-------- Original Message --------
Subject: Re: number of specimens
Date: Tue, 6 May 2008 09:44:40 -0700 (PDT)
From: Joseph Kunkel <[EMAIL PROTECTED]>
To: [email protected]
CC: Joseph Kunkel <[EMAIL PROTECTED]>
References: <[EMAIL PROTECTED]>
Hamid,
Sample size depends on the significance of your result and if there is
any real difference tested by your design.
I use a sample size of 3 as a minimum for any sample of anything. The
median of 3 protects you from outliers occurring at some small
frequency such as 1 in 20. The probability of seeing 2 outliers in a
sample of three is on the order of 0.05*0.05*0.8.
Next I like a sample of 5 since the range of 5 observations is an
approximate 95% confidence interval.
For populations of wing landmarks I have generally used 30 individuals
and that has allowed me to consistently see differences for instance
between European populations of Gypsy moth males and North American
populations. However a sample size of 30 does not allow me to see any
significant differences between the NA populations of Gypsy moths.
Perhaps there are some-such differences but I would need to increase
my sample size to make the difference significant. I would also have
to have an hypothesis that needed me to discriminate between NA
populations. I have used a sample size of 50 males and 50 females for
populations of Drosophila melanogaster that allowed me to produce a
discriminant score predictor for whether a particular wing came from a
male or female. This predictor allowed me thereafter using that
discriminant score to achieve 95-100% in predictability of sex from
the wing alone. For those experiments the need for a higher certainty
drove my investment in a higher sample size.
Often we do not know what sample size is sufficient until we do an
initial experiment. If we do not get a significant result we then ask
if we are willing to invest the energy or money in increasing the
sample size to further test our hypothesis.
Joe
On May 5, 2008, at 1:44 PM, morphmet wrote:
-------- Original Message --------
Subject: number of specimens
Date: Mon, 5 May 2008 10:41:07 -0700 (PDT)
From: Saber Sadeghi <[EMAIL PROTECTED]>
To: [email protected]
Dear Morphometricean
I am a student and I do not have enough experience in Geometric
morphometrics method.
I want to know how many specimens in a sample is enough
statistically in
this method? and if I have few specimens in each location how I
can statistically take into account several close localities as a
single one and their specimens as a single population?
Thanks a lot in advance
hamid
------------------------------------------------------------------------
Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try
it now.
<http://us.rd.yahoo.com/evt=51733/*http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
>
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
-·. .· ·. .><((((º>·. .· ·. .><((((º>·. .· ·. .><((((º> .··.·
=- =º}}}}}><
Joseph G. Kunkel, Professor
Biology Department
University of Massachusetts Amherst
Amherst MA 01003
http://www.bio.umass.edu/biology/kunkel/
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
