General

• Communicate with me and be pro-active: If you see a better way of doing things than I have proposed, tell me. If you find mistakes in my work, tell me.
• Don’t let assignments languish. If you run into a roadblock, contact me quickly so we can keep moving. 
• We may set up a regular meeting time. You should be working steadily in the interim, not cramming in the work right before the meeting.
• Do not communicate with outside entities (e.g., units of government, funders, data administrators, media, and so forth) unless we have discussed prior. When in doubt, ask first.

Data Handling & Programming

Work is to be carried out in Stata except for special situations where there are clear advantages to working in another program. Discuss with me first.

All work goes in the shared folder, preferably Box. It is critical to keep the folder orderly.

• Create a folder for all your work. Share this file with me so I can access all of your work. Do not keep work product in your ‘private’ folders, nor on your own laptop (backups/working copies excepted). Never write over the files in another RA’s folder. Keep all of your work in your own RA folder. The first set of subfolders in your main folder should be project-level.
  • Within projects, it is usually most orderly to place data in one file, do files in another, and other information (e.g., PDFs with policy info, literature reviews, and so forth) in another. Within these folders, there may be many subfolders.
  For every folder, make a “Readme” Word Document that describes the content of the folder and how the files in the folder relate to the project and to each other. They should also contain contextual information about things like data sources. The Readme should describe file-naming conventions. Indicate the date that a Readme.doc was last updated in the document. Great care needs to be taken with do files.
  • Give do files intuitive names (i.e., they relate to the task and the topic – e.g., “DataPrep.do” or “MainAnalysis.do”). Include copious comments within every do file. Indicate the project the do file belongs to, describe the data, describe the steps, etc… When in doubt, err on the side of making “too many” notes. It often makes sense to have multiple do files, but always provide a “master.do” file that runs the individual do files in the correct order. It should be possible to replicate the project from beginning (data construction) to end (final tables) by double-clicking “master.do”.