Statistics Toolbox Release Notes

# Chapter 2Statistics Toolbox 3.0 Release Notes

New Features

This section introduces the new features and enhancements added in the Statistics Toolbox 3.0 since the Statistics Toolbox 2.2 (Release 11.0).

Summary of Enhancements

Expanded Support for Linear Models

Version 3.0 expands the Statistics Toolbox support for linear models in general, and analysis of variance in particular. The following are the major changes for the Statistics Toolbox 3.0:

• Improvements in one-way analysis of variance (`anova1`)
• Higher-way analysis of variance (`anovan`)
• Analysis of covariance (`aoctool`)
• Multiple comparisons of means or other estimates (`multcompare`)
• Multivariate analysis of variance (`manova1`, `manovacluster`)
• Graphics functions useful for examining data used for multivariate analysis of variance (`gscatter`, `gplotmatrix`, `gname`)
• Response surface fitting with multiple responses (`rstool`)
• Nonparametric analysis of variance (`friedman`, `kruskalwallis`)
• More flexible calculation of confidence bounds (`polytool`, `nlintool`, `nlpredci`)

Other Enhancements

In addition, the following changes do not involve linear models:

• Generalized linear models (`glmfit`, `glmval`)
• Robust regression (`robustfit`, `polytool`)
• Distribution testing and plotting (`cdfplot`, `lillietest`, `kstest`, `kstest2`)
• Fractional factorial design generation (`fracfact`)
• Importing numeric and text data from tab-delimited files (`tdfread`)
• More flexible handling of grouping variables (`boxplot`, `grpstats`)
• Multivariate t random number generation (`mvtrnd`) and improvements to other t distribution functions

Numerous other functions received enhancements, as described in the following sections:

 Known Software and Documentation Problems New Functions