12-01-2015 05:55 PM
Hello SAS Community
I was curious as to whether any of the members knew of an approximate degrees of freedom test for ANCOVA in SAS. Lix and Keselman (1995) and Keslman, Wilcox, and Lix (2003) discuss the ADF solution for ANOVA designs and even the use of bootstrapping with the solution. However, I have been unable to find any applications of an ADF solution used with ANCOVA analyses.
Thank you for your time. Any assistance would be appreciated.
12-02-2015 09:35 AM
SAS is committed to the Kenward-Roger (KR) degree of freedom adjustment for the linear model problem. You can read about it in the documentation for MIXED or GLIMMIX. It is not available in GLM. KR is recommended for analysis of covariance. The book SAS for Mixed Models, 2nd edition (2006) demonstrates in detail.
12-02-2015 08:51 PM
Thank you for your response. I will look into the resouces you suggested.
Can the Kenward-Roger (KR) degree of freedom adjustment handle violations to normality? With the ADF solutions I mentioned for ANOVA, trimmed means/bootstrapping can be added to the syntax to handle issues of non-normality.
12-30-2015 12:53 PM
IN PROC GLIMMIX (SAS/STAT13.2 and later, I think), there is an option KR2 for degrees of freedom that is substantially more robust to the assumption of normality, based on a correction for nonlinear covariance structures. It does still assume that the distribution being fit is from the exponential family, so if you roll your own link or variance function, be careful about applying this method for calculating degrees of freedom.