Release 3.5 of the Real Statistics Resource Pack

Happy New Year.

I am pleased to announce Release 3.5 of the Real Statistics Resource Pack. The new release is now available for free download (Download Resource Pack) for Excel 2007, 2010 and 2013 environments.

The Real Statistics Examples Workbook has been updated to reflect the new release. You can also download this file for free (Download Examples). The website is in the process of being updated to reflect the new features. These changes will be made over the course of the next few days.

New features include:

Resampling data analysis tool

A new data analysis tool has been added which provides the following capabilities:

Bootstrapping (p-values, histograms and confidence intervals) for one sample, two sample, paired sample, multiple group (ANOVA) and correlation coefficient tests

Randomization (p-values and histograms) for two sample, paired sample, multiple group (ANOVA) and correlation coefficient tests

Confidence intervals for effect size and statistical power

The website shows how to build confidence intervals for various effect sizes using the noncentral t, chi-square and F distributions. The following functions, which calculate the noncentrality parameter, are used to support this capability:

NT_NCP(p, df, x, m, iter) = ncp such that NT_DIST(x, df, ncp, TRUE, m) = p

NCHISQ_NCP(p, df, x, m, iter) = ncp such that NCHISQ_DIST(x, df, ncp, TRUE, m) = p

NF_NCP (p, df1, df2, x, m, iter) = ncp such that NF_DIST(x, df1, df2, ncp, TRUE, m) = p

In addition, the following inverse noncentral distribution functions have been added:

NT_INV(p, df, p, m, iter) = x such that NT_DIST(x, df, ncp, TRUE, m) = p; i.e. inverse at p of the noncentral t distribution

NCHISQ_INV(p, df, ncp, m, iter) = x such that NCHISQ_DIST(x, df, ncp, TRUE, m) = p; i.e. inverse at p of the noncentral chi-square distribution

NF_INV (p, df1, df2, ncp, m, iter) = x such that NF_DIST(x, df1, df2, ncp, TRUE, m) = p; i.e. inverse at p of the noncentral F distribution

Here m = the number of iterations used to calculate the noncentral distribution (1 – 170, default 40) and iter = the number of iterations used to calculate the inverse or noncentrality parameter (default 40).

Note too that the default values of m for NT_DIST and NF_DIST have also been changed to 40.

Revisions to chi-square distribution and inverse functions

The CHISQ_DIST and CHISQ_INV functions have been revised so that non-integer values of the degrees of freedom are supported.

The CHISQ_INV function is now used for the One Random Factor ANOVA data analysis tool instead of the standard Excel CHIINV (or CHISQ.INV) function since it is useful to support fractional values of df.

Revised Additional Non-parametric Tests data analysis tool

The Other Non-parametric Tests data analysis tool has been renamed the Additional Non-parametric Tests data analysis tool.

This data analysis tool now supports both the frequency table and raw data input formats of the Two Sample Kolmogorov-Smirnov Test.

Bug fixes

Fixes an overflow error in the value of alpha for users who have set their decimal delimitator to a comma instead of a period (i.e. full stop) which occurs when using the Contrast option of the One Factor ANOVA data analysis tool with a Bonferroni or Dunn/Sidák  correction factor.

In addition, default alpha values will be displayed as 0,05 instead of 0.05 on data analysis tool dialog boxes for users who have set their decimal delimitator to a comma (and similarly for default power and cutoff values).

Fixes an error in the two sample Kolmogorov-Smirnov test function KS2TEST with two ranges. The function should now work properly.

Fixes some errors in the tab order for a few of the data analysis tool dialog boxes.

Fixes the Help for the T Test and Non-parametric Equivalents data analysis tool, which was truncated. Also some minor changes were made to the help messages of other data analysis tools.