Each software, like each composer, has its strong and weak points.
SPSS started as software for surveys, and became a very efficient PC and LAN package even with millions of records. I estimated that about 10K would cover virtually anything I might want to do for survey work, and even data preparation where the work can be done by administrative assistants, checking for basics. Also the graphics and reporting parts of the software were outstanding.
SAS' strength was in the mainframe and Unix arena, particularly in large organizations where code can be centrally stored and shared. In order to run stats on an internal clients data, one only had to get the SAS code already stored for accessing the data. It had a monopoly on FDA work and was particularly strong in having cutting edge algorithms.
95% of the time spent on any data analysis project (from start to finish) is spent on simply getting the data into shape to be analyzed. For this part of the process, SPSS is probably the easiest and cheapest product to use. SPSS also wins hands down for best documentation and online help.
SAS and SPSS can read each other's files. So a combination of Base SPSS with SAS STAT would probably work nicely.
So the best and cheapest solution depends on the size of the organization, the skill level of people involved, and the ultimate goals of what the data reporting is to do.
The real battle in many business organizations is about things like using Excel to analyze hundreds of thousands of records, and Powerpoint as the sole graphic reporting tool.
-------------------------------------------
Elizabeth Smith
-------------------------------------------