statistiXL is an extremely powerful, feature rich data analysis package that runs as an add-in to Microsoft’s popular Excel spreadsheet program. It provides access to a wide variety of both parametric and non-parametric data analysis tools and statistical tests. By working from within Microsoft Excel, statistiXL is able to leverage the existing features of Excel and provide a host of additional benefits including…
- A familiar and powerful user interface for entering and manipulating data
- A wide variety of formatting options for altering the appearance of results
- The presence of a sophisticated charting package that allows both the manipulation of charts produced by statistiXL and the manual creation of new charts based on statistiXL’s output.
- The ability to perform further ad hoc analysis using Excel’s own built in functions and calculating abilities.
In modern versions of Excel, the features provided by statistiXL can be accessed from a dedicated ribbon, depicted below. Date analysis modules are grouped into categories and can be accessed by the menu associated with each category button. Buttons are multifunctional; clicking on the bottom part will display the menu of modules associated with that category, while clicking on the top part will automatically re-open the most recently used module within that category.
statistiXL also includes a Quick Access Toolbar button that provides access to all of the items found in the statistiXL Ribbon. You can pin this button to you title bar to gain quick access to statistiXL regardless of which ribbon you currently have selected. The list of most recently used tests is also available as a separate item that can also be pinned to the Quick Access Toolbar.
The data analysis features provided by statistiXL fall into the following categories (click on each heading or the menu to the right to see a more detailed description of the feature)
Analysis of Variance (ANOVA) – statistiXL provides both univariate and multivariate ANOVA and ANCOVA. Full factorial and user specified models are supported as are fixed and random factors, nesting and repeated measures.
Contingency Tables – Both 2-way and multi-way contingency data can be analysed.
Correlation Analysis – Simple, Partial, Multiple and Canonical correlation is supported with graphs of Canonical Variates available for the latter.
Descriptive Statistics – Descriptive Statistics are available for both linear and circular data sets. Linear descriptives include a choice of 18 statistics such as Mean, Standard Error and Mode, as well providing Box and Whisker plots and Error Bar plots for a graphical representation of the data. Circular descriptives provide 9 statistics including Mean Angle Circular Variance and Angular Variance.
Discriminant Analysis – Both Grouping and Classification methods of Discriminant Analysis are supported. For Grouping Discriminant Analysis, scatterplots of case scores can be produced for each pair of components. For Classification Discriminant Analysis, an alternate dataset can be classified based on the discriminant functions determined for the primary set.
Factor Analysis – Factor Analysis can be performed on either the correlation or covariance matrix of the raw data set. A variety of component extraction and rotation methods are supported and both scree and scatterplots of case scores can be produced.
Goodness of Fit – A wide variety of tests for the Goodness of Fit of datasets to theoretical distributions are provided including those for Binomial, Circular, Normal, Poisson and Uniform distributions. The level of fit to user specified distributions can also be calculated.
Hierarchical Clustering – Hierarchical clustering of binomial, quantitative and mixed datasets is supported as is clustering based on a predetermined distance matrix. A wide variety of similarity/distance estimates and clustering methods are available and the resultant clustering strategy can be graphically displayed as both a text based and/or graphical dendrogram.
Linear Regression – Simple and Multiple Linear Regression is supported as are comparissons between simple regressions. Plots of regression models and residuals can be produced.
Nonparametric Tests – Numerous Nonparametric Tests are supported including Friedman, Kruskal-Wallis, Mann-Whitney, Mood’s Median, Sign, Spearman, Wald-Wolfowitz and Wilcoxon Paired-Sample tests.
Principal Components – Principal Component Analysis is provided as a means for the reduction of large multivariate data sets into simpler structures. Scree plots and Scatterplots of case scores can be produced.
t Tests – One and two sample, univariate and multivariate t tests are supported.