Jeremy Miles reports
on a free version of SPSS.
The GNU (Gnu’s Not Unix: http://www.gnu.org) organisation was set up in 1984 by the Free Software Foundation (http://www.fsf.org) to develop a free operating system and a collection of software to go with this operating system. One of the latest programs, one which is being developed by Ben Pfaff, was originally referred to as FIASCO (Fiasco Implements Accurate Statistical Computations or Fiasco Is An SPSS Copy) but is now increasingly being referred to as PSPP.
What is Free Software?
The issue of free software is more complex than you might imagine. Free software does not, necessarily, cost nothing. Free software may be available for no money, but will usually cost something. Software that is not free may also be available for no money (for example MS Internet Explorer). When the FSF say "free," they mean "Free software is software that comes with permission for anyone to use, copy, and distribute, either verbatim or with modifications, either gratis or for a fee. In particular, this means that the source code must be available." (http://www.fsf.org/philosophy/categories.html). The FSF do not copyright their free software, rather, they "copyleft" it. This means, in short, that anyone is free to edit and alter the software, but they must also make it free to distribute.
What Can PSPP Do?
PSPP is a project in development - the current version of it is in alpha
test mode (the current version, at the time of writing, is 0.1.22, although
0.2 is due out in the very near future) .It is very limited in its functionality.
Although it has most of the data handling characteristics of SPSS – it
can read and write data files, handle the dictionary, and carry out most
transformations, on the statistics side, it is still very much "in development,"
the only procedures currently developed are t-tests and crosstabs.
Is PSPP a Good Thing?
Many people may question the utility of PSPP – why do we want a copy of SPSS when we have SPSS? Won’t this lead to the end of SPSS as we know it (why should anyone pay for SPSS when there is a free version available?) What about all those people at SPSS – they are going to all the hard work of developing the system, and someone else produces another program that does the same thing?
The assumptions these questions are based on are not correct. In fact, PSPP celebrates the ubiquity of SPSS – the very fact that the Free Software Foundation has deemed it worth creating a program with the functionality of SPSS shows that SPSS has come of age, and is accepted as "a classic."
The two great successes of the FSF are, in my opinion, GNU/Linux and the Gnu C Compiler (GCC). GNU/Linux (commonly, but incorrectly, referred to as Linux) is a free operating system that can be used on a wide range of platforms – most commonly for people who wish to run Unix-like operating systems on PCs. The existence of GNU/Linux has not led to the collapse of other PC operating system producers – Microsoft may be having some local difficulties, but if there is one area where Microsoft is dominant, it is in the world of operating systems. The makers of other flavours of Unix also tend to distribute the machines to run them on, so they feel no threat.
Similarly, GCC is a C Compiler that has almost become a standard in the world of software distribution. The existence of GCC enables software distributors to distribute platform-independent source-code, happy and safe in the knowledge that people can obtain (or, more likely, will already have) the software that can produce the binaries appropriate for the particular brand and type of system they are using.
Both of these systems, useful thought they are, will never replace the major commercial players. What they have done has helped to maintain Unix and C as standards in computer software. People doing serious proprietary development work will never move wholesale to GCC from the likes of Borland Visual C++, but people who are learning to program, or developing non-commercial software may use GCC.
In much the same way that programmers and developers have not abandoned the standard tools of their trades, large marketing companies, universities and other public bodies will not cancel their subscriptions to SPSS to do a wholesale move to PSPP. PSPP is a voluntary effort, and will never keep up with SPSS in terms of either the development of the statistical features or the user interface. The development of PSPP means that SPSS commands, output, and procedures will be able to become standards. If I have carried out a complex analysis, and others would like to confirm it ensuring that they have done everything the same way, I can send them my syntax, and they can run it. It does not matter if they have SPSS, because if they don’t they can use PSPP. If they have PSPP and use it, then, when they come to purchase a statistics package, they will know some SPSS syntax and understand how to interpret SPSS output, and on the basis of the understanding and knowledge they have gained by PSPP, they may be encouraged to decide to to purchase SPSS.
Can I Help?
Given that PSPP is to be a free program, the development is entirely
a voluntary effort. If you are (or are willing to act as) a statistician,
programmer, numerical analyst, technical writer or tester, the developers
would be very happy to hear from you. To join, see the instructions at
the web page http://www.gnu.org/software/fiasco/help-wanted.html.