-------- Original Message --------
Subject: RE: SAMPLE SIZE CALCULATION
Date: Mon, 1 Jun 2009 20:23:43 -0700 (PDT)
From: F. James Rohlf <[email protected]>
Reply-To: [email protected]
Organization: Stony Brook University
To: [email protected]
References: <[email protected]>
I agree. Estimating the proper sample size for a multivariate study is
not easy. Quite a bit of information is needed. If the purpose is to
test for differences in the means of two or more populations one needs
the following:
1) The assumption of some statistical distribution (e.g., multivariate
normal).
2) An estimate of the within-population covariance matrix. Clearly one
cannot estimate this without collecting some initial data.
3) The selection of the minimum difference that you are interested in
detecting. Difficult to do this without first collecting data on two or
more close groups that you believe are actually distinct.
4) The selection of a type I error rate, e.g., 0.05.
5) The selection of the desired power if the true difference is as small
as the difference specified in 3) above. A common choice is 0.80.
For many studies 3) will be the most difficult to specify.
If you can determine the above then one can use software to compute the
necessary sample size needed for a future study.
If on the other hand you simply plan to do an exploratory study using a
principal components analysis to reveal the patterns of diversity in
your data then there is no real way to estimate an optimal sample size.
------------------------
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: Monday, June 01, 2009 1:25 PM
To: morphmet
Subject: Re: SAMPLE SIZE CALCULATION
-------- Original Message --------
Subject: Re: SAMPLE SIZE CALCULATION
Date: Mon, 1 Jun 2009 10:09:40 -0700 (PDT)
From: Alexander Bjarnason <[email protected]>
To: [email protected]
References: <[email protected]>
Ali
Before one can begin to answer this question I think it is important to
consider what biological question you are asking, why you are using a
geometric morphometric approach, what groups you are hoping to study
and
what is actually out there (if you sampling specimens from collections
rather than specimens at your own disposal).
At what level are you going to be investigating your research question-
sex, population, subspecies, species, genera?
If your question includes the hominid fossil record your sampling is
bound
to much smaller than for extant hominoids. If you are looking at living
primates you will want greater sample sizes if you are comparing
subspecies than if you are comparing genera. You also need to think
about
whether for your particular question you will need more specimens than
landmarks used.
Personally, my major worry is often related to logistics- unless you
have
massive funds at your disposal and equally large amounts of time for
data
collection, you need to carefully consider the tradeoff between what
you
want to do and what you need to do in order to address the questions
you
are interested in.
I wish I could give a more straight forward answer but I don't have
one,
perhaps someone else can.....
-Alex (Bjarnason)
>
>
> -------- Original Message --------
> Subject: SAMPLE SIZE CALCULATION
> Date: Fri, 29 May 2009 05:23:54 -0700 (PDT)
> From: ali mahmood <[email protected]>
> To: [email protected]
>
>
>
>
>
> Dear ALL
>
> I need to calculate my sample size before starting new
geometric
> morphometric study, any one can tell me the steps that I need to
> follow to do this ? . with my regards.
>
>
> yours
>
> Ali
>
>
>
>
> --
> 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
