I had a drink with a friend who works in health statistics. She uses SAS at work and asked me what kind of software I use to do my statistics. R and MATLAB, I responded. MATLAB because it’s fast and good and R because it’s free and has heaps of additional packages to extend its use. A few days later she asked me if I’d seen an article on R-bloggers predicting the end of the use of SAS in academic circles, with R overtaking SAS some time in 2015/16.
Even discounting the R package system, the fact remains that R is far less cheaper than SAS or SPSS. As the GFC continues to bite hard and governments, universities and other large institutions look to shed unnecessary costs, perhaps R’s price ($0) will lead to its adoption. Institutional licenses for SAS and SPSS (and let’s throw MINITAB in there, as it’s also used) can’t be cheap and cutting out expensive software when a mature, free statistics environment (R+RStudio) is available would be a very simple way to reduce some ongoing costs. Support is available through companies like Revolution Analytics if the argument is that SAS support the software they sell.
Yes, I’m a bit of an R evangelist, particularly in my research group where people don’t use SAS, Stata, SPSS, MINITAB or MATLAB but instead use Microsoft Excel (one of the worst pieces of software for statistical analysis). I would love to see R displace SAS, SPSS and other proprietary software packages in the next few years, but there’s another parallel objective; the quest to stop scientists using Excel for data analysis and modelling. It’s slow and based on an accounting spreadsheet. If we can get people off Excel and on to R (rather than SAS or SPSS, which would be attractive choices because they’ve heard of them somewhere and because they cost a lot of money they must be good) then academia as a whole will benefit immensely.