In the last few weeks I’ve been using Stata a lot. For simple (and no so simple) tasks I think it is great.
So, the question is: should I quit R?
If Stata does everything you need and you find it easier to use than R, then why isn't the answer to your question self-evident?
Well?
In the last few weeks I’ve been using Stata a lot. For simple (and no so simple) tasks I think it is great.
So, the question is: should I quit R?
Whatever makes your work easier. R is free though, but if you get a license through work it does not matter.
For teaching, Python if you want your students to get jerbs.Are there actually people teaching metrics in Python?
Yes.
Where?
Any language may be used for econometrics. All you need is to be able to solve X'X * beta = X'y somewhow. Statsmodels has that built-in and a lot more, including summary displays very similar to Stata and R. If you want to take a hardcore approach you could do this from first principles using LAPACK in C or Fortran.
Any language may be used for econometrics. All you need is to be able to solve X'X * beta = X'y somewhow.
That's how you find a point estimate. How do you get standard errors, F-tests, CIs, residuals, etc.? Are you really going to teach students to hand-code all of that?
S.T.A.T.A will be obsolete in the next five years.Even Fortran is still around. But I hope the stiffer competition lowers its price.
That will never happen because nobody actually buys Stata themselves, the same way nobody buys their own office desk. Hence LRMs will continue to use it and teach it to undergrads and masters, who then have to jump in at the deep end when they enter their industry data analysist jobs.
Any language may be used for econometrics. All you need is to be able to solve X'X * beta = X'y somewhow.That's how you find a point estimate. How do you get standard errors, F-tests, CIs, residuals, etc.? Are you really going to teach students to hand-code all of that?
Statsmodels does all of that automatically. But learning how to do all of that manually is time well spent I think.