Statistics and Social Sciences

SPSS Version 8.0
What Are You Waiting For?

Frank LoPresti and Bob Yaffee

[Ed: Links to web pages and/or e-mail addresses which have become inactive since the publication of this article have been enclosed in curly brackets { }. Replacement links have been provided where possible.]

For users of SPSS, NYU's most widely used statistical package for social scientists, the upgrade from version 6.x to 7 was a big change. Version 7 was introduced with a radical output window that upset the sleepy world of statistical programmers.

Now, for those who have paid their dues, the move to version 8.0 will be much more comfortable. Everyone, no matter which version they currently use, should upgrade to this version because it really is easier to use, and newer versions of Windows won't run version 6.x.

Of course, there are also some new features not to be missed, such as the many new and useful Help tools.The main statistics enhancement of version 8.0 is the ability to calculate Intraclass Correlation (ICC) coefficients. But before jumping into your first SPSS 8.0 session, make sure to inspect this version's new graphics.

Graphics and Enhancements

Pull down "Graphs" from the menu bar to find two new selections—"Gallery" and "Interactive." Choose "Gallery" to go to a chart gallery (shown below) where the different types of graphs are displayed. This feature will be very useful to students.

The "Interactive" option walks the user through the creation and editing of all types of graphs. During the first session, SPSS uses a "drag" hand to move variables into the graph dialog window. This dialog box helps assign variables, legends, summary functions and appearance attributes. Many statistical programmers felt that SPSS's earlier graphing fell very short of SAS's long-standing graphing capabilities. Perhaps SPSS has now closed the gap. Much more time must be invested in using the new graphs features before we can judge them fairly.

Enhancements in Reliability Analysis

Some research methods involve using a panel of expert judges to classify scenarios or data to be used later in an experiment. The reliability or consistency of the judges is rated using Intraclass Correlation (ICC) coefficients.

Reliability Analysis within SPSS 8.0 has been enhanced by the inclusion of six types of ICC coefficients, taken from Professors Pat Shrout and Joseph Fleiss's 1979 article, "Intraclass Correlations: Uses in Assessing Rater Reliability," in Psychological Bulletin, Vol. 86, 420-428. In that article, Shrout and Fleiss provide two versions of each of three models of an ICC coefficient. The three models are a one-way random effects model, with a random selection of judges; a two-way random effects model, where judges are randomly selected from a larger population; and a two-way mixed effects model, where the judges doing the rating are fixed. A single measure or an average measure are applied to each model.

For the details concerning the formulation, statistical application, sample size requirements, and SPSS programming invocation of these different coefficients, see "Enhancements in Reliability Analysis" by Robert A. Yaffee on the Statistics and Social Science Group's publications section, at www.nyu.edu/its/socsci/Docs/intracls.html.

Other New Features

Database Capture Wizard is an interactive engine which sorts retrieval queries from Excel, dBase, Access and ASCII files to create SPSS datasets. It uses ODBC, an Open DataBase connectivity standard that allows data to travel across software. Steps include selection of the data source, developing criteria to limit the cases retrieved, and ordering the variables in the SPSS data set. SPSS syntax for the query is created and may be saved to be edited and rerun.

There is also a new feature for variable selection when performing tasks interactively. An option is now available to show variable labels rather than variable names when running statistical tasks from the menus. Labels can be up to 256 characters long, providing much more information than the usual eight-character names.

Results Coach helps you understand your results by creating annotated sample output for procedures you have run. It is available from the Output window by double-clicking the table from a procedure. Double-clicking also still allows us to edit the table. These figures came from using the Results Coach after running a bivariate correlation.

At ACF's PC labs, SPSS will be upgraded to Version 8.0 in time for the summer session. (The documentation for Version 8.0 for Windows was previously online at http://www.nyu.edu/acf/pubs/SPSS_Win). The new document will be enhanced to include a section detailing the use of a survey data set that has been stored on a local file server.

Faculty wishing to deposit data sets on ACF servers for use within their course curriculum can contact Frank LoPresti at frank.lopresti@nyu.edu. [ C ]

Frank LoPresti heads the Statistics and Social Sciences Group, where Dr. Yaffee was a statistician at the time of this article's publication.
frank.lopresti@nyu.edu
{bob.yaffee@nyu.edu}