--- title: "Statistical software I have used!" description: "This is a bit of a niche entry but it's an issue that's intrigued me for many years. My point here is that the availability of statistical software packages has fundamentally shaped the quantitative side of psychotherapy research over the…" url: "https://www.psyctc.org/psyctc/glossary2/statistical-software-2/" updated: "2026-11-01" category: All OM book glossary entries --- # Statistical software I have used! This is a bit of a niche entry but it's an issue that's intrigued me for many years. My point here is that the availability of statistical software packages has fundamentally shaped the quantitative side of psychotherapy research over the time I've been involved in it (about 40 years). "Well of course" you say! However, what I am suggesting here is that the shaping of the field is not just about making statistical analyses easier and more of them, available, it's that as packages emerge and grow in popularity that this shapes what analyses get done and what get published and that shaping is not entirely benign. #### Details Over the last 40 years I have used: - Something I wrote in [dBase](https://en.wikipedia.org/wiki/DBase) II (an early database management system that I ran under CP/M on an Osbourne II "portable" computer, those were early days, about 1983 I think!) - [SPSS](https://www.psyctc.org/psyctc/glossary2/spss/) (over many of its iterations and many years, it was my real introduction to statistical software as for many, perhaps particularly in the psychological and social science realms). - [BMDP](https://en.wikipedia.org/wiki/BMDP) (Bio-Medical Data package). For me I only used it very briefly as SPSS on personal computers became more available for me. As the name suggests it was fairly "biomedical" in its leanings. - [GLIM](https://en.wikipedia.org/wiki/GLIM_(software)). I can't remember how I got to GLIM but it required a better mathematical brain than mine though I think it helped me realise that the menu or specific function approach of SPSS (and BMDP if I remember correctly) were limiting the way I thought about data. - [GENSTAT](https://en.wikipedia.org/wiki/Genstat). I used this only to be able to use Patrick Slater's principal coordinate analysis (an extension of principal component analysis) for disagreement grids (that's not what Patrick called them and I felt they were too naïvely linear in their way of dissecting disagreement so I never really got into GENSTAT though I could see it was hugely powerful and I think back then (late 1980s?) it was free if you had access to a mainframe that could run it and if the mainframe operators were willing to install it: true at the time for the University of London Computing Centre. - [SAS](https://en.wikipedia.org/wiki/SAS_(software)) and particularly [SAS/IML](https://www.sas.com/en_us/software/iml.html) its matrix handling component which was the first piece of software that allowed me to create new potential analyses despite my very limited mathematical and programming abilities. - [Matlab](https://en.wikipedia.org/wiki/MATLAB). Well only ever using an open source derivative when I wanted to use someone's "real" Matlab code. I think I was actually using [GNU Octave](https://octave.org/). It led to me finding an error in the original code and my first, and still only, credited correction in a paper ... and I've forgotten what the paper was but remember that the authors were very straightforward but also generous about it! I thin I was partly drawn to it because a job change meant I no longer had access to SAS. - [Systat](https://en.wikipedia.org/wiki/SYSTAT_(statistics_package)) (now [Grafiti](https://grafiti.com/)). If I have the right system, I liked that you could mix text with the code to produce analyses and good graphics. So it was the precursor of "weaving" and "knitting" (i.e. mixing text and code so that an entire paper can be created entirely in the package). I think I couldn't afford it but there may have been other reasons I didn't continue with it. - [XLispStat](https://en.wikipedia.org/wiki/XLispStat). I admired XLispStat a lot and it was pioneering open source statistical software but I was never really able to think in the LISP way, I had self-taught myself poor functional programming whereas I think of LISP as "declarative". Oddly enough I think there is something "declarative" about using shiny an R package to create interactive online "apps" which I now use a lot. Shiny has forced me to think a bit more in this mode. - [S-plus](https://en.wikipedia.org/wiki/S-PLUS). This was the turning point but it was commercial and I only had access for a short period but it led me to ... - R, or strictly the [R Project for Statistical Computing](https://www.r-project.org/) and to its huge number of packages which build on the main "core" of R. It is open source so there are no purchase fees so I had found something that I would no longer lose if I changed jobs or academic connections. It also, in many ways combined the "weaving"/"knitting" of Systat with the radical extensibility that I had seen in SAS/IML, Octave and XLispStat though it used a grammar/syntax that I could use in my largely functional way of programming. - I did also use other things at times that offered particular niches of statistical analyses: - HLM, MLM and MLWin (early packages, and continuing, for multilevel or hierarchical linear modelling) - LISREL, AMOS (early, and continuing software for confirmatory factor analysis and more general structural relations modelling) - Winsteps for Item Response Theory (IRT) analyses There were others whose names I don't remember though they would generally have been for niche analyses. For some years now I have vowed only to do things I can do in [R](https://www.psyctc.org/psyctc/glossary2/r-project/) as it is open source, hugely powerful and can read data in an enormous number of formats including from SPSS and other commercial packages. So what? Well, as I created that, very roughly chronological, list I was reminded of how the big packages, currently probably SPSS, Stata and Mplus dominate published quantitative work in our field (though R has a respectable and increasing share now). The huge problem is that SPSS created what I can now see was a "cookbook", unquestioning use of the NHST (Null Hypothesis Significance Testing) paradigm. I have never used Stata but I think it is less constraining but, I think, also, less easy to use with minimal training. Mplus, which again I have never used is specifically for the structural relations modelling and confirmatory factor analysis area and clearly powerful and introducing methods into that area ahead of other systems but it still seems to lead to many papers, at least in our field, driven by fascination with the methods rather than by any questioning of real work implications. See [Statistical software and Actor Network Theory](https://www.psyctc.org/psyctc/glossary2/statistical-software/) for another glossary entry that goes into some of the problems this has created for our fields. I suspect that a sociologically inclined historian (or v.v.) could do a PhD and then a lifetime's work looking at how as systems come and go, and as the currently dominant ones introduce new capabilities, how these drive what is published in our quantitative literature largely through methodolatry and excitement to be seen to be using the latest thing. The use of AI in our field, though not entirely statistical, or sometimes not at all statistical, looks to be the latest iteration of this pattern. #### Try also - [Confirmatory factor analysis (CFA)](https://www.psyctc.org/psyctc/glossary2/confirmatory-factor-analysis/) - [Item Response Theory](https://www.psyctc.org/psyctc/glossary2/item-response-theory-irt/) - [Null Hypothesis Significance Testing (NHST) paradigm](https://www.psyctc.org/psyctc/glossary2/null-hypothesis-significance-testing-nhst-paradigm/) - [R](https://www.psyctc.org/psyctc/glossary2/r-project/) - [SPSS](https://www.psyctc.org/psyctc/glossary2/spss/) - [Statistical software and Actor Network Theory](https://www.psyctc.org/psyctc/glossary2/statistical-software/) #### Chapters Not covered in the OMbook. #### Online resources I can't think of anything currently!! #### Dates First created 20.iii.26, improved 30.x.26 and 1.xi.26.