While school officials, politicians and researchers often clash over the correct policy to reduce dropout rates, the right way to measure those figures can be

just as controversial.

"Looking at dropout rates is a very tricky thing to do," said Jane Hannaway, a vice president at the American Institutes

for Research. "If a student moved from, say, D.C. to Prince George's County, they could easily be counted as a dropout."

Relocation is especially prevalent among low-income students, who are also often the most at risk of dropping out.

"Kids move around," said Education Sector senior fellow Peter Cookson Jr. "They go from school to school. Especially kids in poverty."

Some organizations use "promoting power" -- the percentage of freshman who make it to senior year -- which was developed by researchers at Johns Hopkins University. Even those numbers, however, can be affected by anomalies like a large influx of students when a nearby school closes.

"We like to think of 'promoting power' as more of a check-engine light," said Jason Amos, a spokesman for the Alliance for Excellent Education. "There's a sign that there's something wrong with the school."

Daryl V. Williams, chief of student services for

Prince George's County Public Schools, said uncertainty over student departures and other small calculations can build up to a larger effect on the way districts and schools are perceived.

"We use the state's definition for dropouts: any student who leaves school for any reason except death before graduation, even if it's not known if a student is enrolled at another school or program," Williams said. - Matt Connolly