The latest

With costs ranging from $100 to more than $1,000, statistics software packages such as SPSS can be too pricey for many graduate students to buy for at-home data crunching. However, an increasingly popular group of computer programs, known as "open source," could be a boon to cost-conscious students who want to analyze their data outside of their university computer lab or who need precise control over their calculations.

Students can download the open-source software-such as "R," which performs many of the functions of the popular statistics program SPSS-off of the Internet for free. Another free program, "Mx," can be used in place of commercially available structural equation modeling software.

In addition to being free, these programs can create charts and graphs to students' exact specifications, says Ken Kelley, a fifth-year quantitative psychology graduate student at the University of Notre Dame. However, R lacks SPSS's automatic data cleaning function-which, for example, flags results outside of a given range.

A nonprofit foundation funds R's development, and psychologist Michael C. Neale, PhD, a professor of psychiatry and human genetics at Virginia Commonwealth University, wrote Mx with the support of a grant that stipulated that the software not be sold. As a result, anyone can download and view both R's and Mx's code, and add to it-either on their own computers or on the programs' online databases.

Because an international community of researchers continually reviews and updates both Mx and R, the programs incorporate cutting-edge statistical methods before similar commercial software does, says Kelley, who uses both programs in his own research.

But while commercial programs often include easily understandable "point and click" controls, working with R requires some knowledge of computer programming and logic, Kelly notes.

"It's definitely harder to learn R than SPSS, but once you know it you can do so much more," says Kelly.


 The R Web site is www.r-project.org. The Mx Web site is www.vcu.edu/mx.

—S. DINGFELDER