The estimation of causal effects from observational data is a critical challenge in social science research, as experimental designs are often infeasible. Christopher Winship and Stephen L. Morgan review the major developments in the past two decades by statisticians and econometricians in this area. They introduce the counterfactual framework, which provides a conceptual basis for understanding causal effects. This framework posits that each individual has two potential outcomes: one under treatment and one under control. However, only one of these outcomes is observed for each individual, making direct estimation of causal effects difficult. The paper discusses various methods for estimating causal effects from observational data, including regression, matching, and propensity score techniques. It highlights the importance of controlling for confounding variables and the limitations of standard estimators when treatment assignment is not random. The authors emphasize that the average treatment effect is not always the most relevant measure, as the average treatment effect for the treated may be more important in policy contexts. The paper also addresses the issue of selection bias, which arises when treatment assignment is correlated with potential outcomes. This can lead to biased estimates of causal effects. The authors discuss methods to address this bias, including the use of instrumental variables and the analysis of covariance. They argue that while these methods can help reduce bias, they are not foolproof and require careful consideration of assumptions. The paper concludes by emphasizing the importance of these methods in improving the quality of quantitative empirical research in sociology. It highlights the need for researchers to understand the assumptions underlying different methods and to use them appropriately. The authors also note that while there are many techniques available, no single method is universally applicable, and the choice of method depends on the specific context and data available. Overall, the paper provides a comprehensive overview of the current state of research on estimating causal effects from observational data and offers valuable insights for researchers in the field.The estimation of causal effects from observational data is a critical challenge in social science research, as experimental designs are often infeasible. Christopher Winship and Stephen L. Morgan review the major developments in the past two decades by statisticians and econometricians in this area. They introduce the counterfactual framework, which provides a conceptual basis for understanding causal effects. This framework posits that each individual has two potential outcomes: one under treatment and one under control. However, only one of these outcomes is observed for each individual, making direct estimation of causal effects difficult. The paper discusses various methods for estimating causal effects from observational data, including regression, matching, and propensity score techniques. It highlights the importance of controlling for confounding variables and the limitations of standard estimators when treatment assignment is not random. The authors emphasize that the average treatment effect is not always the most relevant measure, as the average treatment effect for the treated may be more important in policy contexts. The paper also addresses the issue of selection bias, which arises when treatment assignment is correlated with potential outcomes. This can lead to biased estimates of causal effects. The authors discuss methods to address this bias, including the use of instrumental variables and the analysis of covariance. They argue that while these methods can help reduce bias, they are not foolproof and require careful consideration of assumptions. The paper concludes by emphasizing the importance of these methods in improving the quality of quantitative empirical research in sociology. It highlights the need for researchers to understand the assumptions underlying different methods and to use them appropriately. The authors also note that while there are many techniques available, no single method is universally applicable, and the choice of method depends on the specific context and data available. Overall, the paper provides a comprehensive overview of the current state of research on estimating causal effects from observational data and offers valuable insights for researchers in the field.