Arun Advani, Toru Kitagawa, and Tymon Sloczynski, "Mostly Harmless Simulations? Using Monte Carlo Studies for Estimator Selection", Journal of Applied Econometrics, Vol. 34, No. 6, 2019, pp. 893-910. All files are ASCII files in DOS format and are zipped in aks-files.zip. Unix/Linux users should use "unzip -a". The folders ASCII and Stata contain, respectively, the original datasets that we used and the computer programs that produce all tables of the paper and the supplementary appendix. The folder ASCII contains six datasets: cps_controls, nswera_control, nswera_treated, nswre74_control, nswre74_treated, and psid_controls. cps_controls and psid_controls contain information on CPS and PSID comparison units, respectively. nswre74_treated and nswre74_control are the subsamples of the NSW treated and control units, respectively, that were created by Dehejia and Wahba (1999). nswera_treated and nswera_control are the subsamples of the NSW treated and control units, respectively, that were created by Smith and Todd (2005). The folder Stata contains six further folders: Scenario_1, Scenario_2, Scenario_3, Table_1, Table_2, and Table_3. The computer programs in Scenario_1 produce the results reported in Subsubsection 3.2.1 and in Table E1 of the supplementary appendix. The computer programs in Scenario_2 produce the results reported in Subsubsection 3.2.1 and in Table E2 of the supplementary appendix. The computer programs in Scenario_3 produce the results reported in Subsubsection 3.2.2 and in Table E3 of the supplementary appendix. The computer programs in Table_1 produce the results reported in Table 1 as well as in Table F1 of the supplementary appendix. The computer programs in Table_2 produce the results reported in Table 2 as well as in Table F2 of the supplementary appendix. The computer programs in Table_3 produce the results reported in Table 3 as well as in Table F3 of the supplementary appendix. Each of the six folders contains several Stata do-files as well as a number of further subfolders. To replicate the results reported in the paper, the do-files should be run in the order in which they are numbered; e.g., 1_samples, 2_estimates, etc. When there is a break in the numbering, the actual Monte Carlo simulations need to be run. The corresponding Stata do-files are stored in subfolders whose names consist of "bootstrap", "placebo", or "structured" as well as two digits; e.g., bootstrap06, placebo03, or structured08. Tymon Sloczynski Assistant Professor of Economics, Brandeis University Research Fellow, Institute of Labor Economics (IZA) Phone: (+1) 781-736-8550 E-mail: tslocz [AT] brandeis.edu Website: http://people.brandeis.edu/~tslocz/