How school districts measure homelessness among their students has implications for accountability and funding, as well as for supporting student success. Yet, measuring homelessness among high school students is challenging because students move in and out of experiencing it. Using administrative student-level data from a mid-sized public school district in the southern United States, we show that different commonly used procedures to measure which students are considered homeless can yield markedly different estimates of high school graduation rates for these students. This is largely because of differences in how districts classify students who experience homelessness but later become housed. To address the potentially negative effects of housing insecurity on academic achievement, it is important to first identify a common way to diagnose the problem.