also part of the interesting problem people with non-negative doping tests face. Once the B sample comes back positive, they bear the burden of explaining the results. This is why we need really, really good doping tests.

Proof is a word rarely used in scientific circles. I basically ignore anyone who uses the word; supports, confirmation, evidence of… these are the preferred terms.

Most lab tests are done on a 95% confidence interval (+/-2SD), so one in 20 tests are “wrong.” Test results that elicit a strong emotional response or that have high clinical relevance are requested to be closer to 99% (eg HIV tests). Normally the stats are improved just by doing replicate testing.

Proof is a word rarely used in scientific circles. I basically ignore anyone who uses the word; supports, confirmation, evidence of… these are the preferred terms.


One of the hardest things to get across students is that they will find it difficult to "prove" anything, much less to do so in a 10-15 page paper. That is one of my biggest pet peaves...

Proof is a word rarely used in scientific circles. I basically ignore anyone who uses the word; supports, confirmation, evidence of… these are the preferred terms.

Most lab tests are done on a 95% confidence interval (+/-2SD), so one in 20 tests are “wrong.” Test results that elicit a strong emotional response or that have high clinical relevance are requested to be closer to 99% (eg HIV tests). Normally the stats are improved just by doing replicate testing.

excellent. you should post this in every one of the doping threads that are currently running

The methods that are being used for testing samples are quite sophisticated and accurate. Even if the accuracy is not 100%, it is close to that. With 95% per sample and multiple samples, that becomes almost 100 %.

False positives are very rare and, although almost every accused rider claims that something must have gone wrong with the test, that is just a normal reaction of the person caught with their pants down.

Can you prove a negative: you can prove that you have not been using a particular substance at the given time. That is simple. Not that you have never been using it, that is much harder to prove.

In my mind, and I wish I knew as much about cycling as I do about chemistry, when someone is tested positive, the picture is clear, unfortunately very clear.

It depends on who has the burden of proof. If Lance has the burden of proof to show he didn't dope, very difficult. But if the burden of proof is on someone or an agency that accuses him of doping, then all he has to do is show that they haven't met their burden of proof. Without eye witness evidence or test results showing he doped, then he wins because the accuser hasn't proved his case.

Thats why in the court of public opinion, spin meisters will never let their client get into a situation where he or she has to prove they are innocent.

False positives are very rare and, although almost every accused rider claims that something must have gone wrong with the test, that is just a normal reaction of the person caught with their pants down.

Or the reaction of a person who was only doing what everyone else was doing... but they slipped up on the doses and got a positive.

isn't this why people (in the American criminal justice system) are supposed to be innocent until proven guilty, and not the other way around?

The methods that are being used for testing samples are quite sophisticated and accurate. Even if the accuracy is not 100%, it is close to that. With 95% per sample and multiple samples, that becomes almost 100 %.

Considering that of all the athletes currently on the chopping block for doping only one (Landis) had a positive test, I think it's quite safe to say that your claim of near-100% accuracy for doping tests is complete bullsh!t.

But of course, Lance's favorite line is that he "never tested positive." :rolleyes:

The methods that are being used for testing samples are quite sophisticated and accurate. Even if the accuracy is not 100%, it is close to that.
I love how you are so precise with exactly how accurate the tests are. Would you care to to share your definition of "close"? I couldn't find it in any of my science texts. Also, how about a source for this statement, or did it just come to you in a flash of brilliance?

Logically you can't. At least universally.

But you can use Occam's Razor (http://en.wikipedia.org/wiki/Occam%27s_razor).

Simplest explanation: Lance didn't dope

Complex explanation: Lance used a highly sophisticated technology to avoid all of those doping test and paid everyone that involved in his doping scheme huge amount of money for them to keep quiet.

Simplest explanation: Santa Claus brings presents through your chimney every Christmas.

Complex explanation: After being manufactured in various places around the world, toys are strategically situated in retail outlets where they are purchased by parents who subsequently package them in colorful paper and place them under the Christmas tree.

:rolleyes:

Simplest explanation: Lance didn't dope

Zabel, Riis, and Basso have confessed to doping and had never tested positive.

Zabel, Riis, and Basso have confessed to doping and had never tested positive.

Thats because they have magic fairies who slant the results. Fairy wee and magic dust changes everything.

Considering that of all the athletes currently on the chopping block for doping only one (Landis) had a positive test, I think it's quite safe to say that your claim of near-100% accuracy for doping tests is complete bullsh!t.

But of course, Lance's favorite line is that he "never tested positive." :rolleyes:


Once again, one can prove that they have not been using a particular substance at the given time. False positives are rare, but false negatives are much more common.

I love how you are so precise with exactly how accurate the tests are. Would you care to to share your definition of "close"? I couldn't find it in any of my science texts. Also, how about a source for this statement, or did it just come to you in a flash of brilliance?

Try math books: 1 minus 0.05 (error) to the power of n approaches 1.0 asymptoticly.

Once again, one can prove that they have not been using a particular substance at the given time. False positives are rare, but false negatives are much more common.

So basically you're freely admitting that the tests being used to screen for doping are garbage. Thanks for taking care of that for me. Saves me the whole long explanation of sensitivity and specificity...

So basically you're freely admitting that the tests being used to screen for doping are garbage. Thanks for taking care of that for me. Saves me the whole long explanation of sensitivity and specificity...

All what I am trying to say is that if one has been tested positive, it is quite (see the discussion above for definition of quite) certain that they have been using the stuff. I guess one can see that as garbage too, but that garbage proves a lot.

Or the reaction of a person who was only doing what everyone else was doing... but they slipped up on the doses and got a positive.

Can't agree more.

Simplest explanation: Santa Claus brings presents through your chimney every Christmas.

Complex explanation: After being manufactured in various places around the world, toys are strategically situated in retail outlets where they are purchased by parents who subsequently package them in colorful paper and place them under the Christmas tree.

:rolleyes:

LMAO!

All what I am trying to say is that if one has been tested positive, it is quite (see the discussion above for definition of quite) certain that they have been using the stuff. I guess one can see that as garbage too, but that garbage proves a lot.

Here's a quote from a very recent issue of Forensic Science International, talking about T:E ratio for detection of testosterone use.


Four hundred and thirty-two negative, respectively 88 positive, samples were collected on the 28 male athletes who participated to the clinical studies. On these data, a population-based limit at 4.0 for the T/E makes 24 false positives for 34 true positives, i.e. a positive predictive value (PPV) of only 59%. A high PPV is particularly important for the evaluation of the evidence in forensic sciences, because the PPV indicates the true probability that the subject is culpable in case of a positive test [21] and [22].

I guess some would consider 59% to be "near 100%..."

Here's a quote from a very recent issue of Forensic Science International, talking about T:E ratio for detection of testosterone use.



I guess some would consider 59% to be "near 100%..."

What I find astoundingly funny is that if many were the subjects of a drug test, and their jobs depended on the results of such test, perspectives might be a bit different.

Suspended without pay until we get around to figuring out if your positive result was really true. And we are really in no hurry to do so. Whenever...

Hey, 59% is closer than, say, 58%...:D

Here's a quote from a very recent issue of Forensic Science International, talking about T:E ratio for detection of testosterone use.



I guess some would consider 59% to be "near 100%..."

Clinical Chemistry 43: 731-735, 1997;
This Article

Our goal in this study was to determine whether the urinary ratio of testosterone to luteinizing hormone (T/LH) as an indicator of exogenous anabolic steroid (AS) use is superior to the urinary ratio of testosterone to epitestosterone (T/E). After 2 weekly placebo injections, 19 subjects were given testosterone cypionate (TC) injections of 250 or 500 mg/week for 14 weeks followed by 14 weekly placebo injections. Patients were considered to have ceased taking TC if they tested negative 9 weeks after their last injection. For detection of illicit or supraphysiological TC (AS) use, the urinary T/E ratio of 6 yielded a false-negative rate of 46% and a false-positive rate of 4%. However, a urinary T/LH ratio of 30 produced a false-negative rate of only 24% and a false-positive rate of 13%. We conclude that the urinary T/LH ratio of 30 is a more sensitive marker of AS use than the urinary T/E ratio of 6 and remains sensitive for twice as long as urinary T/E.

What I find astoundingly funny is that if many were the subjects of a drug test, and their jobs depended on the results of such test, perspectives might be a bit different.

Suspended without pay until we get around to figuring out if your positive result was really true. And we are really in no hurry to do so. Whenever...

Hey, 59% is closer than, say, 58%...:D

I am sorry, but in case of pro athletes, much more depends on their use of drugs than just their jobs: The message they send to young people starting in the sport, etc...

BTW, most of us can be subjects to a drug test and our jobs do depend of the result. Didn't you know that before?

Clinical Chemistry 43: 731-735, 1997;
This Article

Our goal in this study was to determine whether the urinary ratio of testosterone to luteinizing hormone (T/LH) as an indicator of exogenous anabolic steroid (AS) use is superior to the urinary ratio of testosterone to epitestosterone (T/E). After 2 weekly placebo injections, 19 subjects were given testosterone cypionate (TC) injections of 250 or 500 mg/week for 14 weeks followed by 14 weekly placebo injections. Patients were considered to have ceased taking TC if they tested negative 9 weeks after their last injection. For detection of illicit or supraphysiological TC (AS) use, the urinary T/E ratio of 6 yielded a false-negative rate of 46% and a false-positive rate of 4%. However, a urinary T/LH ratio of 30 produced a false-negative rate of only 24% and a false-positive rate of 13%. We conclude that the urinary T/LH ratio of 30 is a more sensitive marker of AS use than the urinary T/E ratio of 6 and remains sensitive for twice as long as urinary T/E.

And if cyclists used anabolic steroids, this would be relevant to the topic at hand.

Also, by raising the "positive" cutoff to 6:1 for the T:E ratio, you lose sensitivity and get more false negatives.

The best screening test is one with a high sensitivity, i.e. the one that gives you the most true positives in the presence of doping. Only then, after a positive SCREENING result, do you go the highly specific confirmatory test, where you want the lowest possible false positive rate, or the highest true-negative rate (specificity).

This all started because you used the term "accuracy," which is 1. loaded and 2. muddies the waters when talking about effective lab tests.

Zabel, Riis, and Basso have confessed to doping and had never tested positive.

Same with Virenque in 1998 !

Humor me here for a moment. If Lance never doped, how could he ever PROVE that? You can't prove a negative. So even if Lance is as pure as the driven snow, a large number of observers will presume him guilty by association.

And that's as big a potential problem as anything else in the entire mess.

At least in the US, Armstrong doesn't appear to have this problem. Even in light of the recent unpleasantness, I'd bet a significant majority of the American public doesn't believe Armstrong doped.

Once again, one can prove that they have not been using a particular substance at the given time. False positives are rare, but false negatives are much more common.

That depends entirely on where you draw the line for a positive test. But with where it seems the lines are drawn for drug use in sport I'd agree false positives are far far rarer than false negatives.

isn't this why people (in the American criminal justice system) are supposed to be innocent until proven guilty, and not the other way around?It's not just the American (criminal) justice system but logic, statistics, and most formal thinking systems. In logic the party making an assertion (doping) must prove that assertion. In statistics you must allow for mistakes and choose to either making false positives (type I error) or false negatives (type II errors)- it is recommended to have false positives less often. This thinking also shows up in folk wisdom- err on the side of caution.
Once again, one can prove that they have not been using a particular substance at the given time. False positives are rare, but false negatives are much more common.False negatives are less damaging and preferable than false positives. Additionally, how can "one" prove "that they have not been using a particular substance at the given time"? If an athlete has a false positive in lab tests then what other methods can prove their innocence- short of 24 hour film surveillance or documented utter isolation?
At least in the US, Armstrong doesn't appear to have this problem. Even in light of the recent unpleasantness, I'd bet a significant majority of the American public doesn't believe Armstrong doped.I bet the significant majority of the American public doesn't care if Armstrong was doping or not and that many would have to really think in order to remember who Armstrong was. Many people believe Landis was convicted and stripped of his win in July and have long since moved on to watching other entertainment.

At least in the US, Armstrong doesn't appear to have this problem. Even in light of the recent unpleasantness, I'd bet a significant majority of the American public doesn't believe Armstrong doped.

I don't care if every other cyclist/team mate Lance ever associated with confesses to doping during Lance's TdF period. It would serve ZERO purpose for him to come clean, moreover it would undermine one of the more effective cancer fundraising charities in the world (Lance Armstrong Foundation).

We all know cycling has a problem (along with ALL sports on every level), and has for many years. We knew that before Riis admitted to doping. That didn't serve a purpose either in my view. I'm sure Riss fessed up as the noose got tighter around his proverbial neck.

I think it's a fair assumption that there is a reasonable possibility Lance used EPO, but if he did it was on a level playing field where all his key competitors (and their supporting casts) were also doping. The emerging picture is that anyone who DID dope in pro cycling during this era would appear to have had little choice if they wanted a podium spot or a JOB for that matter. The notion that this was some dastardly evil ethical lapse by these riders is ridiculous. Where is the outrage about the UCI turning a relative blind eye in what was an obvious doping culture? This appears to have been a rider's choice between 1995 and 2006 (or still today perhaps):

•Dope and ride up to your potential against other doping cyclists

•Ride clean and finish 18th instead of 1st-3rd or if you were a supporting rider go find a day job

I know what I would have done, in a heartbeat. Please. This is PRO $PORT$, not the National Honor Society, the Vienna Boy's Choir or the Doc Ray School of Delusional Ethics. You can take all that 'great tradition of cycling' crap and blow it our of your Toupe while you survey the global TV ratings for the Tour de France along with all the marketing execs. At the end of the day it's a friggin business on many levels.

You want clean cycling? That's the UCI's job. The Honor System only works in... wait, it actually doesn't work in anything.

Why is there more doping in cycling? That's easy. Because it's the HARDEST SPORT. Duh.