In the past several years, a slew of reports on graduation rates have been issued from such organizations as the Urban Institute, Education Week, the Economic Policy Institute (EPI), the U.S. Census Bureau, the National Governors Association (NGA), the Manhattan Institute and the National Center for Education Statistics (NCES).

With so many reports from such a variety of institutions, opinions naturally vary about which source to use, the appropriate methodology and how to calculate the numbers. The debate hinges largely on what data yield the most accurate estimates and how or whether to account for transfers, retentions, dropouts/drop-ins and the incarcerated population.

A variety of databases are used. The U.S. Department of Education uses the Common Core of Data (CCD) and national longitudinal studies, such as the National Education Longitudinal Study (NELS) and the High School and Beyond (HS&B) survey. The CCD is an annual census of all public schools with uniform definitions for data across states, but it misses information about in-grade retention and transfers.

The Current Population Survey (CPS) from the Bureau of Labor Statistics, the Department of Labor, and the U.S. Census is another source. This is an authoritative supplier of economic indicators, a long-standing source of data on adult educational attainment. Yet, it is a sample survey with coverage problems for some populations (incarcerated, transient) and cannot reliably distinguish diplomas from GEDs. Additionally, state-reported statistics are used, but these are self-reported and not comparable across states (and are consistently higher than independent research). Researchers who study the issue use all or some of these sources, which adds to the differing results.

According to various reports, the national graduation rate could be anywhere from approximately 70 percent to 86 percent (see Figure 1). This means that anywhere from about 500,000 to 1.2 million members of the class of 2003 dropped out before receiving their standard diplomas. A spring 2006 report from the Economic Policy Institute (Mishel and Roy), which calculates a much higher graduation rate than most other estimates, has especially generated controversy and stirred up public debate on this topic. The only consensus is that a precise and definitive calculation of the graduation rate is impossible until there is a system to track each student in the country from the time he or she starts school until he or she finishes. This, in turn, has reinvigorated calls for states to create longitudinal data systems and for the federal government to help them do so.