University of Southern California

ITS Information Technology Services

A division of the Office of the Chief Information Officer

Statistics Using SPSS


SPSS offers a full range of pre-packaged statistics as part of the standard installation at USC. You are encouraged to explore the possibilities either using your own data or one of the many datasets included with SPSS.

These statistics are usually installed on Windows under

C:\Program Files\SPSSInc\Statistics17\Samples

or on a Macintosh in

Applications\SPSSInc\Statistics17\Samples

All statistics available in SPSS at USC are selected through the Analyze menu. There are 18 options, from Reports to ROC curves, under the Analyze menu. The arrows at the side of each of these options indicate that there are additional choices within that category. When you select any one of the options, a new drop-down menu will allow you to choose which subset of these options you wish to use.

SPSS Analyze

Types of Statistics Available with SPSS

SPSS statistics in the Base Package installed at USC fall into three global categories.

  1. Descriptive Statistics - Statistics used to describe your sample can be obtained using either the Reports option or the Descriptive Statistics option on the Analyze menu. To produce simple reports with the mean, sum, minimum, maximum or other basic statistics, you can select the Reports option. Please go to our Creating Reports Using SPSS page for more information. The Descriptive Statistics option allows you to select Frequencies, Descriptives (includes mean,variance, sum, kurtosis) and other basic statistics), Cross-tabulation, Explore and a few simple plots.

  2. Parametric Statistics - Parametric statistics can be selected under options such as Compare Means, Regression, Correlate, and the General Linear Model. These methods are extremely common but do require meeting certain assumptions, particularly that the data are normally distributed.

  3. Non-Parametric Statistics - Non-parametric statistics are used for data that cannot meet the distribution assumptions for parametric methods. A common use for non-parametric statistics are studies with dichotomous dependent variables, such as mortality. You either are dead or not dead, there really is not a bell-shaped curve for mortality. Many of these statistics can be found under the Non-parametric Tests option. Logistic regression and probit models can also be found under the Regression option.

Last updated:
February 19, 2010

SPSS

The use of all USC computing resources is governed by the USC Computing Policies.