QUOTE(one-mint-julich @ Jul 9 2007, 08:16 PM)
In the first place, these are pro cyclists, not controls. We have no reason to believe that some if not all of them were not on some doping program when the measurements were made. Even if they were not doping, their training/racing program would be expected to result in considerably more variation, up and down, than for non-athletes. The same could be said for Floyd, too, except in his case we know that one measurement came just before the TDF, the other during it, a relationship that, as SM has emphasized, should tend to result in a decreased second value.
This is team CSC, you know, one of the three with a really comprehensive and public anti-doping program. If they aren't clean, who is?
In the second place, the CSC results, even if they were typical for non-athletes, don't show all that much variation. Measurements were made every two weeks over a period of six months. Assuming random variability of the values, the spread of values for any particular rider should be about +/- 1.5 SDs. For all of the riders that I can see, except for one very anomalous appearing value for one rider (that was addressed in this forum earlier), the spread is at most about 5 units. So a typical rider has a mean HT of roughly 43%, and a SD of 1.7.
I didn't do any calculations, but practically every single rider has at least one and generally more than one where the HCT changes by more than 3 units between consecutive measurments out of 7-8 measurments. So,as for the odds of any two consecutive measurements showing a change as large as three, I'd say they're pretty darn good, i.e. around a quarter to a third of the time.
Given that typical value, what are the odds that two consecutive measurements would differ by 3.4 HT units? That is two SDs, so the odds are about 20:1. And indeed, if you look at the graph, very rarely, as far as I can tell in this confusing tangle of lines, do two consecutive measurements for one rider differ by this much.
Are you looking at the same graph I am? For almost every rider, I can pick out at least one point that is more than 3 different from one of its immediate neighbors.
Newly Minted Assistant Mechanic. Me, Floyd and Lance--we have 8 TdF titles between us!