This repo contains the replication material for “Can Human Reading Validate a Topic Model?”
We first provide replication materials for the findings documented in the main text. Codes to replicate the online appendix will be added later when they are re-organized.
Files should be run in the sequence indicated by their filenames.
You do not necessarily run the model estimation in "3_surveying_and_alignment.R". It takes time and we already saved related local models for you.