This post is the first in what I hope to be a series of posts regarding our profession of “statistical programming.” As I pondered what to write about first, I did a mental map of the journey I’ve taken from formal education to the current job – and what programming tools I’ve used along the way. I thought I’d share a few memories of these tools. I’m interested in comments from readers – what statistical software have you used then or now that you’ve found particularly helpful, unique, or interesting?
In my college career, I recall three main statistical programming tools: SPSS, Minitab, and BMDP.
I can’t remember all the details – it’s been a really, really long time – but I think it was a PDP-11 that was the computing workhorse. I DO remember being truly wowed by SPSS! Being able to do a simple linear regression with a handful of statements was remarkable. Up until that time, I think I did all the arithmetic with a calculator and many sheets of scratch paper – with some stat book always nearby for reference.
I also recall being particularly impressed that Minitab could do the inverse of a matrix in … one line? Is that right? [ I should look this up ] Previously, I was forced to do the work by hand, and it was painful. That’s my Minitab college memory – working with vectors and arrays as objects themselves (quite a bit different than dealing with FORTRAN-based approaches).
BMDP came into the picture – I think I remember this right – because it offered some analysis of variance routines (maybe repeated measures?) that SPSS or Minitab couldn’t handle at that time. It was a pretty big deal!
SAS didn’t enter my world (or vice versa) until my first job – I remember being told to learn to program in SAS on my own – with the help of the manuals (SAS version 82.3, I think – again, is that right?). I was so proud of myself after mastering PROC UNIVARIATE.
Since then, it’s frankly mostly been SAS – with a brief interruption for work with RS/1 ( a statistical programming language – also vector/array based – which seems to have fallen by the wayside). Not surprising, I suppose – I worked for the RS/1 software vendor!
Lately, I’ve been trying to teach myself R – which is heavily used in the academic area and seems to be making a lot of headway in commercial/industrial work.
In our profession, there’s always a frontier – always something to learn.
By the way, bonus points to readers who get the reference in the post’s title! J