- I posted this on 8/22  to sci.stat.math.  However, the original
request was posted in three different groups, and the other two 
are both more appropriate for the topic; and have had Replies.

This is a (re)post to sci.stat.edu  and sci.stat.consult.
======= as posted before

On 20 Aug 2003 19:59:17 -0700, [EMAIL PROTECTED] (han) wrote:

> I am very sorry to make a typo. The type 1 error is 0.05. 
> 
> The original document is as below: 
> 
> Freiman and colleagues were interested in assessing whether RCTs with
> negative results had sufficient statistical power to detect a 25% and
> a 50% relative difference between treatment interventions. Their
> review indicated that most of the trials had low power to detect these
> effects: only 7% (5/71) had at least 80% power to detect a 25%
> relative change between treatment groups and that 31% (22/71) had a
> 50% relative change, as statistically significant (alpha=.05, one
> tailed).

I don't know who Freiman is, or the area he is covering.
But for Randomized Clinical Trials (RCTs)  with *human*
subjects, (I suspect that) those comments should be regarded 
as  tendentious and not altogether serious.

Who is making the assumption, or who is implying, that
any large fraction of studies should have power to detect 
"minuscule"  or "tiny"  effects?  25% relative change, such as,
10%  to 12.5%,  would take a HUGE  study.  100%  is 
more believable as a current convention, if you want to use
that terminology.

A change of Survival rates from 20% to 40%  is 100%
improvement, by simple proportion.
Or, it could be a change of odds from  1:4    to 2:3, so,
using that better measure of Odds ratio, it is .25  to .67...
something rather larger than 100%.


 - If Freiman is describing past studies, as he seems to be, 
the fractions of studies with 80%  power is an indicator that 
the studies have *not*  sought to have power for those changes.
Pragmatically, they *were*  treated as being too small to matter.

Now, if he wants to say that there would be a cost-benefit to
seek improvements that are so small as 25%,  then he has
an argument to make.  And he has to be more detailed about
what makes-up  success, too.  But are you going to take an
expensive drug (or other treatment), with unexplored side 
effects, if the best-guess for its efficacy was that it should 
raise your chance of not-relapsing (say)  from 20%  to 25%?   

Further, as to the figuring:  Would that be the same bet, as
wagering on a change from 80 to 85%? 

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
"Taxes are the price we pay for civilization."  Justice Holmes.
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to