Date:         Mon, 13 Apr 1998 11:40:49 -0400
Reply-To:     "Eric B. Pickering" <pickering.14@OSU.EDU>
Sender:       Vanagon Mailing List <Vanagon@vanagon.com>
From:         "Eric B. Pickering" <pickering.14@OSU.EDU>
Subject:      Head survey and stats classes
Comments: To: vanagon@vanagon.com
Content-Type: text/plain; charset="us-ascii"

>Only 40 survey responses thus far.  We need at least 100.  There are 850


>Oh yeah, 100 will make that a statistically meaningful sample of the entire
>vanagon wasserboxer population ever produced.

Not entirely meaningful because it is not truely random!  It may be biased
in that a majority of the responders are those that HAD problems so pay more
attention to it, or those that didn't so are worried and thus paying
attention. Plus, the whole list is a confound in comparing to the gen.
Vanagon population because we probably CARE a bit more, or have more $
(unlikely) to do e-mail, or more time or blah blah blah.  To make it truely
stand up to stiff research criticism one would need to randomly select 100
listees.  Even then you would have to include the caveat: the data is
generalizable to the a list population and not the vanagon population.

REGARDLESS, the survey is definitley worthwhile!  Some data is better than
none, just know its limits!

I'm in a graduate psychology statistics class right, thus I jumped on this
soapbox.

Eric