home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!sun-barr!ames!pacbell.com!network.ucsd.edu!sdcc12!cs!kc
- From: kc@cs.ucsd.edu (k claffy)
- Newsgroups: sci.math.stat
- Subject: measure of goodness of sample
- Keywords: sampling, estimator, goodness, efficiency
- Message-ID: <37767@sdcc12.ucsd.edu>
- Date: 7 Sep 92 02:45:24 GMT
- Sender: news@sdcc12.ucsd.edu
- Lines: 28
- Nntp-Posting-Host: beowulf.ucsd.edu
-
- Can anyone help me with the problem of
- assigning a "goodness" grade to a given sample
- of a population, for example based on one
- or more of the following:
- the error of the estimated parameter (mean/median),
- the variance of the sample (as compared with the
- variance of the parent population?), and
- taking into account the sample size (so that,
- ceterus paribus, smaller samples get better grades?)
-
- i have Cochran but i can't find any universal
- measure (right, there probably isn't any).
- i know there are standard formulas for variances
- of means of normal populations, for example,
- but i have a fairly bizarre parent population,
- and i want to assign ranks to 20 samples (taken
- with different methods and different
- sampling fractions).
- (i want to use the term "efficiency",
- but i assume someone else has defined this
- and i'm just not reading the right books.
- if so, i wouldn't mind knowing. One of my
- books mentions 'Gaussian efficiency' in passing
- but doesn't define it. argh.)
-
- email appreciated -- i'll summarize if desired,
- thanks much,
- k
-