Stata vs R

[Daniel Sutcliffe]

Hello,

My software of choice when undertaking econometric modelling - or analysis more broadly is Stata. However, I have been asked to look at other options that the team I work for could use - perhaps for dealing with different analytical issues/dataset management.

We have access to R Stata and Python.

I was wondering if people had any thoughts about what the pros and cons were of Stata versus R.

I find that Stata is very helpful when manipulating data prior to analysis - I am not sure how R compares to Stata in this regard - plus there is plenty of online support when using Stata.

For those people who have used both Stata and R - are there any points at which one is better than the other? Has anything been published comparing the two when using very large datasets. The datasets I use are often > 2gb in size.

I do hope this question is OK to post on this forum.

[Carlo Lazzaro]

Daniel:
you might be interesetd in https://www.inwt-statistics.com/read...and-stata.html

[Nick Cox]

I guess you could ask the same question in a R forum.

Here I want to point out that there is plenty of on-line support for R. The centre of gravity seems to have shifted away from R-help to Stack Overflow. The range of views on R support is just as wide as it is about Stata forums, with praise for supposed saints and snark and hate at the other extreme.

R perhaps has a larger group of user-programmers, but seems more of a jungle. The number of R packages I would guess to exceed 10000 but that cuts both ways. (My guess for Stata is say 3000 just as a half-order of magnitude things.) Finding your way to a package that does what you want and is well documented and well maintained is far from trivial in Stata, but it seems harder in R.

R also seems to be dividing into subcults. ggplot2 and the tidyverse stir up lots of passions and prejudices.

But making comments doesn't mean that I am any kind of R expert. I have just used it occasionally and read more about it than I use it.

Carlo's link talks about Stata having a 3 year cycle, which is wrong if trivially so.