home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!ogicse!uwm.edu!daffy!uwvax!uchinews!spssig.spss.com!nichols
- From: nichols@spss.com (David Nichols)
- Newsgroups: sci.math.stat
- Subject: 1) Bootstrap 2) Standard deviation
- Message-ID: <1992Aug14.193227.9029@spss.com>
- Date: 14 Aug 92 19:32:27 GMT
- Article-I.D.: spss.1992Aug14.193227.9029
- Sender: news@spss.com (Net News Admin)
- Organization: SPSS Inc.
- Lines: 18
-
- Bootstrapping refers to a methodology in which error estimates are
- produced by resampling with replacement from a finite set of data,
- in order to produce unbiased or less biased estimates of standard
- errors. This may be a simplistic defintion that can be improved
- upon by the more mathematically oriented, but it's the jist of it.
-
- The reason that n-1 is used in the denominator when calculating
- standard deviations of samples is that the squared standard deviation,
- the variance, is estimated in an unbiased manner by using n-1 instead
- of n. You've estimated a parameter (the mean) in calcuating the sum
- of squared deviations about the mean, so you have one less degree of
- freedom than when you started. Incidentally, either divisor produces
- a biased estimate of the population standard deviation.
- --
- David Nichols Statistical Support Specialist SPSS, Inc.
- Phone: (312) 329-3684 Internet: nichols@spss.com Fax: (312) 329-3657
- *******************************************************************************
- Any correlation between my views and those of SPSS is strictly due to chance.
-