Stephen Black [mailto:[EMAIL PROTECTED]
> Given some hints in Leibovici's background, my guess is that this is
> a deliberate hoax (note its presence in the special Christmas issue
> of BMJ) intended to provoke discussion. Yet I don't think he
> falsified data. So how he did he do it?
Run the randomization over and over again until you get the results you
want? With "retroactive prayer", that'd be pretty easy to do, right? Especially
if you're just saying one prayer for the list as a whole, as the article says.
I imagine it'd only take a few minutes each time you did it (with most of that
time used in saying the prayer, if you automated the statistical analyses), so
you could do it the fewer than 100 or so times needed to make it likely that
you'd get these results. Randomize, pray, check the results, randomize, pray,
check the results, etc. Stop when you get the ones you want. If you're not too
worried about being honest, it's technically not falsified data (just a failure
to report nonsignificant studies).
I've seen plenty of published "data fishing expeditions" (one IV, 40 or
more DVs, make a big deal about the two or three significant ones) that aren't
really any more honest than this would be, if that's what he did.
Paul Smith
Alverno College
Milwaukee
---
You are currently subscribed to tips as: archive@jab.org
To unsubscribe send a blank email to [EMAIL PROTECTED]
