home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #26 / NN_1992_26.iso / spool / bit / listserv / statl / 1907 < prev    next >
Encoding:
Internet Message Format  |  1992-11-07  |  1.5 KB
  1. Path: sparky!uunet!stanford.edu!bu.edu!olivea!sgigate!sgiblab!swrinde!zaphod.mps.ohio-state.edu!sdd.hp.com!nigel.msen.com!emory!europa.asd.contel.com!paladin.american.edu!auvm!MIZZOU1.BITNET!SOCLYMAN
  2. From: SOCLYMAN@MIZZOU1.BITNET
  3. Newsgroups: bit.listserv.stat-l
  4. Subject: Precision Probability for Chi Square statistics
  5. Message-ID: <STAT-L%92110520010443@VM1.MCGILL.CA>
  6. Date: 6 Nov 92 00:55:27 GMT
  7. Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
  8. Lines: 26
  9. Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
  10.  
  11.  
  12.   I have a set of chi square statistics (goodness of fit measures)
  13. that I wish to compare.  I have used the PROBCHI function in SAS
  14. (probability = PROBCHI (value, df)) to estimate the probability
  15. of each chi square value.  I've run into the problem that the
  16. function cannot produce a probability value precise enough to
  17. distinguish between the significance of chi square values.
  18. For example,
  19.    chi-square     df      probability (from probchi)
  20.       25.62       1       0.00000
  21.       43.0        3       0.00000
  22.       48.90       4       0.00000
  23.       42.80       5       0.00000
  24.  
  25. Obviously there's not much difference in the "fit" of these
  26. models, but I would like to be able to distinguish among
  27. them, if possible.
  28.   Does anyone know of a way to make probchi give greater
  29. precision, or of an alternative program, or a *book* or
  30. (heaven forfend!) a formula for *calculating* the differences
  31. myself?
  32.  
  33. Katherine Lyman
  34. Dept. of Sociology
  35. U. Missouri-Columbia
  36. soclyman@mizzou1
  37.