Real Statistics Release 7.7

I am pleased to announce Release 7.7 of the Real Statistics Resource Pack. The new release is now available for free download at Download Resource Pack for Excel 2010, 2013, 2016, 2019 and 365 in both Windows and Mac environments.

The Multivariate, Basics, Distributions, ANOVA 1, ANOVA 2, Regression 1, and Regression 2 examples workbooks will be updated later today for compatibility with the new release. Over the course of the next few days, the website will also be updated for compatibility with the new release.

If you are getting value from the Real Statistics website or software, I would appreciate your donation to help offset the costs of the website by going to Please Donate.

The following is an overview of the new features in Release 7.7:

Ordinal Regression

A new Ordinal Regression data analysis tool has been added. To support this capability, the following new functions have also been added: OLogitCoeff, OLogitCov, OLogitConverge, OLogitLL, and OLogitCorrect.

Data may be entered in raw or summary format. To convert data in raw format to summary format, the function OLogitSummary can be used.

The existing OLogitPredC function has been revised. The previous version is no longer valid. In general, I try to avoid changing the format of Real Statistics functions but have done so in this case since the previous version was probably seldom if ever used.

Binary Logistic Regression and Probit Regression Enhancements

The existing Logistic and Probit Regression data analysis tool has been enhanced primarily for faster processing, especially for data in raw format. Data sets with more than 65,536 rows can now be supported. The existing LogitSummary and MLogitSummary functions have been revised for much faster processing (although the format of these functions remains the same).

In addition to the current output format, a new abbreviated format is now supported as the default.

A bug in the calculation of the Probit coefficients has also been fixed.

The following new functions have also been added:

LogitCov, LogitConverge, LogitLL, LogitCorrect, LogitROC, LogitAUC

ProbitCov, ProbitConverge, ProbitLL, ProbitCorrect, ProbitROC, ProbitAUC

ANCOVA

A new ANCOVA Light data analysis tool has been added that supports One-way ANCOVA even when the parallel slopes assumption is not met. This version supports one or two covariants.

A new ANCOVA1Light(R1, head, center) function has also been added that converts the data in R1 in standard (stacked) format into a format suitable for regression analysis. R1 contains of 3 or 4 columns where column 1 = categorical values, column 2 = dependent variable values, column 3 = values of the covariant, column 4 (optional) = values of a second covariant.

If the first column of R1 contains k categorical values, then the first k-1 columns of the output consist of the dummy codings for the categorical variable, the next column contains the values from the second column of R1, the next one or two columns contain the values for the covariants, and the remaining columns contain the values for the interaction between the dummy variables and the covariant(s).

If head = TRUE then the first row of R1 is assumed to contain column headings and is retained in the headings of the output. If head = FALSE (default) then R1 doesn’t contain a row of column headings; in this case, a row of headings is added to the output.

If center = TRUE (default), then the data corresponding to the covariant(s) are centered in the output (by subtracting the mean of the data for that covariant).

Friedman-Rafsky-Smith-Jain Test

The FRSJ test is now supported via the MNORMTest function. This test uses the Friedman-Rafsky test (FR_TEST) to determine whether a set of multivariate values is multivariate normally distributed.

Enhanced Lp Regression Support

The LAD Regression data analysis tool has been enhanced by adding the Lp norm parameter p to the dialog box, in which case L_p regression is performed for the specified value of p. When p = 1 (default), this is equivalent to LAD regression. 

New Reformatting and Sorting Capabilities

The following sorting array function has been added:

SORTRows(R1, k, head): returns an array of the same size and shape as R1 which sorts the rows in R1 based on the sort keys consisting of the first k columns of R1 (default 1). If head = TRUE (default FALSE) then the first row of R1 is assumed to contain column headings and is retained as the heading of the output

The following compaction and binary compaction array functions have also been added.

BCOMPACT(R1, head): returns an m × (k+1) array which is a summary table for the data in the h × k array R1 whose kth column can only contain the values 0 or 1. The first k-1 columns of the output consist of the unique values in the first k-1 values of R1. The k^th column consists of the counts of the matching rows from R1 that contain 1 in the k^th column and the k+1^th column consists of the counts of the matching rows from R1 that contain 0 in the k^th column.

If head = TRUE (default FALSE) then the first row of R1 is assumed to contain column headings, in which case the elements in the first k-1 columns of the first row of R1 are retained as headings in the output and the k^th and k+1^th columns of the first row contain the values “Success” and “Failure”.

Note that LogitSummary(R1, head) is identical to BCOMPACT(R1, head).

COMPACT(R1, head): returns an m × n array which is a summary table for the data in the h × k array R1. The first k-1 columns of the output consist of the unique values in the first k-1 values of R1. The remaining columns consist of the counts of rows in R1 that match the non-negative integer in the k^th column of R1.

If the minimum value in the kth column of R1 is mn and the maximum value is mx, then the output contains an additional mx-mn+1 columns.

If head = TRUE (default FALSE) then the first row of R1 is assumed to contain column headings, in which case the elements in the first k-1 columns of the first row of R1 are retained as headings in the output and the headings for the remaining columns of the output are mn, mn+1, …, mx. If mx-mn > 24, then an error is generated.

Note that MLogitSummary(R1, head) and OLogitSummary(R1, head) are equivalent to COMPACT(R1, head).

Also, the CountRowsUnique function has been revised for quicker processing.

One-sample t-test

Excel’s T.TEST function supports paired-sample and two-sample t-tests, but not the one-sample t-test. This release adds the T1_TEST(R1, hyp, tails) function that outputs the p-value of the one-sample t-test for the data in array R1 based on the hypothetical mean hyp (default 0) where tails = 1 or 2 (default).

Gage R&R

Revised the Gage R&R option of the Two Factor ANOVA data analysis tool for consistency with the results from Minitab and other software packages. Also, fixed a bug that crept in a few releases ago.

Changes to Main Menu

In addition to adding a new Ordinal Regression option on the Reg tab and a new ANCOVA Light option on the Anova tab, the Survival Analysis option has been moved from the Reg tab to the Misc tab.

Bug Fixes

• Fixed a bug in the DELROWS function
• Fixed a bug in the PBTEST function