This README file applies to the paper titled "Online Speech and Communal Conflict: Evidence from India" by Sebastian Schutte, Daniel Karell, and Ryan Barrett. All questions and comments should be sent to sebastian@prio.org.

---

There are two files in our replication packages. The first file is "dataframe.Rda". This file contains the data needed to reproduce our main results, supplemental tests, and robustness checks. The second file is "repo_code.R". This coding script uses the dataframe to generate the results presented in various figures and tables in the main text and supplemental information of the paper.

The analysis was conducted using R version 4.3.0 (2023-04-21) -- "Already Tomorrow". The script "repo_code.R" employs the "groundhog" package (version 3.2.0) to ensure that any future replication uses the versions of the other packages we used while conducting the analysis (which occurred during 2023). The code should complete in less than two minutes on standard hardware.

Note that the results generated by the coding script are not in the same order as they are presented in the paper. 

Please also note that Table S10 in the Supplemental Information document was constructed "manually" by using the results generated in Lines 214-299 of "repo_code.R". This was the most efficient way of constructing the table because of its necessary layout.