Paul D. Allison

Issues in Multiple Imputation of Missing Data

The talk begins with a brief introduction to the principles and practice of multiple imputation. It then proceeds to consider several issues of considerable practical importance: maximum likelihood vs. multiple imputation, imputation of categorical variables via linear models, sequential models vs. comprehensive models, and congeniality of imputation model and analysis model.

To request disability accommodations, contact the Office of the ADA Coordinator ten days in advance of the event. 543-6450 (voice) 685-3885 (FAX) 543-6452 (TDD) access@u.washington.edu