In this work, the theoretical results of Desai and Sen (2006, 2008) are used to describe several types of analyses of randomly incomplete data without imputation of any sort. This book describes how to characterize Fisher information when data are randomly incomplete. This book includes several examples of analyses of randomly incomplete data without imputation and using some of the most commonly encountered types of models such as repeated-measures MANOVA, tests of independencein an R X C contingency table, the general linear model, and logistic regression. This book uses simulation results wherever possible to compare analyses of data with and without imputation. INDICE: Introduction. Fisher Information in Randomly-Incomplete-Data Likelihoods. Methods for Normal Data. Methods for Categorical Data. The General Linear Model. Logistic Regression. References. Index.
- ISBN: 978-3-642-23504-7
- Editorial: Springer Berlin Heidelberg
- Encuadernacion: Rústica
- Páginas: 95
- Fecha Publicación: 30/09/2011
- Nº Volúmenes: 1
- Idioma: Inglés