You should also be cautious any good news about seemingly improving graduation rates — especially when it is being reported that the numbers are better than graduation rate numbers of more than three decades ago. One reason lies with the fact that, as with so much in education, there was little in the way of accurate graduation data until At the same time, you should ignore dropout rates altogether as a measure of how well states and districts are doing in improving student achievement. Why? Because those numbers often only account for those struggling kids who officially dropped out in a given year, and not those who are shunted out of traditional high schools into alternative school ghettos where they waste away educationally until they are too old under law to attend high school, usually around age 21 or so.

wpid10020-wpid-this_is_dropout_nation_logo2.pngSo when media outlets breathlessly quoted the U.S. Department of Education’s press release yesterday that the new graduation rate data released by the U.S. Department of Education’s National Center for Education Statistics was supposedly “the highest” since 1974, your editor was immediately skeptical. And not just because NCES’ report is provisional, or essentially a first draft of what will finally be released later on. The graduation rate now being reported by NCES, called the Average Freshmen Graduation Rate, is fundamentally different from the graduation rate calculation the agency had been using until a decade ago.

For years, NCES culled what is called Status Completion Rates from the Current Population Study, the annual survey conducted by the U.S. Census Bureau. The survey wasn’t very good at providing accurate graduation rate data. It merely surveyed a sample of 18-to-24 year olds who told Census Bureau staffers whether they graduated or not. Given that the surveyed adults didn’t have to show a sheepskin of any kind, one could simply say they “graduated” from high school, even if they only had a GED (which, contrary to Chris Rock’s famed joke, is hardly a “good enough diploma”) or worse, had dropped out of high school before getting one. As a result, NCES would report (as it did for the nation’s Class of 2000) that 87 percent of high school freshmen graduated — even if it wasn’t exactly so. Not that NCES could get the data in the first place; in 1989, NCES began coaxing states into developing accurate and uniform high school dropout data two years after it found that they varied wildly in how they reported such data.

The better approach would have been to report graduation rates based on the percentage of high school freshmen who eventually left high school with a diploma in four years. NCES itself began reporting dropout rates using this approach back in 1992 (and has been further refined since). Starting in 2002, it began report so-called high-school completer rates, breaking separating the high school graduates from those who received GEDs and so-called certificates of completion given to high school seniors who didn’t have enough credits to graduate. But it took the passage of the No Child Left Behind Act, and its elevation of graduation rates (along with student test score data) as the primary measures of success in improving student achievement, to force NCES in the right direction. Researchers such as Jay P. Greene (now of the University of Arkansas) and then-Urban Institute scholar Christopher Swanson (now the head of the research unit of Education Week‘s parent company) that states were inflating their high school completion numbers also played its part, as did revelations by NCES that only 10 states could provide high school completion data from the 1994-1995 to 1997-1998 school years.

Simply put, there’s no way that anyone can know with any certifiable accuracy how many high school graduated 100 years ago, much less back in the 1970s. While researchers, most-notably Nobel Laureate James Heckman and recently, Richard Murname of Harvard University, have attempted what can be best called guestimates of graduation rates, there are flaws to their efforts. Murname’s study, for example, uses data from the National Educational Longitudinal Study of 1988, which is merely a sample of students that is hardly representative of the nation’s student population as a whole; for example, special ed students were excluded from NELS88, so one could not extrapolate how many kids in special ed ghettos managed to leave school with a diploma. Ultimately, we don’t know how many kids actually graduated before, at best, before the 1980s (when states such as Indiana began collecting and disseminating enrollment and diploma recipient data, which researchers can then use to deduce graduation rates). Simply put, the Department of Education was rather irresponsible in issuing that statement, and reporters were equally thoughtless in running with the spin without doing a little leg work.

All that said, the NCES report itself doesn’t make any claims about graduation rates being higher or lower than in the 1970s. If anything, the latest NCES report does offer some important information on what states are not doing when it comes to reporting accurate graduation rate data. Connecticut was exposed by federal officials for submitting high school diploma numbers that were “high compared to other states and data from earlier years”. How high? The Nutmeg State was reporting a graduation rate of 98 percent for its Class of 2010, higher than the 80 percent graduation rate reported for the Class of 2009; Connecticut’s numbers were also 10 percent higher than the number of high school seniors in the final year of class. Little wonder why Connecticut officials declined to confirm the accuracy of their data. NCES also suppressed school district dropout reports from Kentucky, Maine, and Mississippi because they accounted for less than 85 percent of what was reported by states, and refused to accept unusually low dropout rates for either the District of Columbia (whose numbers would seemingly show that only one percent of students officially drop out over a four-year period) or for Puerto Rico. School reformers in those states and territories now know that they need to push those states to improve their school data system; pushing those states to move to collecting individual student-level data instead of collecting aggregate information from schools and districts would certainly improve accuracy of data being reported.

Just as importantly, there has been improvement in graduation rates. As EdWeek reported last year in its annual Diplomas Count study, the nation’s graduation rate increased from 66 percent for the Class of 1999 to the 73 percent for the Class of 2009; this means that 100,000 fewer high school graduates dropped out into poverty in prison in 2009 than a decade ago. Dropout Nation‘s own Five-Year Promoting Power estimates based on eighth-grade enrollment (a version of the index developed by Johns Hopkins researcher Robert Balfanz) shows that four percent fewer eighth-graders dropped out between 2005 and 2009, than between 1995 and 2009 — even as the percentage of eighth graders attending school increased by 16 percent within the decade. Even if one assumes that NCES’ report of a 78 percent graduation rate is inflated (largely because of gamesmanship by states, including decisions to report students graduating with diplomas from district-run GED programs, and the creation of low expectations-driven diploma tracks), the good news is that two decades of reforms that culminated with No Child’s passage in 2001 have helped more kids stay off out of economic and social despair.

But simply graduating more high school students isn’t enough. As I noted last June in a Dropout Nation Podcast, far too many high school graduates have not gotten the strong, comprehensive college preparatory curricula they need to graduate from a traditional college, technical school, apprenticeship or other form of higher education. Only 13 percent of American high school seniors of all backgrounds were taking college-preparatory courses, while another 25 percent took high school coursework that didn’t involve any form of algebra or other form of college preparation, according to data from the 2009 National Assessment of Educational Progress. Thanks to Zip Code Education policies and the condescending view among many traditionalists that there are some kids (usually black, Latino, Native American, or poor white and Asian) who are just not capable of college-preparatory work, many children are being subjected to educational neglect and malpractice that is intolerable.

There is no way we can know how many high school freshmen graduated in 1974, or even in 1924. What we do know is that graduating from high school alone isn’t good enough in the increasingly knowledge-based economy. And that our children deserve better.