I am pleased to announce the latest release of the Real Statistics Resource Pack. The release provides the following new features:
- A reorganization of the data analysis tools
- New data analysis tools that automate Shapiro-Wilk and create new capbilities for identifying outliers and missing data
- Some new functions
- Improvements to existing capabilities and a few bug fixes
Release 1.8 of the Real Statistics Resource Pack can be downloaded immediately for free from the Free Download page. The release is backwards compatible with previous releases. You can also download the Real Statistics Examples Workbook which is also compatible with this release. I will be updating the website shortly to reflect all the changes in the new release.
I welcome your feedback about this new release and any other features that you would like to see added in future releases. As previously mentioned, a new release, R2.0, will be issued later this month. It will provide multivariate statistics support, including MANOVA and Factor Analysis.
Specific new capabilities in Release 1.8 are as follows:
Data Analysis Tools: The data analysis tools provided in the Real Statistics Resource Pack have been reorganized so that capabilities that are used together are accessible from the same dialog box. In particular, we note the following changes:
Descriptive Statistics: when this option is selected you will have the choice of generating any or all of the following: the existing descriptive statistics, Box plots (improved), the Shapiro-Wilk test (new) or identification of potential outliers (new) and missing data (new).
T Tests: when this option is selected you will have the choice of performing a one-sample analysis, a two sample paired analysis or a two sample independent analysis. You can also choose to perform either the t test or the corresponding non-parametric test (Mann-Whitney or Wilcoxon signed rank tests) or both.
Single Factor Anova: when this option is selected you have the choice of performing any or all of the following analyses: the omnibus ANOVA test itself, contrasts, Tukey HSD test or Scheffe test.
Anova with Repeated Measures: when this option is selected you have the choice of performing the omnibus ANOVA test (including correction factors for sphericity) or contrasts or both.
Logistic Regression: A change has been made to the Logistic Regression data analysis tool which allows much larger data sets to be analyzed. Previously, data sets larger than about 5,000 records could not be analyzed. I want to thank Richmond for identifying this problem.
MAD(R1): A change has been made to the existing supplemental Median Absolute Deviation function so that it will work even when R1 contains empty cells or cells with non-numeric values.
SortUnique(R1): This new function sorts a range R1, omitting any duplicates. This is similar to the existing NoDupes supplemental function, but uses a different algorithm.
COUNTAU(R1): This new function counts the number of unique non-empty cells in range R1. This complements the existing COUNTU(R1) supplemental function which counts the number of unique numeric cells in range R1 and the standard Excel function COUNTA(R1) which counts the number of non-empty cells in range R1.
SWCoeff(n, j, b): The SWCoeff(n, j) function which calculates the Shapiro Wilk coefficient a(j) for samples of size n using the Royston algorithm has now been extended to also support the original SW algoithm. If b is True or is omitted then the Royston algorithm is used, as before. If b is False then SWCoeff(n, j, False) outputs the a(j) coefficients found in the SW Coefficient Table.