(Ebook) Replication and Evidence Factors in Observational Studies by Paul R. Rosenbaum ISBN 9780367751708, 9781000370027, 9781003039648, 9780367483883, 0367751704, 100037002X, 1003039642, 0367483882

(Ebook) Replication and Evidence Factors in Observational Studies by Paul R. Rosenbaum ISBN 9780367751708, 9781000370027, 9781003039648, 9780367483883, 0367751704, 100037002X, 1003039642, 0367483882

Association does not imply causation, yet some causal conclusions are firmly established based on associations found in observational studies. How does that happen? A study has two evidence factors if it provides two statistically independent tests of one causal hypothesis, susceptible to different biases. Two evidence factors can jointly provide quantifiably stronger evidence than either factor can provide on its own. The first book about evidence factors. Examples are drawn from epidemiology, economics, medical research and other fields. Data from these examples is available in a companion R package that reproduces many of the analyses. Self-contained, presenting needed background from causal inference, statistics and mathematics. Part 1 of the book presents concepts, methods and applications using limited mathematics. The theory of evidence factors is presented in a separate, second part of the book. Mathematics required for the theory is presented from the beginning.