I won't extend much beyond the many comments already published in
BMJ.
I used to teach statistics to medical students and faculty when I was
on the faculty of the Medical College of Georgia. I really had to pound
it into their heads that even at p<_0.05 you="you" are="are" going="going" to="to" be="be" wrong="wrong" _5="_5" out="out" of="of" _100="_100" times.="times." p="p"/> There is a rule of thumb (I think it could be formalized if anyone
wants to make the effort)that I (think I) made up: In a dichotomous
randomization trial, your main results
must exceed the delta (i.e. measured change, or differences) obtained from
a nonselected naturally dichotomous variable in the raw data or
subpopulation.
When I looked at the data presented in tables 1, and 2 of Leonard
Leibovici’s paper on prayer in BMJ, I see that his randomization yielded a
sex ratio and some locus of infection differences with distribution
spreads equal to or larger than his principle results.
So, applying my "rule of thumb," to conclude that Leibovici’s prayer
result is "good" we are also forced to conclude that the "focus of
prayer," by implication God, is particularly focused on male urinary tract
infections. I doubt that this is the case. Rather, I believe that male
urinary tract infections are perhaps easier to treat resulting in lower
mortality, and shorter hospital stays. Or even more likely, the
experiment proves nothing at all. Even the hospital stay data are weak on
other grounds.
Competing interests:
No competing interests
05 March 2002
Gary S. Hurd, Ph. D.
Director of Education, Orange County Natural History Museum, CA USA
Rapid Response:
Rule of thumb
I won't extend much beyond the many comments already published in
BMJ.
I used to teach statistics to medical students and faculty when I was
on the faculty of the Medical College of Georgia. I really had to pound
it into their heads that even at p<_0.05 you="you" are="are" going="going" to="to" be="be" wrong="wrong" _5="_5" out="out" of="of" _100="_100" times.="times." p="p"/> There is a rule of thumb (I think it could be formalized if anyone
wants to make the effort)that I (think I) made up: In a dichotomous
randomization trial, your main results
must exceed the delta (i.e. measured change, or differences) obtained from
a nonselected naturally dichotomous variable in the raw data or
subpopulation.
When I looked at the data presented in tables 1, and 2 of Leonard
Leibovici’s paper on prayer in BMJ, I see that his randomization yielded a
sex ratio and some locus of infection differences with distribution
spreads equal to or larger than his principle results.
So, applying my "rule of thumb," to conclude that Leibovici’s prayer
result is "good" we are also forced to conclude that the "focus of
prayer," by implication God, is particularly focused on male urinary tract
infections. I doubt that this is the case. Rather, I believe that male
urinary tract infections are perhaps easier to treat resulting in lower
mortality, and shorter hospital stays. Or even more likely, the
experiment proves nothing at all. Even the hospital stay data are weak on
other grounds.
Competing interests: No competing interests