Statistical Analysis with Measurement Error or Misclassification

Strategy, Method and Application

verfasst von: Grace Y. Yi

Verlag: Springer New York

Buchreihe : Springer Series in Statistics

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This monograph on measurement error and misclassification covers a broad range of problems and emphasizes unique features in modeling and analyzing problems arising from medical research and epidemiological studies. Many measurement error and misclassification problems have been addressed in various fields over the years as well as with a wide spectrum of data, including event history data (such as survival data and recurrent event data), correlated data (such as longitudinal data and clustered data), multi-state event data, and data arising from case-control studies. Statistical Analysis with Measurement Error or Misclassification: Strategy, Method and Application brings together assorted methods in a single text and provides an update of recent developments for a variety of settings. Measurement error effects and strategies of handling mismeasurement for different models are closely examined in combination with applications to specific problems.
Readers with diverse backgrounds and objectives can utilize this text. Familiarity with inference methods—such as likelihood and estimating function theory—or modeling schemes in varying settings—such as survival analysis and longitudinal data analysis—can result in a full appreciation of the material, but it is not essential since each chapter provides basic inference frameworks and background information on an individual topic to ease the access of the material. The text is presented in a coherent and self-contained manner and highlights the essence of commonly used modeling and inference methods.
This text can serve as a reference book for researchers interested in statistical methodology for handling data with measurement error or misclassification; as a textbook for graduate students, especially for those majoring in statistics and biostatistics; or as a book for applied statisticians whose interest focuses on analysis of error-contaminated data.
Grace Y. Yi is Professor of Statistics and University Research Chair at the University of Waterloo. She is the 2010 winner of the CRM-SSC Prize, an honor awarded in recognition of a statistical scientist's professional accomplishments in research during the first 15 years after having received a doctorate. She is a Fellow of the American Statistical Association and an Elected Member of the International Statistical Institute.

Inhaltsverzeichnis

Frontmatter

1. Inference Framework and Method

Abstract

This chapter sets the stage for the development of the book. The discussion in this chapter concerns the standard context in which mismeasurement is absent. This chapter lays out a broad framework for parametric inferences where estimation is of central interest. §1.1 outlines the inference framework and the objectives. Important issues concerning modeling and inferences are discussed in §1.2. Representative and useful estimation methodology is reviewed in §1.3. Strategies of handling model misspecification are described in §1.4, and the extension to the regression setting is included in §1.5. Brief bibliographic notes are presented in §1.6.

Grace Y. Yi

2. Measurement Error and Misclassification: Introduction

Abstract

In Chapter 1, we provide an overview of statistical modeling and inference methods. There is a critical condition underlying the development: variables included in the models must be measured precisely. This condition is, however, frequently violated.

Grace Y. Yi

3. Survival Data with Measurement Error

Abstract

Survival analysis is commonly challenged by the presence of covariate measurement error. Biomarkers, such as blood pressure, cholesterol level, and CD4 counts, are subject to measurement error due to biological variability and other sources of variation. It is known that standard inferential procedures often produce seriously biased estimation if measurement error is not properly taken into account. Since the seminal paper by Prentice (1982), there has been a large number of research papers devoted to handling covariate measurement error for survival data.

Grace Y. Yi

4. Recurrent Event Data with Measurement Error

Abstract

Recurrent event data arise commonly in public health and medical studies. While analysis of such data has similarities to that of survival data for many settings, recurrent event data have their own special features. Compared to the extensive attention given to survival data with covariate measurement error, there are relatively limited discussions on analysis of error-prone recurrent event data. In this chapter, we discuss several models and methods to shed light on this topic.

Grace Y. Yi

5. Longitudinal Data with Covariate Measurement Error

Abstract

Longitudinal studies are routinely conducted in various fields, including epidemiology, health research, and clinical trials. A variety of modeling and inference approaches are available for longitudinal data analysis. The validity of these methods relies on an important requirement that variables are precisely measured. This assumption is, however, often violated in practice.

Grace Y. Yi

6. Multi-State Models with Error-Prone Data

Abstract

Multi-state stochastic models are closely related to survival and longitudinal data analysis. They may be used to describe survival data from a perspective different from what is discussed in Chapter 3 They also provide a useful framework for analyzing longitudinal data when interest lies in dynamic aspects of the underlying process.

Grace Y. Yi

7. Case–Control Studies with Measurement Error or Misclassification

Abstract

In epidemiological research case–control studies provide an important method to investigate factors contributing to certain medical conditions, such as disease statuses. Case–control studies are quick and cheap to conduct. They enable us to study rare health outcomes without having to follow up a large number of subjects over a long period of time. Analysis of case–control studies dates back to Broders (1920) and Lane-Claypon (1926). Various statistical analysis methods for case–control data have been developed since the landmark paper by Cornfield (1951). Those methods are, however, vulnerable to measurement error and misclassification that commonly accompany case–control studies. This chapter deals with this topic and discusses inference methods for handling error-prone data arising from case–control studies.

Grace Y. Yi

8. Analysis with Mismeasured Responses

Abstract

In many settings, precise measurements of variables are difficult or expensive to obtain. Both response and covariate variables are equally likely to be mismeasured. Measurement error in covariates has received extensive research interest. A large body of analysis methods, as discussed in the aforementioned chapters, has been developed in the literature. Issues on mismeasured responses, on the other hand, have been relatively less explored.

Grace Y. Yi

9. Miscellaneous Topics

Abstract

Many methods discussed in this book are motivated by research problems arising from various fields, including nutrition studies, cancer research and environmental studies. Methods and application of measurement error models are vast in the epidemiology literature. Although the book discusses some research in this field, the coverage is far from complete.

Grace Y. Yi

Backmatter

Titel: Statistical Analysis with Measurement Error or Misclassification
verfasst von: Grace Y. Yi
Verlag: Springer New York
Electronic ISBN: 978-1-4939-6640-0
Print ISBN: 978-1-4939-6638-7
DOI: https://doi.org/10.1007/978-1-4939-6640-0