home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!howland.reston.ans.net!paladin.american.edu!auvm!KENTVM.BITNET!WINCHELL
- Message-ID: <SAS-L%93010414323258@OHSTVMA.ACS.OHIO-STATE.EDU>
- Newsgroups: bit.listserv.sas-l
- Date: Mon, 4 Jan 1993 14:17:33 EDT
- Reply-To: Janice Winchell <WINCHELL@KENTVM.BITNET>
- Sender: "SAS(r) Discussion" <SAS-L@UGA.BITNET>
- From: Janice Winchell <WINCHELL@KENTVM.BITNET>
- Subject: White Noise test stats in proc spectra
- Lines: 17
-
- Would anyone out there be able to give me information on the two test
- statistics generated by the WHITETEST options on Proc Spectra (SAS ETS 6.07)?
-
- Specifically, when attempting to determine the probability associated with
- either, what should be used as the degrees of freedom ? The number of
- observations in the time series ? Or some combination of the number of
- variables with the number of obs in the time series ? I tried the Davis and
- Bartlett references and did not find them helpful.
-
- Tia
-
- Jan Winchell
- Systems Analyst
- Academic Computer Services
- Kent State University BITNET: WINCHELL@KENTVM
- Kent, Ohio USA 44242 INTERNET: WINCHELL@KENTVM.KENT.EDU
- 216-672-2736
-