exarch
Anti-homeopathy illuminati member
- Joined
- Jul 4, 2003
- Messages
- 7,184
Okay, thanks for the calculation, Linda. However, as I'm having trouble understanding the model, let me ask you this: If all 107 of the homeopathic trials discussed in the 1991 British Medical Journal article had been positive, what overall percentage of a true positive would the model calculate?
Well Rodney, no point in avoiding the real issue here.
I think what you have just found out, as explained by fls, is that there is a 1.1% chance that there is a real positive trial among those 81 listed. In other words, the chance of finding a true positive trial (i.e. one showing an actual effect above placebo taking place) is 1.1% of 81 trials.
Of course, I'm not sure if I'm allowed to make calculations with the numbers like that, but here goes:
1.1% of 81 is just under 0.9, in other words, we're not even sure that just one of those 81 positive trials may be for real.
And on top of that, even if we knew for certain that at least one trial was a true positive for sure, we don't know which one that might be.
So overall, it's far more likely to conclude that even that one potentially true positive trial is just as bogus as the rest of 'em, and just a result of dumb luck making it look like something's happening. Don't forget that the trials making publication are probably the best of a series of trials, so they're already selected because they look better than average.
How does this change your opinion on homeopathy though, Rodney?