Data Errors
by

More than 90 percent of New Jersey’s high school freshmen graduate on time. That’s according to statistics reported by the state to the public and to the U.S. Department of Education, but it isn’t so. At best, just eight in ten freshmen graduate on time and even those projections may too high. Dropouts who get General Educational Development diplomas can also receive regular high school diplomas if they score at least 225 on all the required tests.

New Jersey isn’t the only state in which the number of students earning diplomas is given an overly — shall we say? — optimistic spin. America is in the midst of an educational crisis, with high numbers of high school dropouts and low academic achievement. More often than not, school performance data reported by state and federal officials fails to convey this reality.

Because of bad reporting, it can be difficult to get a handle on how poorly traditional public schools are doing in teaching children, keeping them safe, or knowing if they are even attending school. This makes it more difficult for school reformers, especially those in the standards-and-accountability movement, to get parents to embrace the stiff medicine that they tout as necessary reforms.

This spotlight on data quality comes courtesy of the No Child Left Behind Act, whose accountability rules have spawned and revived an array of statistical measurements. For the National Education Association and suburban school districts, which don’t want to surrender their control over curricula, the federal law has been a constant source of consternation. The standards movement sees these measurements as keys to bolstering their arguments.

For reform-minded education researchers of all ideological stripes, No Child has not only given them the ability to fully evaluate school performance, but even get a handle on how states and school systems track student achievement. What they have learned proves the old adage that there is often little difference between lies and statistics.

TAKE GRADUATION RATES, which, along with the test scores, are mandated by No Child as key measures of achievement. Forty-seven states reported graduation rates in 2003 that were far higher than reality, according to the Education Trust in a 2005 report.

In 2006, 25 states didn’t report graduation rates for four categories of students as mandated under No Child. And some states and local school districts report numbers that are suspect. Indianapolis Public Schools in Indiana reported that 33 percent more 12th graders graduated from its woeful dropout factories in the 2005-06 school year than were actually enrolled. Texas may have artificially boosted its 2005 graduation rate by labeling 12,700 missing students as “data errors.”

The national numbers issued by the federal government are not much better. The 88 percent national graduation rate reported by the National Center for Education Statistics is likely inflated by at least ten points, according to Nobel laureate James Heckman and Paul LaFontaine of the American Bar Association in a report issued last month by Germany’s Institute for the Study of Labor. Another estimate, derived from the Census Bureau’s Current Population Survey, is even more inflated because it includes GED graduates and excludes prison inmates and those in the armed services.

If tracking graduates is tough, keeping tabs on school safety is even more difficult. Incident reports issued through Florida’s School Environmental Safety Incident Reporting system, for example, can be rife with errors, especially after reports were sent from individual schools to district administrators, according to Florida A&M researcher Nancy Fontaine in a 2003 study.

Standardized test score reports are especially troublesome. Just three of the 20 states that reported increases in student scores on reading tests since 2003 also showed progress on the National Assessment of Educational Progress, the gold standard in testing, according to a report by the conservative-leaning Thomas B. Fordham Foundation. Another report, by standards-and-accountability think tank Achieve Inc., shows that just five states — Louisiana, Maine, Massachusetts, South Carolina, and Wyoming — have testing regimes equal to or more stringent than NAEP’s “proficient” standard.

The rules and formulas schools use in keeping track of student performance, shaped during battles between school reformers and the educational establishment, is one reason why school data can be so faulty. Local school officials want to avoid being penalized for low graduation numbers or attendance rates, especially when the penalties include smaller budgets and public ridicule. So they create loopholes in such measures, giving their schools the appearance of high achievement.

Another culprit, surprisingly, is No Child itself, which allows states to set their own performance goals — thus allowing them to game the system to their advantage. This is why 28 states can set “any progress” in graduation rates as a goal. As a result, as Heckman and LaFontaine point out, schools have “strong incentives to raise graduation rates by any means possible,” including cheating.

For school reformers, the poor quality of school data hamstrings their efforts to get parents on board with their reforms. But they can take heart in revealing that traditional public schools are struggling as mightily in offering honest statistics as they are in educating students.

Sign Up to receive Our Latest Updates! Register