**1 README 1.1 Overview** This document describes the replication package for Di Falco and Vieider (2022a). The replica- tion folder contains 6 data files and 3 analysis files. We describe each combination of data and analysis file in turn: • The main panel data used for analysis are contained in the file data.csv. The results in the main paper can be reproduced using the file analysis_Stata.do. Note that the analysis file needs to be set to the local directory to run. The analysis was carried out on Stata 15 on Max OS 11.1. Some outputs may not display properly in older Stata versions. The analysis file requires the user-written packages esttab , colrspace, grstyle, and palettes. The analysis file contains installation code for the latter three. The analysis produces all the tables in the main paper, but tables 1 and 2 are not automatically exported and need to be created by hand. It also produces and exports figures 8, 9, and 10. The expected running time is approximately 22 minutes, but may vary depending on hardware. • All figures except for 8, 9, and 10, which are generated by the Stata code described above, are produced by the file generate\_figures.R. The only exceptions are the maps in figure 1 and figure S1, which requires GPS data that cannot be shared due to EU privacy protection law. The code uses all the different data files to generate different output, in particular it uses data.csv , data\_idio.csv, data\_models.csv, and data\_rain\_historical.csv. The output was generated in RStudio version 1.3.616, R version 3.5.2 on Mac OS X version 10.14.6. It requires the packages tidyverse, data.table , rstan, rethinking , and ggrepel. The expected running time is 30 seconds, but may vary depending on hardware. • The data file data_crop_yields.csv contains plot-wise data on agricultural yields. These data are analyzed in the file analysis_yield_data.do to reproduce the analysis in the Online Appendix. The analysis was carried out on Stata 15 on Max OS 11.1. Some outputs may not display properly in older Stata versions. The analysis file requires the user-written packages esttab , colrspace, grstyle, and palettes. The expected running time is approximately 6 minutes, but may vary depending on hardware. All tables will be deposited in the tables subfolder. Figures will be deposited in the figures subfolder. **1.2 Data Availability and Provenance Statement** The data are available at the following source: Di Falco and Vieider (2022b). The main data on risk preferences are primary data collected in individual-level experiments with farming house- holds in rural Ethiopia. One of the decisions was paid for real money according to the instructions available in the Online Appendix. These experimental data were combined with survey data, obtained in individual interviews, and with historical rainfall data. The posted datasets include all data collected except for the GPS data used to identify the households, which may not be stored according to EU data storage and privacy regulations. **References** Di Falco, S. and Vieider, F.M. (2022a). ‘Environmental adaptation of risk preferences’, Economic Journal, forthcoming. Di Falco, S. and Vieider, F.M. (2022b). ‘Environmental adaptation of risk preferences’, Replica- tion Data: https://osf.io/b3sr7/.