home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!biosci!DUVM.OCS.DREXEL.EDU!MCCAINKW
- From: MCCAINKW@DUVM.OCS.DREXEL.EDU (Kate McCain)
- Newsgroups: bionet.software
- Subject: Re: re:re:population analysis sw
- Message-ID: <9301242040.AA09160@net.bio.net>
- Date: 24 Jan 93 20:22:39 GMT
- Sender: daemon@net.bio.net
- Distribution: bionet
- Lines: 33
-
-
-
- SAS -- in this context -- is prob. a large set of statistical routines that are
- generally available on mainframe computers (though there are likely to be
- versions for PC's that may not run as large data sets). Equivalents are
- SPSS (Statistical Package for the Social Sciences) and BMDP. EAch of these
- offers a remarkably full range of descriptive statistical routines, parametric
- & non-parametric statistical analyses, multivariate analyses, time series,
- LISREL, etc etc etc. Each has its own "language," data formatting requirements
- and so on. I personally prefer SPSSX (because that was the one I learned first
- and because I find SAS command structure more obscure). SAS offers--in addition
- to the "standard," supported programs--a very large library of goodies that
- they mount but do not support. These are documented in various obscure [8-)]
- manuals; some are of great interest for certain rare applications. Each
- package has its own fans. Each will run on various platforms, though perhaps
- not be available in the same version, as the upgrades are developed and
- distributed. Each has an extensive set of published manuals for novices, power
- users, etc. Many social scientists of my acquaintance are SAS-freaks; many
- biologists prefer BMDP. I have found that the same "functional procedure"
- (e.g. principal components analysis) may be implemented differently in the
- different packages -- in terms of defaults, bells & whistles, options, etc.
- BMDP got a bad rap in the 1970's for poor programming & some statistical
- problems. SAS for a long time would only run on IBM's (I understand) --I
- believe one can actually still enter fortran code; I think they are more alike
- than different now. That was NOT the case in 1983 when I was doing a
- dissertation and needed procedures & features from all three. Very confusing.
-
- There are also some nice standalone statistical packages formicros-- like
- SYSTAT -- and dynamic system modelling software -- like Dynamo/StellaII.
-
- Kate McCain "bibliometrics R us"
- College of Information Studies
- Drexel University mccainkw@duvm.ocs.drexel.edu
-