This paper reviews and provides recommendations on making causal claims in social science research. Authors John Antonakis, Samuel Bendahan, Philippe Jacquart, and Rafael Lalive discuss the conditions under which causal inferences can be made from non-experimental data. They emphasize that causal claims are often based on correlational data where the independent variable is not exogenously manipulated. The paper highlights that endogeneity—such as omitted variables, selection bias, simultaneity, common-method bias, and measurement error—can render estimates causally uninterpretable. The authors present methods for testing causal claims when randomization is not possible, including fixed-effects panel, sample selection, instrumental variables, regression discontinuity, and difference-in-differences models. They review 110 leadership-related articles from top-tier journals over the past decade and find that researchers fail to address up to 90% of design and estimation conditions that make causal claims invalid. The paper argues that researchers must not shy away from making causal claims, as they are crucial for society. The randomized experiment is the gold standard for causal inference, but it is often infeasible in social science settings. However, recent methodological advances allow researchers to make causal inferences in the field. The authors provide 10 recommendations for improving non-experimental research, emphasizing the importance of addressing endogeneity and ensuring methodological rigor. The paper also discusses the importance of understanding causality in non-experimental settings, the challenges of endogeneity, and the need for appropriate statistical methods to test causal claims. It highlights the importance of controlling for omitted variables, fixed effects, and measurement error to ensure consistent estimates. The authors conclude that causal inference in non-experimental settings requires careful consideration of design and estimation conditions, and that researchers must use appropriate methods to ensure the validity of their findings.This paper reviews and provides recommendations on making causal claims in social science research. Authors John Antonakis, Samuel Bendahan, Philippe Jacquart, and Rafael Lalive discuss the conditions under which causal inferences can be made from non-experimental data. They emphasize that causal claims are often based on correlational data where the independent variable is not exogenously manipulated. The paper highlights that endogeneity—such as omitted variables, selection bias, simultaneity, common-method bias, and measurement error—can render estimates causally uninterpretable. The authors present methods for testing causal claims when randomization is not possible, including fixed-effects panel, sample selection, instrumental variables, regression discontinuity, and difference-in-differences models. They review 110 leadership-related articles from top-tier journals over the past decade and find that researchers fail to address up to 90% of design and estimation conditions that make causal claims invalid. The paper argues that researchers must not shy away from making causal claims, as they are crucial for society. The randomized experiment is the gold standard for causal inference, but it is often infeasible in social science settings. However, recent methodological advances allow researchers to make causal inferences in the field. The authors provide 10 recommendations for improving non-experimental research, emphasizing the importance of addressing endogeneity and ensuring methodological rigor. The paper also discusses the importance of understanding causality in non-experimental settings, the challenges of endogeneity, and the need for appropriate statistical methods to test causal claims. It highlights the importance of controlling for omitted variables, fixed effects, and measurement error to ensure consistent estimates. The authors conclude that causal inference in non-experimental settings requires careful consideration of design and estimation conditions, and that researchers must use appropriate methods to ensure the validity of their findings.