10.7910/DVN/ZVC9W5 Liu, Licheng Licheng Liu https://orcid.org/0000-0002-9183-163X Massachusetts Institute of Technology Wang, Ye Ye Wang https://orcid.org/0000-0001-5127-4245 Xu, Yiqing Yiqing Xu https://orcid.org/0000-0003-2041-6671 Stanford University Harvard Dataverse 2022 Social Sciences Imputation methods Counterfactual estimators Twoway fixed effects Parallel trends Interactive fixed effects Matrix completion Equivalence tests Placebo tests Time-series cross-sectional data Panel data Xu, Yiqing Yiqing Xu Xu, Yiqing Yiqing Xu 2022-03-02 2022-05-02 1355 672 4173 3021 279311 116133 9518 3649 323610 8067 230600 660 19285 4414 26583510 320453 23186 937 506086 14684436 3084 23154 8428 5029 15537 112 1818360 610008 5218 4646569 5583 0 137921 268852 type/x-r-syntax text/plain text/plain type/x-r-syntax text/tab-separated-values application/gzip type/x-r-syntax type/x-r-syntax application/x-rlang-transport type/x-r-syntax application/x-rlang-transport text/plain type/x-r-syntax type/x-r-syntax application/x-rlang-transport application/x-rlang-transport application/x-rlang-transport text/plain application/gzip application/x-rlang-transport type/x-r-syntax application/x-rlang-transport text/plain type/x-r-syntax text/plain text/plain application/gzip text/tab-separated-values type/x-r-syntax application/x-rlang-transport type/x-r-syntax text/plain application/pdf application/pdf 1.0 Custom terms specific to this dataset This paper introduces a simple framework of counterfactual estimation for causal inference with time-series cross-sectional data, in which we estimate the average treatment effect on the treated by directly imputing counterfactual outcomes for treated observations. We discuss several novel estimators under this framework, including the fixed effects counterfactual estimator, interactive fixed effects counterfactual estimator, and matrix completion estimator. They provide more reliable causal estimates than conventional twoway fixed effects models when treatment effects are heterogeneous or unobserved time-varying confounders exist. Moreover, we propose a new dynamic treatment effects plot, along with several diagnostic tests, to help researchers gauge the validity of the identifying assumptions. We illustrate these methods with two political economy examples and develop an open-source package, fect, in both R and Stata to facilitate implementation. This dataset underwent an independent verification process that replicated the tables and figures in the primary article. For the supplementary materials, verification was performed solely for the successful execution of code. The verification process was carried out by the Odum Institute for Research in Social Science at the University of North Carolina at Chapel Hill. <br></br> The associated article has been awarded Open Materials and Open Data Badges. Learn more about the Open Practice Badges from the <a href="https://osf.io/tvyxz/wiki/home/" target="_blank">Center for Open Science</a>.<br></br> <img src="https://odum.unc.edu/files/2020/03/OpenData_PR-1.png" alt="Open Data Badge" height="77" width="80"> <img src="https://odum.unc.edu/files/2020/03/OpenMaterials_PR-1.png" alt="Open Materials Badge" height="77" width="80">