Explains the value of quasi-experimental techniques that can be used to approximate randomized experiments. The goal is to describe the logic of causal inference for researchers and policymakers who are not necessarily trained in experimental and quasi-experimental designs and statistical techniques.

Researchers often estimate causal effects in experimental or observational studies after coarsening continuous measures of treatments. In the statistical matching context, in particular, non-discrete interventions are frequently discretized to facilitate pair-stratification using traditional matching approaches for binary treatments. A well-known issue in studying coarsened interventions is that any coarsening induces measurement error that attenuates estimates, while inflating estimator standard errors. While this bias is known, there is yet no standard correction for it. This research note illustrates the error-in-variables structure underlying the use of discrete transformations of non-discrete (or dose) interventions. It also recommends the use of the standard IV estimator to recover an unbiased estimate of the uncoarsened treatment effect. Particular attention is given to the problem of matching with a continuous intervention, which motivates simulations.

A daily glass of wine prolongs life—yet alcohol can cause life-threatening cancer. Some say raising the minimum wage will decrease inequality while others say it increases unemployment. Scientists once confidently claimed that hormone replacement therapy reduced the risk of heart disease but now they equally confidently claim it raises that risk. What should we make of this endless barrage of conflicting claims? Observation and Experiment is an introduction to causal inference by one of the field’s leading scholars. An award-winning professor at Wharton, Paul Rosenbaum explains key concepts and methods through lively examples that make abstract principles accessible. He draws his examples from clinical medicine, economics, public health, epidemiology, clinical psychology, and psychiatry to explain how randomized control trials are conceived and designed, how they differ from observational studies, and what techniques are available to mitigate their bias. “Carefully and precisely written...reflecting superb statistical understanding, all communicated with the skill of a master teacher.” —Stephen M. Stigler, author of The Seven Pillars of Statistical Wisdom “An excellent introduction...Well-written and thoughtful...from one of causal inference’s noted experts.” —Journal of the American Statistical Association “Rosenbaum is a gifted expositor...an outstanding introduction to the topic for anyone who is interested in understanding the basic ideas and approaches to causal inference.” —Psychometrika “A very valuable contribution...Highly recommended.” —International Statistical Review

Social scientists and policy makers continue to put increased emphasis on identifying causal effects in their research, employing the myriad of novel approaches that have been developed in recent years. With this rise in the use of causal analysis tools, the importance of understanding the assumptions underlying such methods has increased. Methods and estimates are occasionally stretched beyond the limits of the underlying assumptions. This has increased the need for a thorough understanding of what assumptions are necessary to identify causal estimates outside of a purely experimental framework. Additionally, it has lead to a need for validation tests that allow researchers to provide sound evidence that estimates are unbiased. This dissertation focuses on one area, methods for increasing the external validity of experimental estimates, of interest to researchers and policy makers alike. This problem is addressed from multiple angles, both theoretical, practical, and from a design perspective. First, I outline the assumptions necessary to identify population treatment effects from experimental data. I then discuss a new estimating strategy for testing the key identifying assumption central to most causal methods, the notion of exchangeability. This new method leverages tests of equivalence to redefine the approach to validation tests such as balance and placebo tests. Finally, a new approach to improving the underlying data from which we explore questions and estimate quantities of interest is provided. This new method, response rate sampling, allows researchers to collect more representative survey data. Combined, these tools will allow researchers to move more fluidly between experimental and observational estimates.

This User’s Guide is a resource for investigators and stakeholders who develop and review observational comparative effectiveness research protocols. It explains how to (1) identify key considerations and best practices for research design; (2) build a protocol based on these standards and best practices; and (3) judge the adequacy and completeness of a protocol. Eleven chapters cover all aspects of research design, including: developing study objectives, defining and refining study questions, addressing the heterogeneity of treatment effect, characterizing exposure, selecting a comparator, defining and measuring outcomes, and identifying optimal data sources. Checklists of guidance and key considerations for protocols are provided at the end of each chapter. The User’s Guide was created by researchers affiliated with AHRQ’s Effective Health Care Program, particularly those who participated in AHRQ’s DEcIDE (Developing Evidence to Inform Decisions About Effectiveness) program. Chapters were subject to multiple internal and external independent reviews. More more information, please consult the Agency website: www.effectivehealthcare.ahrq.gov)

An overview of experimental research and methods in public management, and their impact on theory, research practices and substantive knowledge.

Matched sampling is often used to help assess the causal effect of some exposure or intervention, typically when randomized experiments are not available or cannot be conducted. This book presents a selection of Donald B. Rubin's research articles on matched sampling, from the early 1970s, when the author was one of the major researchers involved in establishing the field, to recent contributions to this now extremely active area. The articles include fundamental theoretical studies that have become classics, important extensions, and real applications that range from breast cancer treatments to tobacco litigation to studies of criminal tendencies. They are organized into seven parts, each with an introduction by the author that provides historical and personal context and discusses the relevance of the work today. A concluding essay offers advice to investigators designing observational studies. The book provides an accessible introduction to the study of matched sampling and will be an indispensable reference for students and researchers.

The modern proliferation of large observational databases -- in fields such as e-commerce and electronic health -- presents challenges and opportunities for applied researchers. Such data can contain rich information about causal effects of interest, but the effects can only be estimated if we make untestable assumptions and carefully model the assignment mechanism. Experimental data provides a "virtuous" counterpart for the purposes of inferring causal effects, but randomized trials are often limited in size and, consequentially, lack precision. In this thesis, we consider problems of "data fusion, " in which observational and experimental datasets are used together to estimate causal effects. The problem is considered from three angles. First, we develop methods for merging experimental and observational causal effect estimates in the case when all confounding variables are measured in the observational studies. Next, we remove the unconfoundedness assumption, which leads to a new class of estimators based on a shrinkage approach. Finally, we propose a novel solution for designing experiments informed by observational studies, making use of the regret minimization framework. Throughout, we deploy tools from disparate areas of the literature, including Empirical Bayes, decision theory, and convex optimization.