Hello, Does R have any capabilities (or are there any add on packages) which can do estimation of standard statistical models (means, regression, logistic regression, etc) which take into account not only weights (e.g. post-stratification weights) but also the sample design, such as stratification and clustering information (to compute a robust taylor linearized variance estimator, for example)? Thanks much for any input, Jason _______________________________ Jason C. Bond, Ph.D. Biostatistician, Associate Scientist Public Health Institute Alcohol Research Group 2000 Hearst Avenue Berkeley, CA 94709 Telephone: (510) 642-7965 Fax: (510) 642-7175
On Wed, 05 Feb 2003 13:50:06 -0800 Jason Bond <jbond at arg.org> wrote:> Hello, > > Does R have any capabilities (or are there any add on packages) which > can do estimation of standard statistical models (means, regression, > logistic regression, etc) which take into account not only weights (e.g. > post-stratification weights) but also the sample design, such as > stratification and clustering information (to compute a robust taylor > linearized variance estimator, for example)? Thanks much for any input, > > Jason > > _______________________________ > > Jason C. Bond, Ph.D. > Biostatistician, Associate Scientist > Public Health Institute > Alcohol Research Group > 2000 Hearst Avenue > Berkeley, CA 94709 > > Telephone: (510) 642-7965 > Fax: (510) 642-7175 > >For ols, logistic, accelerated failure time, and Cox models you can do cluster adjustments using either the robcov (Huber-White-Efron methods) or bootcov (cluster bootstrap) functions in the Design library. See hesweb1.med.virginia.edu/biostat/s/Design.html Doing this as the same time as using weights may not be supported by the functions. -- Frank E Harrell Jr Prof. of Biostatistics & Statistics Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences U. Virginia School of Medicine hesweb1.med.virginia.edu/biostat
On Wed, 5 Feb 2003, Jason Bond wrote:> Hello, > > Does R have any capabilities (or are there any add on packages) which > can do estimation of standard statistical models (means, regression, > logistic regression, etc) which take into account not only weights (e.g. > post-stratification weights) but also the sample design, such as > stratification and clustering information (to compute a robust taylor > linearized variance estimator, for example)? Thanks much for any input,The new `survey' package does this for means, tables, glm and cox models. It would be pretty easy to add support for the parametric survival models if there were any demand. It incorporates weights, stratification, clustering, and the finite population correction to variance if desired. -thomas