Metrics in Action: How Social Media Metrics Shape News Production on Facebook

Replication scripts for the Metrics in Action project by Subhayan Mukerjee, Tian Yang, and Yilang Peng.

To run these scripts

• change the data_folder variable to reflect the path to the folder that contains all the .csv files
• create a subfolder called results/ outside the scripts folder
• create a subfolder called figures/ outside the scripts folder

Your working directory should have three subfolders in it:

• scripts/ containing all the scripts in the repository
• results/ which will contain the results (as .tex tables) after the scripts are executed
• figures/ which will contain the plots (as .svg images) after the scripts are executed

Details about the scripts are as follows:

• 01_longitudinal-analysis.R creates figure 1 (in the main text).
• 02_main-analysis.R creates table 1 and figures 2 and 3 (in the main text).
• 03_test-stationarity.R tests the stationarity assumption of the time series.
• 04_robustness-with-10-periods.R creates figures A2 and A3 (in the Appendix).
• 05_robustness-with-alternative-ideology-score.R creates figures A4, A5, and A6 (in the Appendix)
• 06_robustness_with_quadratic_model.R creates table A2 (in the Appendix)
• 07_robustness-with-post-freq.R creates table A3 (in the Appendix)
• 08_robustness-with-interaction-analysis.R creates table A4 (in the Appendix)
• 09_robustness-with-shares-only.R creates table table A5 (in the Appendix)

Details of which script produces which table and figure are also documented at the beginning of each script.