Posted by Larry Hoover on January 29, 2009, at 7:55:44
In reply to Re: Stats...ooops » Larry Hoover, posted by SLS on January 28, 2009, at 18:29:53
> Oh, yeah. I bet if clinical trials were allowed to run for more than 6 weeks, many of your placebo responders would "relapse", thereby separating active compound from placebo by a larger margin at the completion of the study.
They do have some data on that. There are two words they use for it, and I can only think of one. A continuation trial is what you're looking for, if you want to do a keyword search. I also recall perhaps two meta-analyses of continuation trial data. Again, if I recall correctly, placebo responders are well over twice as likely to relapse, and to relapse sooner than medication responders.
> I have no statistical background, so I must rely on intuition. Lame, I know. I understand much of what you explain, but without having the tools myself, I cannot analyze things from the same perspective on my own. Thanks for all your help.
>
>
> - ScottYou're welcome for the help. I don't mind contributing whatever I can. I know we have an audience, so I try to speak to the whole. I try to be thorough without being overwhelming. Whenever I can find an opportunity, I try to explain concepts in more than one way.
Over time, with practise, you get better and better at it. I can eyeball a chart and get a feel for what the stats should be like, but that wasn't always the case, certainly. It does get my gall up when I see someone using selected statistics to misrepresent a dataset. I try to be balanced in my responses, but I fear my emotion makes me appear overzealous. Anyway, now I'm way off on a tangent.
Take care,
Lar
poster:Larry Hoover
thread:876214
URL: http://www.dr-bob.org/babble/20090104/msgs/876953.html