[Erppcatoolkit-support] EP Toolkit 2.994 released

[Joseph D. <jd...@ma...>]

EP Toolkit version 2.994 has been released:
https://sourceforge.net/projects/erppcatoolkit/

Aside from the usual bug fixes and improvements to the working of the 
user interface, there has been substantial upgrades to the output from 
the ANOVA function, including more comprehensive and up-to-date effect 
sizes, confidence intervals, information used for power analyses, and 
graphs of outliers and effects.  The robust statistics are now current 
as of the perspective represented in Wilcox (2022).

1) Fixed auto option not working for multi-factor structures where the 
first letter represents a factor.
2) Fixed crash when saving ANOVA table name.
3) Added buttons for loading and saving the Window cells table.
4) Added checkboxes to Save function to allow for multiple saves.
5) Fixed Winsorized Standard Errors in output cell tables computed 
incorrectly. Test statistics not affected.
6) Fixed crash when using 10-05 column.
7) Fixed p-value sometimes displayed with more than two significant 
digits by ANOVA function.
8) Fixed not clearing out old column names in ANOVA interface when 
loading in a newer dataset with fewer columns.
9) Fixed crash when loading an ANOVA file with more than one spec.
10) Fixed Winsorized Standard Errors in output cell tables computed 
incorrectly. Test statistics not affected.
11) Fixed robust MSE computed incorrectly for designs with multiple 
factors.  Test statistics not affected.
12) Changed vector scaling test to compute scaling within collapsed 
cells rather than original cells.  Test statistics potentially affected.
13) Added parallel computing calls to accelerate robust statistics.
14) Added buttons to ANOVA output html for selecting between cell means 
tables with SD, Cousineau-Moray SE, robust Cousineau-Moray CI, outlier 
plots, or plots with CI error bars.
15) Fixed crash when loading ANOVA table.
16) Added mean R to ANOVA output for use with power calculations.
17) Fixed Save function checkboxes crashing and also further 
streamlining functionality.
18) Removed Loftus-Masson standard error statistic from ANOVA output.
19) Added dR effect sizes and confidence intervals for pairwise 
within-group effects as well, plus effect sizes for between and within 
group effects with more than two levels.
20) Added bold for non-significant effects with confidence interval 
boundaries less than absolute .2.
21) Moved files list of History pane to a popupmenu so that the contents 
can be better inspected.
22) Changed output of trimmed means to be those of collapsed cells 
(e.g., with 5% trimming, two separate cells with n=10 each would not be 
trimmed whereas a collapsed cell with n=20 would be trimmed. Means 
tables and test statistics potentially affected.

Joe
--------------------------------------------------------------------------------

Joseph Dien, PhD
Senior Research Scientist
Human Development and Quantitative Methodology Department
University of Maryland, College Park
http://joedien.com