The No Child Left Behind Act requires states to publish high school graduation rates for public schools and the U.S. Department of Education is currently considering a mandate to standardize high school graduation rate reporting. However, no consensus exists among researchers or policy-makers about how to measure high school graduation rates. In this paper, we use longitudinal data tracking a cohort of students at 82 Texas public high schools to assess the accuracy and precision of three widely-used high school graduation rate measures: Texas's official graduation rates, and two competing estimates based on publicly available enrollment data from the Common Core of Data. Our analyses show that these widely-used approaches yield inaccurate and highly imprecise estimates of high school graduation and persistence rates. We propose several guidelines for using existing graduation and persistence rate data and argue that a national effort to track students as they progress through high school is essential to reconcile conflicting estimates.