June 11, 2015
Moved the web application from http://www.healthstats.org/rmass/ to http://www.rmass.org.

Updated the user interface so that the application works reliably and consistently on a wide variety of mobile and desktop devices. This included the following minor changes in the labeling and grouping of user interface controls:
  • The Set All to Default button was relabeled Reset All.
  • The Compute and Reset All buttons were moved to the page header.
  • A new Menu button was added to the page header. Clicking Menu displays a panel that provides links to the RMASS Manual, Technical Contact, History of Changes and CHS Projects.
  • The Back to Projects link was relabeled CHS Projects to indicate that it points to the Projects page on the Center for Health Statistics website.
  • All Close this window links were replaced with Close buttons throughout the application.
May 12, 2014
Added a PDF manual which provides a summary of the program and an in-depth discussion of the statistical methodology with examples.
November 12, 2009
The 'Effect size at last timepoint' input field was replaced with a 'Mean difference at last timepoint' (MD) input field. A read-only 'Computed effect size at last timepoint' value was added below the 'Type of mean' radio buttons. This value is automatically updated after modifying either the MD or the 'Treatment by time interaction' (TAU) values.

In addition, when a different value is entered in the MD field, the TAU value is automatically computed by dividing the MD by the last 'Design matrix column 2' value. Similarly, when a different value is entered for TAU, the MD value is automatically updated to the product of the TAU value and the last 'Design matrix column 2' value.
April 9, 2009
RMASS now correctly takes into account a two-sided test type when determining the total number of subjects per center. Previously, the results were always for a one-sided test.
February 22, 2008
The user is now prompted for a valid person-level or center-level covariance value if the value entered exceeds the square root of the product of its intercept and slope variances (values 1,1 and 2,2 in the matrix).
February 19, 2007
Invalid values are now more clearly identified by a different input field color and an "invalid entry" tooltip. Both the field color and the tooltip are reset when clicking on 'Set All to Default'.

The correlation terms are now automatically set to their default values. Each correlation term's default value is also displayed as a tooltip. If the first correlation term has not been modified by the user, its default value changes according to the type of correlation structure of errors selected.