Hello Andrew,

I know how you feel. Without knowing the type of organisation you have, who will use the system, and what you want to achieve, I cannot give specific advice. As a professional statistician I regard the following modules as desirable.

Basics. -------- Base: Cant do anything without it. Stat: Contains virtually all the stats we need

To make life easier. --------------------- Access to PC file formats: The alternative to this is to write your own macros using DDE. FSP: Important if you want to look at your data. I don't like the alternatives but many people manage without FSP. Insight: Interactive look/see module. Saves time getting a feel for the data. For small data sets you can get good results using Pivot tables, and graphs within Excel. QC: Design and analyis of experiments. Alternatively look for a freeware/shareware utility on the web.

To overcome limitations or develop new methods. ------------------------------------------------ IML: Interactive matrix language. We inherited a couple of systems with customised stats and need IML to replicate what has gone before.

Others ------- Graph: Not so important. My clients do not like SAS graphics they prefer Excel quality graphics. If Excel does not have the graph type I use SigmaPlot. Lab: Useful for people without much statistical experiance. Assist: We have this but I have never used it, (I hope someone does).

The absolute minimum for statistical data analysis is Base, however you will not be able to do anything beyond simple summary statistics. In some cases this is all you need.

My own personal opinion is that a true base system should include Base, FSP, Graph, and Access to main formats used by operating system e.g. Excel, dbase, etc for PC. A Base stats system should include Stats, Insight, IML, and QC. If your data volumes are not too extreme you can easily find equivalent stats software for far less than SAS.

Dave.

Andrew Hoy <ahoy@IHUG.CO.NZ> on 19/02/2001 19:07:56 Please respond to Andrew Hoy <ahoy@IHUG.CO.NZ> To: SAS-L@LISTSERV.UGA.EDU cc: (bcc: David Mcnulty/GB/QUES/ICI) Subject: Questions about buying SAS

The organisation that I work for is currently considering purchasing SAS for use in statistical analysis, but we feel we lack information about what components are actually needed. I would appreciate any advice that people on this newsgroup could give:

Basically the SAS institute here offers "base" SAS and SAS STAT for about $NZ10,000 (+sales tax) a year, but after talking to them it emerged that realistically a $NZ15,000 (+sales tax) package with a number of additional components, was more akin to a true "base" system. We are investigating whether we could realistically get away with just the $10,000 system, and whether these other components are that useful.

For a start, is SAS Graph any better than just importing the data into Excel? The SAS sales person was trying to tell us that version 8 of SAS Graph was as good as, if not better, than Excel for graphing. Do readers here have any views on this?

Secondly, how useful is ACCESS TO PC and FSP? What about ACCESS TO ODBC?

Overall, could a user could get away with just base SAS and SAS STAT without anything else? The data that we will be analysing will be available to us electropnically, so we will not have to do data entry).

Finally, is the most recent version of server SAS as easy to use as SAS version 8.1.

Anyway, thanks for any advice that people can give.

Andrew Hoy, New Zealand.

IMPORTANT NOTICE: This email may be confidential, may be legally privileged, and is for the intended recipient only. Access, disclosure, copying, distribution, or reliance on any of it by anyone else is prohibited and may be a criminal offence. Please delete if obtained in error and email confirmation to the sender.