The process to recreate the measure entirely:

1. Download all the C/CES data and place the data files into the folders in the data/ces_data folder.
2. Run ces_data/data_clean.R script which will extract all the requisite survey questions about ideology and place them into the cleaned folder.
3. You can then run data/providing_ids.R which will create a single giant csv with all the data.
4. Run data/dynamic_prep.R which will generate two files into the model folder. The model_data.json file will be used by stan, and the candidate_data.csv is used to connect data at the end.
5. The actual estimation can be accomplished with model/estimation.R. I run this on the Ohio Super Computer and so the model/run.sh is used to do that. This will produce several very large files that contain the entire posterior.
6. To summarize the posterior you can use model/outputs.R which is a very simple script (this takes a lot of RAM). This saves summary stats of the posterior as summary.csv.
7. Finally to produce useful outputs run posterior/post_prep.R which will connect everything together. You will have to make sure that the summary.csv file is in the right spot, and that you’ve downloaded all the supplementary files to data/supplementary_data.