We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
To save content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about saving content to .
To save content items to your Kindle, first ensure coreplatform@cambridge.org
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about saving to your Kindle.
Note you can select to save to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
Given the centrality of counterfactual difference-making relations to the argument of the book, in Chapter 7 we say a few more things about contrary-to-fact conditionals. We offer a primer on the logic and the semantics of counterfactuals, focusing on the two main schools of thought: the metalinguistic and the possible-worlds approach. We also present and examine James Woodward’s interventionist counterfactuals and the Rubin-Holland model. We argue that the counterfactual approach is more basic than the mechanistic, but information about mechanisms can help sort out some of the methodological problems faced by the counterfactual account.
This first concluding chapter discusses the ways that social science literature has analyzed the legacies of colonial rule, and argues that colonial state-building may represent a more fruitful approach to analyzing complex regimes and their long-term consequences. It presents a discussion of the logic of historical legacies in relation to the comparative-historical framework of critical junctures and path dependency. It then discusses the drawbacks to dominant approaches that focus on causal inference in assessing colonial legacies.
We introduce hierarchically regularized entropy balancing as an extension to entropy balancing, a reweighting method that adjusts weights for control group units to achieve covariate balance in observational studies with binary treatments. Our proposed extension expands the feature space by including higher-order terms (such as squared and cubic terms and interactions) of covariates and then achieves approximate balance on the expanded features using ridge penalties with a hierarchical structure. Compared with entropy balancing, this extension relaxes model dependency and improves the robustness of causal estimates while avoiding optimization failure or highly concentrated weights. It prevents specification searches by minimizing user discretion in selecting features to balance on and is also computationally more efficient than kernel balancing, a kernel-based covariate balancing method. We demonstrate its performance through simulations and an empirical example. We develop an open-source R package, hbal, to facilitate implementation.
Disasters have short and long-term negative effects on a large array of physical and mental health outcomes. Epidemiology offers a variety of tools and methodologies for conducting a needs assessment, surveillance, and longitudinal research aimed at identifying adverse outcomes and developing strategies for preventing disease and promoting health. The application of epidemiological methods has advanced our understanding of pervasive morbidity and mortality often experienced in the aftermath of disasters. Findings from epidemiological studies have implications for improving the allocation of resources and developing interventions targeting these adverse outcomes. In this chapter, we briefly highlight developments in the epidemiology of disasters. We present common study designs employed in disaster response and research and provide examples of applications of these methods in studying the consequences of the 1988 Spitak earthquake in Armenia. The chapter concludes with a brief discussion of recent developments in research methodology and their potential implications for disaster researchers and public health practitioners focusing on prevention and mitigation.
“Process tracing and program evaluation, or contribution analysis, have much in common, as they both involve causal inference on alternative explanations for the outcome of a single case,” Bennett says. “Evaluators are often interested in whether one particular explanation – the implicit or explicit theory of change behind a program – accounts for the outcome. Yet they still need to consider whether exogenous non-program factors … account for the outcome, whether the program generated the outcome through some process other than the theory of change, and whether the program had additional or unintended consequences, either good or bad.” Bennett discusses how to develop a process-tracing case study to meet these demands and walks the reader through several key elements of this enterprise, including types of confounding explanations and the basics of Bayesian analysis.
Chapter 2 starts by placing experiments in the scientific process – experiments are only useful in the context of well-motivated questions, thoughtful theories, and falsifiable hypotheses. The author then turns to sampling and measurement since careful attention to these topics, despite being often neglected by experimentalists, are imperative. The remainder of Chapter 2 offers a detailed discussion of causal inference that is used to motivate an inclusive definition of “experiments.” The author views this as more than a pedantic exercise, as careful consideration of approaches to causal inference reveals the often implicit assumptions that underlie all experiments. The chapter concludes by touching on the different goals experiments may have and the basics of analysis. The chapter serves as a reminder of the underlying logic of experimentation and the type of mindset one should have when designing experiments. A central point concerns the importance of counterfactual thinking, which pushes experimentalists to think carefully about the precise comparisons needed to test a causal claim.
Experiments are a central methodology in the social sciences. Scholars from every discipline regularly turn to experiments. Practitioners rely on experimental evidence in evaluating social programs, policies, and institutions. This book is about how to “think” about experiments. It argues that designing a good experiment is a slow moving process (given the host of considerations) which is counter to the current fast moving temptations available in the social sciences. The book includes discussion of the place of experiments in the social science process, the assumptions underlying different types of experiments, the validity of experiments, the application of different designs, how to arrive at experimental questions, the role of replications in experimental research, and the steps involved in designing and conducting “good” experiments. The goal is to ensure social science research remains driven by important substantive questions and fully exploits the potential of experiments in a thoughtful manner.
We describe fundamental challenges to estimating heterogeneous treatment effects in the context of the statistical causal inference literature, proposed algorithms for addressing those challenges, and methods to evaluate how well heterogeneous treatment effects have been estimated. We illustrate the proposed algorithms using data from two large randomized trials of blood pressure treatments. We describe directions for future research in medical statistics and machine learning in this domain. The focus will be on how flexible machine learning methods can improve causal estimators, especially in the RCT setting.
Behavior genetics is a controversial science. For decades, scholars have sought to understand the role of heredity in human behavior and life-course outcomes. Recently, technological advances and the rapid expansion of genomic databases have facilitated the discovery of genes associated with human phenotypes like educational attainment and substance use disorders. To maximize the potential of this flourishing science, and to minimize potential harms, careful analysis of what it would mean for genes to be causes of human behavior is needed. In this paper, we advance a framework for identifying instances of genetic causes, interpreting those causal relationships, and applying them to advance causal knowledge more generally in the social sciences. Central to thinking about genes as causes is counterfactual reasoning, the cornerstone of causal thinking in statistics, medicine, and philosophy. We argue that within-family genetic effects represent the product of a counterfactual comparison in the same way as average treatment effects from randomized controlled trials (RCTs). Both average treatment effects from RCTs and within-family genetic effects are shallow causes: they operate within intricate causal systems (non-unitary), produce heterogeneous effects across individuals (non-uniform), and are not mechanistically informative (non-explanatory). Despite these limitations, shallow causal knowledge can be used to improve understanding of the etiology of human behavior and to explore sources of heterogeneity and fade-out in treatment effects.
This book seeks to narrow two gaps: first, between the widespread use of case studies and their frequently 'loose' methodological moorings; and second, between the scholarly community advancing methodological frontiers in case study research and the users of case studies in development policy and practice. It draws on the contributors' collective experience at this nexus, but the underlying issues are more broadly relevant to case study researchers and practitioners in all fields. How does one prepare a rigorous case study? When can causal inferences reasonably be drawn from a single case? When and how can policy-makers reasonably presume that a demonstrably successful intervention in one context might generate similarly impressive outcomes elsewhere, or if massively 'scaled up'? No matter their different starting points – disciplinary base, epistemological orientation, sectoral specialization, or practical concerns – readers will find issues of significance for their own field, and others across the social sciences. This title is also available Open Access.
Causal inference accordingly becomes more important in the very conditions that make it more difficult. When people feel less confident about discovering causes, they feel less able to control or manage their environment and correspondingly more at risk. They become insecure and vulnerable, and all the more so to the degree that their livelihood, status, or security depends on causal inference. People have strong instrumental and psychological motives to find ways around this dilemma. I identify and evaluate four generic strategies for coping with the causal dilemma in international relations theories
While a difference-in-differences (DID) design was originally developed with one pre- and one posttreatment period, data from additional pretreatment periods are often available. How can researchers improve the DID design with such multiple pretreatment periods under what conditions? We first use potential outcomes to clarify three benefits of multiple pretreatment periods: (1) assessing the parallel trends assumption, (2) improving estimation accuracy, and (3) allowing for a more flexible parallel trends assumption. We then propose a new estimator, double DID, which combines all the benefits through the generalized method of moments and contains the two-way fixed effects regression as a special case. We show that the double DID requires a weaker assumption about outcome trends and is more efficient than existing DID estimators. We also generalize the double DID to the staggered adoption design where different units can receive the treatment in different time periods. We illustrate the proposed method with two empirical applications, covering both the basic DID and staggered adoption designs. We offer an open-source R package that implements the proposed methodologies.
A core part of political research is to identify how political preferences are shaped. The nature of these questions is such that robust causal identification is often difficult to achieve, and we are not seldom stuck with observational methods that we know have limited causal validity. The purpose of this paper is to measure the magnitude of bias stemming from both measurable and unmeasurable confounders across three broad domains of individual determinants of political preferences: socio-economic factors, moral values, and psychological constructs. We leverage a unique combination of rich Swedish registry data for a large sample of identical twins, with a comprehensive battery of 34 political preference measures, and build a meta-analytical model comparing our most conservative observational (naive) estimates with discordant twin estimates. This allows us to infer the amount of bias from unobserved genetic and shared environmental factors that remains in the naive models for our predictors, while avoiding precision issues common in family-based designs. The results are sobering: in most cases, substantial bias remains in naive models. A rough heuristic is that about half of the effect size even in conservative observational estimates is composed of confounding.
Despite the tremendous renaissance of comparative constitutional law, the comparative aspect of the enterprise, as a method and a project, remains under-theorized and imprecise. Methodological self-awareness has not been one of the field’s strengths. In comparative constitutional law (and Constitutionalism in Context more generally) the term “comparative” is often used indiscriminately to describe what, in fact, are several different types of scholarship, each with its own meanings, aims and purposes. What is more, various vocational, jurisprudential, academic, and scientific stakeholders involved in practicing the art of constitutional comparison. This chapter will explore the various types, aims and methodologies deployed in exploring constitutional phenomena comparatively across time and space. In so doing, it will identify some gaps in the field’s contemporary methodological matrix and suggest ways in which these deficiencies may be addressed and overcome.
We consider the following question: given information on individual policyholder characteristics, how can we ensure that insurance prices do not discriminate with respect to protected characteristics, such as gender? We address the issues of direct and indirect discrimination, the latter resulting from implicit learning of protected characteristics from nonprotected ones. We provide rigorous mathematical definitions for direct and indirect discrimination, and we introduce a simple formula for discrimination-free pricing, that avoids both direct and indirect discrimination. Our formula works in any statistical model. We demonstrate its application on a health insurance example, using a state-of-the-art generalized linear model and a neural network regression model. An important conclusion is that discrimination-free pricing in general requires collection of policyholders’ discriminatory characteristics, posing potential challenges in relation to policyholder’s privacy concerns.
If case study comparison is useful in the social sciences, it should be at least as useful as a way of understanding the political preferences and participation of individuals as it is for larger and more complex social categories such as organizations, parties, militaries, or states. Thus, qualitative comparisons involving the political preferences and activity of American billionaires will serve as a benchmark for the plausibility of various comparative frameworks. The chapter will demonstrate that applying Mill’s method-type comparison to billionaires generates inferential absurdities and also shows that regression-type logics of control are a poor fit for individual-level qualitative analysis. More feasible frameworks for understanding qualitative comparison at the individual level come both earlier and later in the inferential process. Cross-case comparisons are highly valuable for concept formation and theory building prior to any systematic effort at causal inference and also play a meaningful role in inferences about moderation effects after process tracing has been successful within each case. Furthermore, multi-case qualitative analysis that is not structured as explicit comparison can make crucial contributions to multi-method research on individuals.
In light of important political events that go beyond the nation state (e.g., migration, climate change, and the coronavirus pandemic), domestic politicians are increasingly pressured to scrutinize and speak out on European policy-making. This creates a potential trade-off between allocating effort to domestic and supranational affairs, respectively. We examine how citizens perceive legislator involvement in European Union (EU) politics with a pre-registered conjoint experiment in Germany. Our results show that Members of Parliament (MPs) are not disadvantaged when allocating effort to European affairs as compared to local and national affairs. In addition, voters tend to prefer MPs who engage in EU policy reform over those who do not. As demand for legislator involvement in European politics is on the rise, we provide empirical evidence that MPs can fulfill this demand without being disadvantaged by the electorate.
In a recent article, I argued that the Bayesian process tracing literature exhibits a persistent disconnect between principle and practice. In their response, Bennett, Fairfield, and Charman raise important points and interesting questions about the method and its merits. This letter breaks from the ongoing point-by-point format of the debate by asking one question: In the most straightforward case, does the literature equip a reasonable scholar with the tools to conduct a rigorous analysis? I answer this question by walking through a qualitative Bayesian analysis of the simplest example: analyzing evidence of a murder. Along the way, I catalogue every question, complication, and pitfall I run into. Notwithstanding some important clarifications, I demonstrate that aspiring practitioners are still facing a method without guidelines or guardrails.
This paper proposes a Bayesian alternative to the synthetic control method for comparative case studies with a single or multiple treated units. We adopt a Bayesian posterior predictive approach to Rubin’s causal model, which allows researchers to make inferences about both individual and average treatment effects on treated observations based on the empirical posterior distributions of their counterfactuals. The prediction model we develop is a dynamic multilevel model with a latent factor term to correct biases induced by unit-specific time trends. It also considers heterogeneous and dynamic relationships between covariates and the outcome, thus improving precision of the causal estimates. To reduce model dependency, we adopt a Bayesian shrinkage method for model searching and factor selection. Monte Carlo exercises demonstrate that our method produces more precise causal estimates than existing approaches and achieves correct frequentist coverage rates even when sample sizes are small and rich heterogeneities are present in data. We illustrate the method with two empirical examples from political economy.
Regression discontinuity (RD) designs have become increasingly popular in political science, due to their ability to showcase causal effects under weak assumptions. This paper provides an intuition-based guide for the use of the RD in applied research. After an intuitive explanation of how the method works, we provide a checklist that can help researchers understand the main robustness checks they should run, and a quick introduction to software implementing the design. We also provide a list of classic designs and examples of their application in political science. We hope this article can constitute a stepping stone from which researchers interested in RD can jump to more advanced literature; and which makes researchers not interested in implementing RDs better consumers of research employing this design.