New Tool Maps School Attendance Zones Across U.S.
Resource may help smooth highly contentious process
Forget the common core or teacher salaries. Attendance zones may be the education topic most guaranteed to launch a debate in any school district in the country.
More than 4 out of 5 U.S. students attend the schools to which they were assigned by neighborhood. Many districts make at least some attendance-zone changes every few years.
“Every time you make a boundary change, someone is going to get angry,” said Tom C. Marshall, a Leesburg, Va., real estate agent. He served on the school board for the now-70,000-student Loudoun County district from 2007-11, a time when it gained about 10,000 students in four years. “A lot of people don’t understand the reasons a boundary is drawn. ... They think somebody just makes a decision behind closed doors.”
Understanding who goes to which area school and why may soon become a lot easier for education officials and community members alike. The U.S. Department of Education plans to release the first nationwide map of school attendance boundaries this summer. Starting in November, school districts will be able to use an online tool to draw or upload their own maps and download or tweak existing maps.
In the process, districts will create the most detailed picture yet of how American schools define their communities. Education officials will also have new tools to plan new schools and transportation routes, and to identify equity problems across the district. In Georgia, for example, state education officials are considering using the data to track where foster children attend school. This could help the state provide more support for such students.
“I think there are very practical reasons schools want to be able to show where their boundaries are,” like qualifying for federal or state programs or relief after a disaster, said Meredith P. Richards, an assistant professor in education policy at Southern Methodist University in Dallas.
“It is very difficult to create a [geographic-information system] map from scratch,” said Ms. Richards, who studies geographic issues in education. “Having a tool that will allow them to do this is fantastic.”
The new data can also help researchers and school leaders pinpoint equity problems in districts. Gerrymandering—the tactic by which state legislatures configure electoral maps to create safe seats for a particular party—can be common in schools, too.
School attendance boundaries often do not include students from the closest neighborhoods, but increasingly complex borders can leave out groups of students and make individual schools less diverse. One recent study compares the actual attendance areas (left) of Loudon County, Va., a rapidly growing outer suburb of Washington, with a depiction of how those boundaries would look without gerrymandering.
“People are talking so much about school choice, but the most common form of school choice is still residential choice,” said Ms. Richards. “Everybody’s concerned about political gerrymandering, but nobody really blinks an eye at gerrymandered school districts.”
In a study in the December issue of the American Educational Research Journal, Ms. Richards and her colleagues looked at gerrymandering in the School Attendance Boundary Information System, a predecessor of the NCES project. The researchers examined the attendance ares of about 15,000 mostly urban schools.
When school enrollment areas are irregular or fractured, students may live far from the school they attend and have to pass other schools to reach their own. Ms. Richards was not able to track changes to individual schools’ catchment areas over time. But she did find that the longer a school’s boundaries had been in place, the more efficient and less gerrymandered it was. This suggests gerrymandering became worse with more frequent boundary changes—as when a district is in the middle of rapid student growth.
“Troublingly, as with electoral gerrymandering, ... [educational] gerrymandering may serve primarily as a means of excluding nonwhite and poor students from whiter and more affluent schools,” she and her co-authors conclude in a related study published this month in the Teachers College Record.
Districts with neighborhoods highly segregated by race had less-gerrymandered school boundaries than districts with more integrated neighborhoods. Moreover, school districts that saw a rapid rise in the percentage of nonwhite students between 2000 and 2010 were more likely to be gerrymandered than districts without such a change. Districts under court desegregation orders had less-gerrymandered attendance areas.
The school boundaries didn’t “have a massive effect on school segregation, but it’s just one more thing having a little effect,” Ms. Richards said.
For example, Loudoun County’s rapid population growth began when the district still operated under a court desegregation order in place since the 1960s, but that order was lifted in 2006. By 2009-10, the school year that Ms. Richards focused on, Loudoun’s actual attendance zones increased separation of white and minority students in 1st grade, compared to models of enrollment without gerrymandering.
Mr. Marshall said attendance planning is almost always a contentious process. For example, groups of parents in matching protest shirts mobbed school board meetings on the Loudoun district’s most recent boundary changes.
“If you always make the boundaries based on what’s closest to schools, it’s really kind of artificial,” Marshall said. “We build schools where we can get the land, so it’s not always in the middle of communities.”
He agreed that balancing demographics in schools can lead to unusual school attendance zones, but said school board members work to balance students’ travel times with preventing overcrowding or demographic disparities in particular schools. “The school system has nothing to do with planning zoning for buildings, so we have to simply respond to the growth we get,” he said.
While parents may pay more attention to their school boundaries than even their congressional districts, the geography of education is a fairly recent addition to federal research. The U.S. Census Bureau has collected data on school district boundaries off and on since 1990, but school district information became standard in the American Community Survey only in 2005, and not until 2010 for the full Census. And those projects did not include enrollment areas for individual schools.
Researchers from the NCES and Sanametrix Inc., a Washington-based analytic firm, previewed the new mapping project at a forum in Washington this month. They outlined 80,000 schools’ enrollment areas in more than 12,800 districts, using a combination of NCES, Census, and district information. To compile that last pool of data, researchers sorted through, in many cases, hand-drawn maps from local officials.
The technical difficulty of updating boundaries, coupled with rapid changes in growing or shrinking districts meant some public information was years out of date. “I can’t tell you how many times I’ve been to a district website with an [attendance] map accompanied by text saying the map is wrong,” said Andrea Conver, a geographic information system manager at Sanametrix who works on the project.
“[Washington] D.C. was messy; New York City was a nightmare,” Ms. Conver said. “In some cases, there were 10 schools on top of each other because that’s just the reality there.”
With the 2013-14 maps, NCES has updated attendance zones for more than 90 percent of U.S. schools. There are some holes in the maps, though, according to NCES statistician Tai Phan. Many magnet and charter schools that have districtwide or criteria-based enrollment aren’t included in the 2013-14 survey, and changing enrollment practices is “an issue that will keep growing and growing,” Mr. Phan said. The researchers plan to include these schools in later surveys.
NCES plans to update the maps every other year, beginning this fall. Census and NCES researchers plan to study relationships among school and district boundaries and other boundaries, like congressional districts, county lines, and Census tracts.
Vol. 34, Issue 37, Pages 16-17