Date: Fri, 7 Sep 2007 14:20:54 -0000
Reply-To: Shawn Haskell <shawn.haskell@TTU.EDU>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Shawn Haskell <shawn.haskell@TTU.EDU>
Organization: http://groups.google.com
Subject: Re: Neyman-Person paradigm
In-Reply-To: <Pine.GSO.4.58.0709061536150.4849@leukothea>
Content-Type: text/plain; charset="us-ascii"
On Sep 6, 2:41 pm, kv...@EMORY.EDU (Kevin Roland Viel) wrote:
> A collaborator asked me to respond to criticism of his grant. He has
> enrolled 74 people in a study and presented the work as preliminary to
> support the grant. The people have four levels, but he grouped them into
> two. The 2X2 p-value was 0.026. He was able to enroll 18 more (rare
> disease). The reviewer wrote "the addition of 18 more samples is
> statistically inappropriate in a Neyman-Person paradigm for analysis."
>
> I did not read the original application. I think the applicant failed to
> stress that we would rather not group and additional patients are
> necessary since we have not controlled for another exposure which is very
> important. Since this is the case, can we appropriate reply that the NP
> paradigm does not apply; that additional cases must be obtained?
>
> Thank you,
>
> Kevin
>
> Kevin Viel, PhD
> Post-doctoral fellow
> Department of Genetics
> Southwest Foundation for Biomedical Research
> San Antonio, TX 78227
hard to say what's going on with this much info. I assume the
reviewer spoke of Jerzy Neyman and either Egon or Karl Pearson. Egon
collaborated a bit with Neyman who was impressed by Fisher's
likelihood approach. Karl Pearson on the other hand was not so
impressed with Fisher and of course gave us his chi-square
distribution. I am unaware of any "Neyman-Pearson paradigm of
analysis". Maybe someone else will shed some light here, but I
suspect your reviewer knows just enough to be dangerous. Check out
David Salsburg's "The Lady Tasting Tea", or maybe Google or Wiki
Neyman and Pearson to maybe figure out what the reviewer is trying to
say. good luck. SH
| 
