oh yeah, randomizing 30 guys to three groups is 'significantly' different. Okay, give me 20 guys and the two groups I suggest and use the same randomizing algorithm to divide them. Or use the simple count off method 1...2...3...1..2..3. Ones to the left, twos stay here, and threes to the right. I'd bet the results would be damn close to the same as what I suggested with just two people. All the randomizing did was to say they didn't look at genetic disposition or other factors that might influence the study. In my example that didn't happen either. Don't try to over emphasize what is a minute issue in comparing these things. The truth is that they took 30 experienced swimmers of unknown quality and did a study to measure improvement. I've coached swimmers who went from over a minute in a 50 down to :33 and less in just over two months. I certainly didn't hand pick my swimmers, they were given to me. I'd say it was random. So was it the right tool, my suberb coaching skills? I could claim that, but I won't. These were Special Olympic participants without any quality coaching prior to me working with them. Without knowledge of what you start with how do you know? Perhaps there is more to this study than what the article says, but in reading the article at the link Brian put out , they don't include anything to suggest that what I have said is wrong.

__________________
Just in case you weren't sure, it's official -- CMAS sucks.
Oh yeah, and now they're broke.