Bias and Causation Models and Judgment for Valid Comparisons

by Weisberg, Herbert I.

• ISBN13: 9780470286395
• ISBN10: 0470286393
• Edition: 1st
• Format: Hardcover
• Copyright: 2010-09-07
• Publisher: Wiley

Summary

The cursory treatment of bias as a topic of serious consideration has resulted in a lack of easily accessible reference material. Bias and Causation organizes and clarifies the diverse and somewhat overlapping types of biases within a coherent framework. It provides a comprehensive discussion of the sources of bias in comparative studies (both randomized and observational) and how to address them, emphasizes systematic errors (i.e. bias) that affect proper interpretation of results, and draws concrete examples from biomedical and social science literature to illustrate and explain how biases arise in everyday practice. This will be a single go-to reference for scientific and legal researchers.

Author Biography

Herbert I. Weisberg, PhD, is founder and President of Correlation Research Inc., a consulting firm that specializes in the application of statistics to various business and legal issues. Dr. Weisberg has over forty years of statistical consulting experience and has published numerous articles related to bias assessment and reduction.

Table of Contents

Preface	p. xi
What Is Bias?	p. 1
Apples and Oranges	p. 2
Statistics vs. Causation	p. 3
Bias in the Real World	p. 6
	p. 23
Causality and Comparative Studies	p. 24
Bias and Causation	p. 24
Causality and Counterfactuals	p. 26
Why Counterfactuals?	p. 32
Causal Effects	p. 33
Empirical Effects	p. 38
	p. 46
Estimating Causal Effects	p. 47
External Validity	p. 48
Measures of Empirical Effects	p. 50
Difference of Means	p. 52
Risk Difference and Risk Ratio	p. 55
Potential Outcomes	p. 57
Time-Dependent Outcomes	p. 60
Intermediate Variables	p. 63
Measurement of Exposure	p. 64
Measurement of the Outcome Value	p. 68
Confounding Bias	p. 70
	p. 71
Varieties of Bias	p. 72
Research Designs and Bias	p. 73
Bias in Biomedical Research	p. 81
Bias in Social Science Research	p. 85
Sources of Bias: A Proposed Taxonomy	p. 90
	p. 92
Selection Bias	p. 93
Selection Processes and Bias	p. 93
Traditional Selection Model: Dichotomous Outcome	p. 100
Causal Selection Model: Dichotomous Outcome	p. 102
Randomized Experiments	p. 104
Observational Cohort Studies	p. 108
Traditional Selection Model: Numerical Outcome	p. 111
Causal Selection Model: Numerical Outcome	p. 114
	p. 121
Appendix	p. 122
Confounding: An Enigma?	p. 126
What is the Real Problem?	p. 127
Confounding and Extraneous Causes	p. 128
Confounding and Statistical Control	p. 131
Confounding and Comparability	p. 137
Confounding and the Assignment Mechanism	p. 139
Confounding and Model Specification	p. 141
	p. 144
Confounding: Essence, Correction, and Detection	p. 145
Essence: The Nature of Confounding	p. 146
Correction: Statistical Control for Confounding	p. 172
Detection: Adequacy of Statistical Adjustment	p. 180
	p. 191
Appendix	p. 192
Intermediate Causal Factors	p. 195
Direct and Indirect Effects	p. 195
Principal Stratification	p. 200
Noncompliance	p. 209
Attrition	p. 214
	p. 216
Information Bias	p. 217
Basic Concepts	p. 218
Classical Measurement Model: Dichotomous Outcome	p. 223
Causal Measurement Model: Dichotomous Outcome	p. 230
Classical Measurement Model: Numerical Outcome	p. 239
Causal Measurement Model: Numerical Outcome	p. 242
Covariates Measured with Error	p. 246
	p. 250
Sources of Bias	p. 252
Sampling	p. 254
Assignment	p. 260
Adherence	p. 266
Exposure Ascertainment	p. 269
Outcome Measurement	p. 273
	p. 277
Contending with Bias	p. 279
Conventional Solutions	p. 280
Standard Statistical Paradigm	p. 286
Toward a Broader Perspective	p. 288
Real-World Bias Revisited	p. 293
Statistics and Causation	p. 303
Glossary	p. 309
Bibliography	p. 321
Index	p. 340
Table of Contents provided by Ingram. All Rights Reserved.