In this paper, we explore the challenges inherent in conducting research with a hidden population -- how to conduct research with teachers who have left the classroom. Capturing the storied experiences of this group is vital to understanding how to effectively recruit, prepare, support, and sustain teachers in US classrooms for our next generation of students. Although not typically considered vulnerable, this population of "teacher leavers" remains hidden-by-choice because of a voluntary severing of ties with former schools and school districts. In this paper, we chronicle our efforts to identify an appropriate sampling method for this population, our selection of snowball sampling because of the hidden nature of individuals, and our efforts to use snowball sampling to construct a national sample of teacher leavers. We conclude by identifying the resulting areas of balance and imbalance in our sample, their possible explanations, and their subsequent implications for informing thoughtful sampling methodology.