PMean: History of SPSS

I’m helping to put together three separate classes, Basic data management and analysis with R [SAS / SPSS]. As part of these classes, I need to discuss the history of these programs, because understanding that history will help you better understand the strengths and weaknesses of each statistical package. Here’s a brief history of SPSS.

SPSS, an acronym for Statistical Package for the Social Sciences, first appeared in 1968. Interestingly, the marketing department today will only refer to SPSS by its acronym because it wants to de-emphasize its original target audience of social scientists and try to encourage (and appropiately so) use of the program by a broader group of researchers. For a couple of years in 2009, the product was renamed PASW (short for Predictive Analytics SoftWare), but that name change was reversed because of the large customer awareness of the SPSS brand name.

The development team, Norman Nie, Hadlai Hull, and Dale Bent, were all working at Stanford University at the time. Dr. Nie moved to the National Opinion Research Center at the University of Chicago and continues distribution of the program from that University. The actual program was first distributed for free, but the authors of the first SPSS manual (Nie, Bent, and Hull) made a fair amount of money from book royalties. In 1975, to avoid jeopardizing the non profit status of the University, Drs. Nie and Hull created SPSS, Incorporated.

SPSS was originally written in FORTRAN, and some of the legacy of that FORTRAN code still remains, such as the distinction between short strings (8 characters or less) and long strings (more than 8 characters). But because SPSS was written in FORTRAN, it was available, even during its early years on a broad range of minicomputers and mainframe computers. By contrast, SAS, written in a mix of PL/1, FORTRAN, and IBM Assembler Language, was limited at first to IBM mainframes.

SPSS was ported to Microsoft computers in 1984 and to the Windows operating system in 1992. While earlier versions of SPSS required knowledge of the SPSS programming language, the Windows version introduced an easy to use menu based interface that is close to comprehensive. This is in sharp contrast to the menu based interfaces for R and SAS which are not as tightly integrated into the program and which only cover a subset of the total number of features of these packages.

Much of the code for the graphical user interface was re-written in Java in 2007 for version 16 to simplify the process of porting to multiple computers and to allow greater flexibility in the use of their dialog boxes. SPSS also recently added Python capabilities to allow for more efficient processing of macro codes.

Over the years, SPSS has added various modules to the package to handle advanced methods like multiple imputation, complex survey data analysis, and text mining. SPSS also marketed AMOS, a popular program for Structural Equations Modeling.

In 2009, SPSS Incorporated was purchased by IBM. In promotional literature, the package is now called called “IBM SPSS Statistics”.

Here are some resources that I found while researching the history of SPSS.


Update: August 19, 2014. I got a nice email from David Nichols, who has been working at SPSS since 1988. He mentions that Norman Nie provided mostly oversight and did not do much, if any, coding, so I changed the phrasing above from “programming team” to “development team.” He also sent me a link to a Chicago Tribune article about a legal dispute over the trademark SPSS between Norman Nie and IBM and noted that the temporary name change from SPSS to PASW occurred before the buyout by IBM. The Chicago Tribune article had links to other articles in that paper. One written in 1990 during the transition of SPSS to PCs, and another written in 2003 that talks about recent growth in the company and reviews the initial development of SPSS.