[R] power? -- Wait a minute!

[Bert Gunter]

Whoa -- wait a minute here! The poster said that "one of the regressors" had
a P-value of .01," and asked if this was believable. One of how many? -- 3?
300? What about multiplicity? How was the regression model selected -- P
Values are essentially meaningless when computed **after** model selection.
And what does the design look like? -- is the regressor highly correlated
with others? 

So I would say the short answer is that the .01 P value is meaningless
without further information.

The longer answer is a rant on the misuse and (ir?)relevance of P values
hypothesis tests, and the like in scientific work, which I will spare you
all from. However, for those who may wonder about such ravings, have a look
at:

http://www.philosophynow.org/issue74/74pigliucci.htm


Cheers,

Bert Gunter
Genentech Nonclinical Biostatistics
 
 

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On
Behalf Of SNN
Sent: Wednesday, October 07, 2009 7:31 AM
To: r-help at r-project.org
Subject: Re: [R] power?


Thanks Simon for the help. 



Simon Blomberg-4 wrote:
> 
> The short answer is Yes. If you reject the null hypothesis based on that
> p-value, then by definition you had enough power to do that. This is
> because there is a precise inverse relationship between the p-value and
> the "observed" power, once you fix the effect size and the sample size.
> In other words, your post-hoc power analysis would be a simple
> re-statement of the p-value. There is no extra information that can be
> gained from such an analysis. See:
> 
> The American Statistician, February 2001, Vol. 55, No. 1, pp 19-24
> 
> Don't bother with your power analysis, unless you are planning a new
> experiment.
> 
> Simon.
> 
> On Tue, 2009-10-06 at 13:49 -0700, SNN wrote:
>> Hi,
>> 
>> I have used multiple linear regression on a data set and one if the
>> regressor was significant with a p-value =0.01
>> 
>> I need to calculate the power for a multiple linear regression. i.e. do I
>> have enough power to believe the above p-value?
>> 
>> 
>> 
>>  
> 
> 
> -- 
> Simon Blomberg, BSc (Hons), PhD, MAppStat. 
> Lecturer and Consultant Statistician 
> School of Biological Sciences
> The University of Queensland 
> St. Lucia Queensland 4072 
> Australia
> Room 320 Goddard Building (8)
> T: +61 7 3365 2506
> http://www.uq.edu.au/~uqsblomb
> email: S.Blomberg1_at_uq.edu.au
> 
> Policies:
> 1.  I will NOT analyse your data for you.
> 2.  Your deadline is your problem.
> 
> Statistics is the grammar of science - Karl Pearson
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
> 
> 

-- 
View this message in context:
http://www.nabble.com/power--tp25776305p25787766.html
Sent from the R help mailing list archive at Nabble.com.

______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.