In its regulatory attention on school-level graduation, the Every Student Succeeds Act (ESSA) continues what is historically a federal effort to define graduation and dropping out. These efforts rely on methodology which may not accurately represent the measured outcome. The purpose of this research article is to examine the historical accuracy of graduation-rate proxies commonly used before the federal definition of the adjusted-cohort graduation rate (ACGR) in 2008. Our findings suggest the historical methodologies used to measure graduation have been less sensitive than necessary to estimate district-level graduation accurately. This complicated relationship should give policymakers pause about over-reliance on insensitive measures believing that "better than nothing" is plausibly accurate. We conclude with implications for policy and recommendations for research.