I recommended SAS, Stata, and R. R is a good place to start simply because it is free, and the company that I work for hires programmers to do mainly SAS even if they only have experience in R (the thinking is, once you learn one language you can learn any language). So use R to get up to speed and get your foot in the door without having to pay for SAS. But for a lot of government work, SAS is still king.

SAS, definitely. What staying power it has. Used it extensively over 20 years ago. Supported by the SAS Institute in NC.

R. Why is open source a problem? I think of it as a plus.

OP here. Thanks for all the suggestions.

I'm surprised that so many are saying R. The fact that it is open-source concerns me, and I'd love to hear about why this isn't a problem for the PPs who recommend it.

R.

Definitely SAS or STATA - especially if you are in govt. These are what most census programmers would be using. Also, you need to learn "hands on" - I would suggest actually starting a long-term project at work that would require you to use the statistics and do the programming.

SAS and R.

However it is difficult to really get decent at unless you use it often.

SAS, Stata, and R.

SPSS is not a good one to invest time into--nobody under 65 in my gov't office uses it and mgmt is threatening to stop renewing the license. Stata or SAS would be better options. Stata will be easier to learn if you're familiar with SPSS and it is a pretty good all-around package unless you'll be using massive datasets, in which case you should learn SAS.

SAS

Although having experience with any of them helps. It's the analytic logic and data molding that requires skill

sas or/and stata

I'm an analyst for a government office that doesn't use statistical software. I think it would be a marketable skill to have and was wondering if it's better to learn SPSS, SAS, or something else. I learned the basics of SPSS in grad school, but that was a while ago. Thanks for any advice!