Projected scores for GLM's


You want to measure the effectiveness of a treatment. This can only be directly achieved by treating individuals and recording their responses. Unfortunately people tend to be extremely variable in their personal characteristics and behaviors and this variability can get in the way of precisley measuring the effectiveness of the treatment. It would be useful if somehow this variability could be accounted for so that accurate estimates could be obtained. Statisticians have found many ways to formulate this problem and below are links to routines that implement one approach to this heterogeneity issue called the projected score method.


Links to the archived files:

The routines are for the SPlus statistical software package.

Help pages

Background and technical material

Return to my homepage.