School assignment flaws detailed (BG), September 12, 2003

The Boston Globe

School assignment flaws detailed

Two economists study problem, offer relief

By Gareth Cook, Globe Staff, 9/12/2003

Boston uses a deeply flawed system for assigning students to its public schools, pushing more students out of their top-choice schools than necessary and giving parents a reason to lie about which schools they want, according to a pair of researchers who recently published their findings in a leading economics journal.

A new system, they say, could greatly reduce the anxiety in the city’s annual school-choice process, in which thousands of parents submit lists of their top choices and await the computer-generated decision that will affect the next year to five years of their child’s education.

The researchers found that once the parents submit their lists, they are subject to a poorly designed method of allocating spots in the top schools. By using a different technique, they say, the city could get more students into one of their top-choice schools while also making the system fairer. The alternate technique, which the researchers outline in the paper, could be put in place with relatively simple, inexpensive changes and would not require the city to change any of its broader policies, according to the researchers and other academics who have seen the paper.

“Once all this is known, I don’t see how they can keep the Boston mechanism,” said Turkish economist Tayfun Sonmez, one of the researchers who studied Boston’s system.

For more than two decades, policymakers have devoted enormous amounts of attention to various ways to assign students to schools, sparking philosophical debates, charges of racial and economic discrimination, and tangled court battles — all of which have played out with particular drama in Boston. But the authors say their work, which also examined districts in Columbus, Minneapolis, and Seattle, is the first rigorous examination of how best to do the actual matching once the policy is decided.

The research has broader implications as well. If more parents were happier with their school assignments, it would help keep them from fleeing for the suburbs and bolster the fortunes of the school district — and the city. Officials with the Boston public schools and the Boston School Committee readily acknowledge that parents are frustrated with the current system, and officials said at a School Committee meeting this week that they would make changing the system a priority. They have not yet considered the method suggested by the economists.

“For every parent who feels frustrated about a policy, there is always a parent who will feel frustrated about an alternative,” said Christopher M. Horan, chief of staff for the Boston public schools. Horan said he was intrigued by the economists’ work and considered their suggestion a serious alternative.

In Boston, most students have to apply to get into kindergarten, first grade, sixth grade, and ninth grade. All students are given a priority ranking at each school, based on whether they have a sibling there, whether they live within walking distance, and a lottery number. And all parents rank the schools they would like their children to attend.

To begin, each school considers all the students who ranked it number one, and gives out seats in order of the student’s priority at that school. Then, each school that still has room considers all the students who ranked it number two, and again gives out seats in order of student priority. This process continues until all the students are assigned a spot or no more schools are left on a student’s list — in which case the student is “unassigned” and ends up at a school that still has an empty spot.

The problem, as parents quickly figure out, is that many of the best schools fill up in the first round, so students who don’t get into their first choice can find themselves crowded out of all their top choices.

“It is crazy,” said Kathy Bear, a West Roxbury mother whose 5-year-old daughter, Kaitlyn, didn’t get any of her top choices for kindergarten, and was assigned to a distant, troubled school in Dorchester.

Parents who learn how the system works begin to “game” it by lying about their top choices, ranking a less desirable school first in the hope that they’ll have a better chance of landing a spot. Now that she understands the system better, Bear says she will be more strategic in her selections for first grade.

Another problem, according to the economists, is that the Boston matching technique leads to situations where a student doesn’t get into a school even though he or she is higher on the school’s priority list than a student who did. This happens because as soon as a school fills up, it no longer considers any more applicants, no matter how high the student might be on the priority list.

A better alternative, said the researchers in the June issue of the American Economic Review, is a system similar to the national “matching” process used to assign residencies to medical students.

This system starts the same way as the method now used in Boston, with each school considering students who rank it number one and accepting them in order of their priority at the school. The critical difference, said Atila Abdulkadiroglu, a co-author of the study and an assistant professor of economics at Columbia University, is that this initial assignment is only temporary.

In the next round, students who didn’t get into their first choice are allowed to apply to their second choice. That school then considers these students, along with all the students temporarily assigned to it, and gives out all of its seats in order of the students’ priority. This process repeats, with any unassigned students applying to the next school on their list, and getting in if there is room — or bumping other students if they are higher on the priority list.

When all students have been assigned, the process ends, and the assignments become final.

This method has several advantages, according to the study. First, the best strategy for parents is always to list the schools in their true order of preference. The technique also eliminates all cases of a student losing a spot to a lower-priority classmate.

An experiment run by Sonmez, who is an associate professor of economics at Koc University in Istanbul, and University of Michigan economist Yan Chen also indicates that the method is likely to make more students happier with their assignment. In the experiment, which used 36 students vying for seven schools, most students wound up at a school that was slightly higher on their list of preferences.

Of course, no new system can create more seats at the most sought-after schools. But all parents interviewed by the Globe said that it would be a huge relief simply to write a truthful answer to the question: What school do you want?

“A lot of the alienation some parents have toward the choice system is solely attributable to the alienation of not making your first choice your first choice,” said Neil Sullivan, the father of four children who have attended Boston public schools.

Gareth Cook can be reached at

© Copyright 2003 Globe Newspaper Company.