On May 1st, Dr. Donna Spiegelman delivered the Fuller Lecture at Iowa State University. Her talk was titled, “Measurement error: from Fuller to the future.” Her abstract is below:
One of the first books I read when I began to study methods for correction for bias in regression coefficients due to covariate measurement error to prepare for my doctoral dissertation research was Professor Wayne Fuller’s 1987 seminal treatment of the topic, “Measurement error models”. Under conditions of normality of the regression model as well as of measurement error in the covariates, he showed that the convergent values of the estimated coefficients of the model in the observed covariates have a simple closed form. In this talk, I will show how this relationship generalizes, asymptotically and sometimes under some additional approximations, to a wide range of models used in biostatistics, epidemiology and public health, including logistic regression models, relative risk regression models, and Cox regression models. Some further extensions to work in progress will also be discussed.