Although I'd prefer not to trigger a "Holy War" (see, e.g., "Word v. WordPerfect" or, perhaps, "Tastes Great v. Less Filling") I invite comments, suggestions, observations, etc. relating to the various stats software packages on the market. In particular, two specific discussion threads strike me as possibly useful. First, which packages work the best for empirical legal scholarship and why. Second, what would you advise a colleague (or student) making their initial stats software selection.
To be sure, products like Stats Transfer make stats software selections far less critical. That said, I analogize the stats software selection to a law prof's casebook selection. Due to a necessary front-end investment of time and energy path dependency tends to explain a lot of post-initial decision behavior. Also, for many (most?) initial software selections pivot on whatever their professors used/required for statistical analysis courses in college or graduate (and, increasingly, law) school.
Two final points. First, a fairly thorough and objective comparison of the Stata, SPSS, and SAS packages is found here. Second, in the interest of full disclosure, insofar as I use both Stata and SPSS on a daily basis I have no "dog in this fight."