Date:         Sat, 20 Sep 1997 23:21:01 -0400
Reply-To:     Mike Davenport <Mike_Davenport@PPD-PHARMACO.CO.UK>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Mike Davenport <Mike_Davenport@PPD-PHARMACO.CO.UK>
Subject:      Re: Need opinion on Statistical Packages
Content-type: text/plain; charset=US-ASCII

I have been using SAS beginning in 1979. In those days the big problem with
SPSS was that it didn't have the precision of SAS in solving normal
equations for ANOVAs and regressions.  Of course trying to write programs
in english language has always been a negative in my opinion. I have not
used anything for statistics except SAS in the ensuing 18 years. For a
while I needed BMDP 4F to do categorical analysis, but SAS has long since
fixed that issue.
 
For statistics today either S-plus or SAS are probably the two best
packages available.  SPSS, BMDP,  Mini-Tab, etc. are also rans.  I would
point out that SAS IS NOT a relational database. All SAS files are
basically hierarchialy in nature.  Thats why you have SAS/ACCESS for Oracle
and other relational DB products.  SAS is great for statistics (Statistical
Analysis System) and report generation, average to poor for grahics and
poor as a database.
 
Proc mixed which allows a great deal of flexability for GLM is the one of
the best tools I have used for statistical modeling and report generation.
 
good luck
 
mike