Estimates of High School Graduation Rates Seen as Inaccurate
By Daniel de Vise
Washington Post Staff Writer
Tuesday, February 28, 2006
Between the start of freshman year and graduation, the class of 2005 at Frederick Douglass High School in Prince George's County shrank by more than one-third. Yet according to school officials, 90 percent of the class graduated.
Driven by a conviction that graduation rates are widely overstated in the region's public schools, legislators in Maryland and Virginia are seeking a more accurate count of how many students earn diplomas four years after they enter high school, one of the most vital -- but most poorly tracked -- indicators in public education. Bills before both legislatures, along with a resolution passed last fall by the D.C. school board, are part of a national effort to adopt a common formula for figuring graduation rates.
"Hopefully, with this bill, we'll begin to capture some accurate data," said Sen. Gwendolyn T. Britt (D-Prince George's), who sponsored the measure in the Maryland Senate. "And it may not be a pretty picture."
Leaders of Maryland and Virginia school systems say they are moving toward tracking individual students, using identification numbers generated by statewide testing to follow students from one school to another. Educators in both states say that by about 2009 they will know with near certainty who graduates and who does not, with or without a change in the law.
For years, Washington area school systems have estimated graduation rates based on the number of students known to have dropped out -- an equation only as accurate as each high school's ability to track dropouts.
Critics of the system say that, try as they might, most school administrators simply cannot track down every student who has left to confirm whether he or she enrolled at another school. Some administrators may not even try all that hard, given that a lower dropout rate and a higher graduation rate make a school look good.
"It's a highly manipulable statistic," said Nettie Legters, a research scientist at the Center for Social Organization of Schools at Johns Hopkins University.
Elaine Mitchell, an Upper Marlboro parent, had a son, a daughter and two nephews who were under her care attend Frederick Douglass High. She said she continued to receive bus routes and bulletins in the mail for the older nephew long after he had moved to Colorado in 2003, even after she did "everything we needed to do" to report his exit. "My nephew was probably counted in the numbers," Mitchell said, "and he wasn't there."
School system administrators defend the accuracy of their graduation and dropout data but acknowledge that it is hard work to follow every student out the door.
"Sometimes they just sort of drop off the radar," said Bruce Hislop, accountability reporting officer for Prince George's schools.
Researchers say a better way of determining graduation rates would be to consider attrition, which calculates the dwindling size of a high school class over time.
A 2004 report by the Urban Institute, a nonpartisan research group, used attrition data to estimate a graduation rate of about 75 percent in Maryland and Virginia and about 65 percent in the District. The estimates were derived from enrollment data for the class of 2001.
Maryland's State Department of Education reported that 54,170 of 63,860 students in the class of 2005, or 85 percent, reached graduation. Virginia reported a graduation rate of 77 percent, or 76,866 of 100,230 students that year. D.C. schools claimed a 72 percent graduation rate in 2004, the most recent data published.
Comparing area graduation rates with those of other states is difficult, however, because many of the figures are deemed inaccurate by researchers. It is estimated that about two-thirds of high school students graduate nationwide.
Some critics cite the Urban Institute report as evidence that graduation rates reported by Prince George's schools -- 89 percent in 2002, 87 percent in 2005 -- may be as much as 20 points too high. The report estimated Anne Arundel's graduation rate at 69 percent, in contrast to the 83 percent reported last year by the school system.
In candid written testimony before the Maryland House, the Anne Arundel school system acknowledged that its graduation and dropout rates did not seem to add up. The system reported 577 dropouts in 2005, for a dropout rate of 2.3 percent. Extended across four years, that figure should complement the system's 83 percent graduation rate but does not.
Del. Ana Sol Gutierrez (D-Montgomery), who introduced the Maryland legislation, spent eight years on the Montgomery County school board. "Every time I asked for dropouts, everybody would roll their eyes," she said. "Nobody wanted to talk about it."
She said she learned that there were "about 53 different codes" for students who left the public school system. "Of those codes, about four or five would be counted as dropouts," Gutierrez said. "In other words, you had to put on a sweatshirt that said 'dropout' and then jump up and down about five or six times to be counted as a dropout."
Public school administrators say their methods for reporting dropouts and graduates, while not perfect, are fundamentally sound.
The education departments in Virginia and Maryland instruct schools to label students as dropouts if they cannot be located and to exclude those who earn General Education Development diplomas from the graduation rolls, resulting in a comparatively conservative count of total graduates.
"We're pretty confident that we're pretty accurate from year to year in Maryland," said Gary Heath, assistant state superintendent for accountability and assessment.
Critics do not entirely disagree. The Urban Institute study put Maryland, Virginia and the District far down the list of states that appear to inflate their graduation rates.
Each camp disputes the other's methods. Incomplete dropout data tend to inflate state-issued graduation rates. Most scholarly estimates of student attrition, in turn, are skewed by a statistical "bubble" of students who repeat the ninth grade, resulting in a graduation rate that may be artificially low.
"It's really difficult for students and parents and, frankly, administrators to measure success and what's working and what's not working," said Virginia Del. William H. Fralin Jr. (R-Roanoke), who introduced legislation to change how the graduation rate is calculated.
The proposed formula, endorsed by the National Governors Association, is fairly simple: divide the number of on-time graduates by the number of students who started the ninth grade four years earlier, and add or subtract students who transfer in or out.
Virginia's Education Department supports the legislation, but lawmakers have squabbled over details. In Maryland, education officials support the measure but favor reducing it to a non-binding resolution. Both bills await decisive votes.
Patricia Wright, acting Virginia schools superintendent, said her system will adopt the new math by 2009, with or without a legislative mandate.
"The bill," she said, "is for emphasis."
View all comments that have been posted about this article.