-------- Original Message --------
Subject: RE: p-values
Date: Fri, 28 Aug 2009 15:42:27 -0700 (PDT)
From: F. James Rohlf <[email protected]>
Reply-To: [email protected]
Organization: Stony Brook University
To: [email protected]
References: <[email protected]>
No rule of thumb but for such a simple calculation and sample sizes of
just 10 and 10 you could use thousands of random permutations very quickly.
What you want to do each time is to take a random split of the 20
observations into two samples of size 10 and then compute your sample
statistic. Note that it does not have to be the t-statistic it can simply be
the difference between the two means. You then count how many times the
observed difference is equal to or larger than those obtained by the random
splits.
------------------------
F. James Rohlf, Distinguished Professor
Ecology & Evolution, Stony Brook University
www: http://life.bio.sunysb.edu/ee/rohlf
-----Original Message-----
From: morphmet [mailto:[email protected]]
Sent: Friday, August 28, 2009 1:58 PM
To: morphmet
Subject: Re: p-values
-------- Original Message --------
Subject: Re: p-values
Date: Fri, 28 Aug 2009 10:28:31 -0700 (PDT)
From: Samor Gandhi <[email protected]>
To: [email protected]
Thank you very much! What I meant with the sample size was about the
following:
If we have two group, in each group 10 sample (i.e. n=20), the t.test
in
R can be used to get run the t-test (and/or Welch), and it gives the
t-value (t-statistic) and the theoretical p-value (from the t-
distribution)! What I am asking about is to get the t-value from the
t.test and run the test let say 100 times. What i wanted to know is
there an rule of thumb that tells how big n1 and n2 (n=n1+n2) should be
to run permutation for each time of these 100 times! So, would 8=10-2
be
good for each permutation! Because what I understand here (I am not a
mathematician) that one should compute first the t-value (i.e. using
t.test in R) and the compute 100 t-values due to permutation and then
see where does my first t-value lies and this would be the p-value! For
example if I have 4.332 as a t-value and this would be equal to the 8th
value in the permutations values (100 values), then my p-value would be
8/100=0.08! Is this correct?
Thank you in advance,
Samor
--- On *Fri, 8/28/09, morphmet
/<[email protected]>/*
wrote:
From: morphmet <[email protected]>
Subject: Re: p-values
To: "morphmet" <[email protected]>
Date: Friday, August 28, 2009, 5:25 PM
-------- Original Message --------
Subject: Re: p-values
Date: Fri, 28 Aug 2009 04:55:24 -0700 (PDT)
From: Dennis E. Slice <[email protected]
</mc/[email protected]>>
To: [email protected]
</mc/[email protected]>
References: <[email protected]
</mc/[email protected]>>
Sokal and Rohlf. 1995. Biometry, 3rd ed.
Section 9.4 - t-test
Box 9.14 - Finding the sample size required and power for
ANOVA
Rohlf and Sokal. 1995. Statistical Tables.
Table B - critical values for t
R Development Core Team (2005). R: A language and environment for
statistical computing. R Foundation for Statistical Computing,
Vienna,
Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.
Function: t.test()
-dslice
morphmet wrote:
>
>
> -------- Original Message --------
> Subject: p-values
> Date: Thu, 27 Aug 2009 15:22:48 -0700 (PDT)
> From: Samor Gandhi <[email protected]
</mc/[email protected]>>
> To: [email protected]
</mc/[email protected]>
>
>
>
> Dear all,
>
> What is the best way to compute the p-vallue from the t-
statisic
> (t-value)? And how big the subsample should be?
>
> Many thanks,
> Samor
>
>
>
-- Dennis E. Slice
Associate Professor
Dept. of Scientific Computing
Florida State University
Dirac Science Library
Tallahassee, FL 32306-4120
-
Guest Professor
Department of Anthropology
University of Vienna
-
Software worth having/learning/using...
Linux (Operating System: Ubuntu, CentOS, openSUSE, etc.)
OpenOffice (Office Suite: http://www.openoffice.org/)
R package (Stats/Graphics environment: http://www.r-project.org/)
Eclipse (Java/C++/etc IDE: http://www.eclipse.org/)
========================================================
-- Replies will be sent to the list.
For more information visit http://www.morphometrics.org
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
