When Jon Krosnick and I were principal investigators of the American National Election Studies (ANES), we learned a number of very important lessons about coding open-ended questions and in this chapter I hope to convey those lessons. What I'm going to describe is based on the work and experiences of a lot of different people. I'll start by outlining some background and describe some challenges in the domain of open-ended coding. I'll tell you about an example that was our trial by fire in this domain. Then, I'll talk about some general attributes of the approach that we have tried to develop and are taking at the ANES to make open-ended coding something that's more legitimate and credible than it has been.