IBM SPS Statistics Extension Command News

Extension Commands for IBM SPSS Statistics supplement the built-in commands with additional statistical, graphical, data management, and other capabilities that look and operate like the built-in commands.  At this writing, 1/8/2026, there are about 160 such commands available on the SPSS Extension Hub.  Some are automatically installed with SPSS, but most need to be installed from the Extension Hub in order to use them.   Once installed, they appear on the menus, and almost all have standard-style SPSS syntax.

You can see what’s available via the Extensions > Extension Hub menu or by running the (natch) extension command Extensions > Installed Extensions Report (STATS EXTENSION REPORT) once that is installed.  EXTENSION REPORT can identify extensions that are not installed as well.

Since you have to look for them, you might not know what you might be missing.  The purpose of this post is to point out the extensions posted in the last year or so to the Extension Hub.  These will all appear under the Analyze menu.

·         STATS NORMALITY ANALYSIS (Descriptive Statistics > Normality Analysis): a set of tests for univariate and multivariate normality and  several distribution plots

·         STATS CITREE (Classify > Conditional Inference Trees):  classification and regression trees. In contrast to the three algorithms available in the SPSS TREES procedure, conditional inference trees use statistical significance test levels on means or categorical data with multiple testing corrections and depth-first measures to find splits. These trees are less likely to overfit without cross validation and are, therefore, more stable.  It also provides flexible tree plots.  The command has flexible ways of displaying large trees as outlines.

·         STATS EARTH (Generalized Linear Models > Multiple Adaptive Regression Splines): MARS regression and classification models that can do variable selection  and detect nonlinearities and interactions in linear and logistic regression

·         STATS BORUTAFEATURES (Descriptive Statistics > Boruta Feature Selection):  find important predictive variables for a scale or categorical dependent variable using the Boruta algorithm, which  compares the importance of original features with a set of created "shadow" features using random forests. 

·         STATS MIXED CLUSTER (Classify > Mixed Type Cluster): clustering for a mixture of categorical and scale variables that may outperform the SPSS TWOSTEP CLUSTER procedure

·         STATS PERM (Regression > Permutation Tests): Permutation (resampling) tests for two-group t tests, anova, and regression models.  Normality is not required, and it works even when the dataset is too small for asymptotic properties to apply - currently in QA.

·         STATS BAYES SELECTVARS (Generalized Linear Models > Bayesian Regression Variable Selection): uses Bayesian model averaging to provide probabilities of nonzero coefficients and shows selected variable patterns in the top five models.  It handles linear, logistic, Poisson, and gamma error term – currently in QA.

All of these are free with SPSS Statistics and are available on all SPSS operating systems.

I hope you will find some of these  useful in your work.