home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.stat
- Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!wupost!sdd.hp.com!usc!rpi!newsserver.pixel.kodak.com!psinntp!psinntp!isc-newsserver!jsvrc
- From: jsvrc@rc.rit.edu (J A Stephen Viggiano)
- Subject: Re: Fwd: Standard Deviation.
- Message-ID: <1992Aug16.145724.11081@ultb.isc.rit.edu>
- Sender: news@ultb.isc.rit.edu (USENET News System)
- Nntp-Posting-Host: rcmain.rc.rit.edu
- Organization: RIT Research Corp
- References: <1992Aug14.172833.11844@cbfsb.cb.att.com> <seX2yRq00Uh785H2EB@andrew.cmu.edu>
- Date: Sun, 16 Aug 1992 14:57:24 GMT
- Lines: 22
-
- In article <seX2yRq00Uh785H2EB@andrew.cmu.edu> dr3u+@andrew.cmu.edu (Daniel Read) writes:
- >---------- Forwarded message begins here ----------
- >
- >Can someone explain why calculating the Standard Deviation (SD),
- >for small samples, with (n-1) in the denominator is better than
- >doing so with (n) in the denominator? I'm sure that there's
- >a perfectly good reason for doing so. But we, lowly engineers
- >aren't usually told the reason. Thanks now, for your response later.;-)
-
- This is done in only in the case of the _sample_ standard deviation. The
- reason for this is that the _sample_ mean is used in the formula, rather
- than the population mean. Using a denominator of (n-1) results in the sample
- _variance_ being unbiased -- the expected value of the sample variance will
- be the population variance. (Unfortunately, the sample standard deviation
- will be slightly biased. It is difficult to come up with an unbiased sample
- standard deviation; I imagine that the square root make the results
- dependent on the underlying distribution. So the goal is to make the sample
- _variance_ unbiased.)
-
- For a proof of this, check any introductory text on mathematical statistics.
- If you'd like to try it yourself, take the expected value of the sample
- variance. Expand the (x - xbar)**2 term inside.
-