If you are at a university, it is likely that they would have a license for SPSS. IMO no other package comes close in terms of human factors. I am a psychologist as well as a math statistician. Full documentation of meta-data, widespread availability of help, ease of use for the portions of analysis like data checking, preparation, description, that make up a subjective 90% of person time on a project are very important to me. YMMV.
The base module covers a wide variety of basic stat. Many specialized packages can read the data files from it or you can save files in the formats for different packages. Also if you need stats that are more in the developmental stage, R procedures can be called directly from SPSS without having to deal with the data prep in R.
SPSS can be expensive to purchase, but having spent many years at the General Accounting Office, I focus on total cost of ownership. In my experience the savings in staff time are many times the purchase/license price.
Also, if you are entering your data in excel, SPSS can read many kinds of files including EXCEL. When you enter your data, I strongly suggest that you fill in most of the information in the data view. IMO that meta-data is important for you to understand your data and for communicating to others about it.
In addition if you want to get into the internals of the stat, all of the algorithms are available for free.
Art Kendall Social Research Consultants
On 6/28/2012 9:21 AM, djh wrote: > Art - > > 1. Agreed on Ray Koopman - he's one of a kind because incredible > patience and incredible smarts don't usually go together .... > > 2. Thanks so much for reminding me I'm on shaky ground with Excel, > even 2010 with the add-in Dr Vidmar suggested. > > Since I'm reasonably competent in PERL, but not R, I fully intend to > install PDL-stats: > > http://pdl-stats.sourceforge.net/ > > as soon as I have a chance. (My problem right now is that I'm a one- > man operation and simple data collection, even as I've recently > automated it "to the nines", has precluded my establishing a > sufficiently robust software base.) > > In the interim, the results I get in Excel will (I hope) at least > provide a large number of very clear-cut successes and failures so > that: i) Ray can see that the overall approach is (or is not) > "working"; ii) run the less-clear-cut cases himself. >