Reflections

Coming into this summer, I expected the technical side (cleaning data, building a matching algorithm, coding interviews) to be the hard part. It was, but not in the way I expected. The harder part was the hurried, important decisions made within these "technical" tasks, such as whether a non-vetted household still counted as a valid control, how to frame a question so it didn't lead the respondent toward the answer I expected, and how to think long-term about each step of a protocol. Good research turned out to be less about finding the right analysis and more about being honest about what the data could and couldn't tell me.