MASS: rlm, MM and errors in observations AND regressors
Hello, I need to perform a robust regression on data which contains errors in BOTH observations and regressors. Right now I am using rlm from the MASS package with 'method="MM"' and get visually very nice results. MASS is quite clear, however, that the described methodologies are only applicable to observation-error only data (p. 157, 4th Ed.). So here's the questions now: a) is there methodology for robust and resistant regression on data with errors in BOTH observations and regressors? b) if "yes" at a): does somebody know of an implementation in R? c) if "no" at b): would somebody reading this be open to implement this? If yes: please get in contact with me off list. Please excuse my unfamiliarity with terminology and subject - I'm very new to statistics. Joh
+----------------------------------------------------------------------+ | Johannes Graumann, Dipl. Biol. | | | | Graduate Student Tel.: ++1 (626) 395 6602 | | Deshaies Lab Fax.: ++1 (626) 395 5739 | | Department of Biology | | CALTECH, M/C 156-29 | | 1200 E. California Blvd. | | Pasadena, CA 91125 | | USA | +----------------------------------------------------------------------+