Paul Gustafson Libri

Focusing on statistical challenges from inaccurate measurement of explanatory variables, this book delves into both continuous and categorical variable errors. It emphasizes the need for adjustments in the presence of imprecise covariates, utilizing Bayesian methods and Markov chain Monte Carlo algorithms. Covering a blend of fundamental and specialized topics, the text includes mathematical details in each chapter's final sections, making it a valuable reference for biostatisticians, epidemiologists, and students alike.

The book explores the Bayesian approach to inference within partially identified models (PIMs), highlighting its effectiveness in such contexts. It provides a comprehensive overview of the statistical theory, properties, and applications related to PIMs, including those for misclassified data and instrumental variables. Drawing from extensive research, the author includes recent real data applications, illustrating the practical relevance and performance of Bayesian procedures in analyzing PIMs.