MAREG is a statistical software tool for estimating marginal regression models. Marginal regression models are an extension of the well known regression models to the case of correlated observations. MAREG currently handles binary, categorical and continuous data with several link functions. Although intended for the analysis of correlated data, uncorrelated data can be analyzed. We supply two different approaches for these problems -- Generalized Estimating Equations and Maximum Likelihood methods. Handling of missing data is also provided.

WinMAREG is a MS Windows user interface for MAREG, allowing method specification, selection and coding of variables, treatment of missing values and selection of general settings.

Keywords: correlated data, generalized linear model, generalized estimating equations, marginal model, maximum likelihood, missing data, software tool, user interface

A more detailled description can be found in Fieger, A., Heumann, C., Kastner, C. (1996): MAREG and WinMAREG (PS file). Features introduced in version 0.2 are described in Kastner, Fieger and Heumann (1999) (PS file).

Statistical background and purpose of the project
WinMAREG quick start
Where can I get it
Installation instructions

MAREG and winMAREG are developed by Andreas Fieger, Christian Heumann, and Christian Kastner as part of their work for the Sonderforschungsbereich 386: Statistische Analyse diskreter Strukturen, Teilprojekt C3 at Institut für Statistik, Ludwig-Maximilians-Universität München
(that was: German Research Foundation - SFB386 C3, Institute of Statistics, University Munich, Germany; links mostly in german language, but some english information also provided).
Suggestions and bug reports to: Christian Heumann or Christian Kastner

Last modified: Thu May 27 16:10:01 CEST 2004