mpleung / dependence_robust Goto Github PK

Replication files for "Dependence-Robust Inference Using Resampled Statistics". The files were coded for Python 3 and require the following dependencies: networkx, numpy, pandas, powerlaw, scipy, and snap.

Contents:

• RS_module.py: Functions implementing our methods.
• gen_net_module.py: Functions for simulating data.
• jackson_rogers_application.py: Empirical application. Prints output as latex table directly to console.
• jackson_rogers_data: Data for empirical application. Obtained from http://www.stanford.edu/~jacksonm/JacksonRogers-Data.zip.
• clustering.py: Clustered data monte carlo (Tables 2 and 3). Prints output as latex table and also saves as CSV in this directory.
• clustering_strongdep.py: Clustered data with strong dependence monte carlo (results verbally summarized in section 6). Prints output as latex table directly to console and also saves as CSV in this directory.
• node_stats.py: Network statistics monte carlo (Tables B.1 and B.2). Prints output as latex table directly to console and also saves as CSV in this directory. First run with simulate_only=True. The rerun with simulate_only=False to get the results.
• power_law.py: Power law monte carlo (Table B.4). Prints output as latex table directly to console and also saves as CSV in this directory.
• tspill.py: Treatment spillovers monte carlo (Table B.3). Prints output as latex table directly to console and also saves as CSV in this directory. First run with simulate_only=True. The rerun with simulate_only=False to get the results.