home *** CD-ROM | disk | FTP | other *** search
- Path: vax.wittenberg.edu!lcrawfor
- From: lcrawfor@wittenberg.edu (Crawf)
- Newsgroups: comp.sys.amiga.applications
- Subject: stat programs
- Date: 24 Jan 96 13:16:50 -0500
- Organization: Wittenberg University, Springfield OH
- Message-ID: <1996Jan24.131650@vax.wittenberg.edu>
- NNTP-Posting-Host: vax.wittenberg.edu
-
-
- I've just read Soper's review of Simple Stat Graph in the Jan 96 issue
- of Amazing Computing. I have no desire to slam SSG specifically, and
- I'm willing to believe it does what it does well.
-
- What I do want to grouse about is the general sorry nature of statistics
- packages for micros. SSG, along with many other programs I have seen,
- does only one-way analysis of variance. Those few that do multi-factor
- ANOVA generally do not do repeated measures. At least in behavioral
- research, it's very common to take multiple (repeated) observations from
- every subject. For example, give one group a drug and give another group
- no drug and give them daily trainB[Cing/testing for several weeks. Those
- daily observations observations are not independent of each other, and
- you need modified statistics to deal with that (e.g., repeated-measures
- ANOVA).
-
- I would *really* love to give up mainframe analyses, but until I get
- a really good ANOVA package, I can't. Even SPSS for Windows makes you
- buy a bunch of the modules/packages at premium prices to get something
- I'd consider both fairly basic and essential. The only package I've seen
- that does what I need is SuperANOVA for the Mac. I can find all of the
- other stuff elsewhere. I refuse to believe it's a coding problem. I knew
- someone who coded a lot of those designs on the old pre-C64 Commodore
- machines (remember the CBM 8032? :-) in BASIC, of all things.
-
- Thanks. I feel much better now. :-)
-
- larry crawford
- lcrawfor@wittenberg.edu
-
- still putting the psycho in psychobiology
-
-
-