- Introduction: Doing Philosophy of Social Science
- Micro, Macro, and Mechanisms
- Mechanisms, Causal Modeling, and the Limitations of Traditional Multiple Regression
- Process Tracing and Causal Mechanisms
- Descriptive-Causal Generalizations: “Empirical Laws” in the Social Sciences?
- Useful Complex Causality
- Partial Explanations in Social Science
- Mechanistic Social Probability: How Individual Choices and Varying Circumstances Produce Stable Social Patterns
- The Impact of Duhemian Principles on Social Science Testing and Progress
- Philosophy and the Practice of Bayesian Statistics in the Social Sciences
- Sciences of Historical Tokens and Theoretical Types: History and the Social Sciences
- RCTs, Evidence, and Predicting Policy Effectiveness
- Bringing Context and Variability Back into Causal Analysis
- The Potential Value of Computational Models in Social Science Research
- Models of Culture
- The Evolutionary Program in Social Philosophy
- Cultural Evolution: Integration and Skepticism
- Coordination and the Foundations of Social Intelligence
- Making Race Out of Nothing: Psychologically Constrained Social Roles
- A Feminist Empirical and Integrative Approach in Political Science: Breaking Down the Glass Wall?
- Social Constructions of Mental Illness
- Cooperation and Reciprocity: Empirical Evidence and Normative Implications
- Evaluating Social Policy
- Values and the Science of Well-Being: A Recipe for Mixing
Abstract and Keywords
This article describes the methods for modeling causal effects in observational social science. It considers the capacity of new graphical methods to represent and then motivates models that can effectively deliver estimates of the underlying heterogeneity of causal effects. The major advancements that have allowed scholarship to move beyond simple regression models are also elaborated. There are two basic goals of writing down a causal graph: (1) to present the set of causal relationships implied by the available state of knowledge, (2) to evaluate the feasibility of alternative estimation strategies. Causal graphs can obscure important distinctions precisely due to their flexibility. Thus, their flexibility enables careful and precise consideration of the challenges of causal effect identification, separated in helpful ways from many specification issues that are less fundamental.
Stephen Morgan is professor of sociology and the director of the Center for the Study of Inequality at Cornell University. He has a PhD in sociology from Harvard University and an MPhil in comparative social research from Oxford University. He has published two books: On the Edge of Commitment: Educational Attainment and Race in the United States (Stanford University Press, 2005) and, coauthored with Christopher Winship, Counterfactuals and Causal Inference: Methods and Principles for Social Research (Cambridge University Press, 2007).
Christopher Winship is the Diker-Tishman Professor of Sociology at Harvard University and a member of the Harvard Kennedy School of Government’s senior faculty. Prior to coming to Harvard in 1992, he was a professor in sociology, statistics, and (by courtesy) economics. At Harvard he is a member of the criminal justice program, inequality program, and the Hauser Center for the Study of Nonprofits. His research interests include models of selection bias, causality, youth violence, pragmatism, and the implications of the cognitive revolution for sociology. With Stephen Morgan, he is author of Counterfactuals and Causal Inference (Cambridge, 2007).
Access to the complete content on Oxford Handbooks Online requires a subscription or purchase. Public users are able to search the site and view the abstracts and keywords for each book and chapter without a subscription.
If you have purchased a print title that contains an access token, please see the token for information about how to register your code.