Klarinet Archive - Posting 000770.txt from 2003/06

From: ormondtoby@-----.net (Ormondtoby Montoya)
Subj: RE: [kl] Misinterpreted statistic (was Left, Right or mixed
Date: Mon, 23 Jun 2003 00:17:31 -0400

Christy=A0Erickson wrote:

So where do I find a control group for this?

The 'skeleton' of a proper experiment would be:

You would select a random sample of musicians, without biases such as
whether they post on the Internet or not, their genetic background....
etc, etc, etc. Determining that your sample of musicians is truly
random in regard to handedness may not be easy. Can you be sure that
college educated people aren't more or less 'handed' than high school
drop-outs? If you can't be sure, which subset do you want to test?
Or both? Etc etc etc.

You would also decide whether your sample of musicians should include
all types of musician, including perhaps vocalists. Or perhaps you
would sample only musicians who use the individual fingers of both hands
in equal manner (thereby including clarinetists and pianists but
excluding violinists). Or perhaps you would include musicians who use
both hands but not necessarily in equal manner (thereby including the
violinists)..... etc etc etc.

Equally as important, you would sample at random a large population that
wasn't biased towards "musicians" (according to whatever definition you
used for "musician" in the first sample). You would need to decide
whether this sample should exclude "musicians" completely or whether it
should merely exclude a bias toward or against "musicians". Once
again, you would need to avoid other potential 'handedness' biases such
as educational level, genetic background, Internet-poster vs.
Internet-avoider.... etc, etc, etc.

Finally, you would test the Null Hypothesis that there is no difference
between the two samples. If you could show mathematically that the
differences between the two populations are statistically significant
(using the Standard Deviation, acceptance of the Normal Curve or some
other distribution, etc etc etc).

....a lot of work, but anything less is (perhaps) just building another
Urban Legend....

---------------------------------------------------------------------
Klarinet is supported by Woodwind.Org, http://www.woodwind.org/