Ever since the New York City Department of Education introduced a new high school admissions process in 2004, there has been much confusion about how it works. The New York Times recently took this confusion to a new level in an erroneous, alarmist article entitled “Couldn’t Get Into Yale? 10 New York City High Schools Are More Selective” (March 10).
Our effort to set the record straight through a prompt letter to the editor of the Times went unacknowledged, as did a follow-up inquiry two weeks later given that neither a letter from another source nor a correction had appeared in the newspaper. As school choice as well as privatization is the province of this research center, this was a story we felt obliged to address.
The Times erred on three fronts. First, high school admissions in New York City is far more a matching than selection process. As the newspaper explained, “Students can apply to up to a dozen schools anywhere in the city, and then an algorithm matches them with one.” What the newspaper failed to explain, however, is that on account of this algorithm, high schools can necessarily admit only a small fraction of students. In fact, if all students exercised their right to apply to 12 high schools, the city’s high schools, in turn, could, on average, each accept only one of 12 applicants, or a meager 8.3 percent.
That the Times discovered 10 schools accepting a lower percentage of applicants than Yale—which last year, the newspaper reported, accepted one of 16 applicants, or 6.3 percent—should have thus come as no surprise. As it happens, if one looks at the profiles of the some 400 high schools and 300 affiliated programs in The 2017 New York City High School Directory, one will find 93 more selective than Yale according to the criteria employed by the Times.
Yet the comparison is utterly misguided. Setting aside that applicants to Yale constitute a filtered cohort of highly academic students, applicants there and to colleges across the country may be accepted by many institutions. By contrast, applicants to high schools in New York City may be accepted, as noted, by only one school—unless they also simultaneously apply to a separate group of what are called Specialized High Schools—eight of which require an exam and one of which requires an audition—in which case they may be accepted by one conventional high school and, if they do well enough on the exam or in the audition, one Specialized High School. (If a student chooses to audition and to take the exam and succeeds in both undertakings, he or she would have three options.)
Such selectivity is thus not a manifestation of selectivity per se but of a matching process designed, in the words of the three economists who developed it, to “relieve the congestion of the previous offer/acceptance/wait-list process” that conferred “some students multiple offers” and “multiple students ... no offers” (Atila Abdulkadiroglu, Parag A. Pathak, and Alvin E. Roth, “The New York City High School Match,” American Economic Review, July 2005).
Second, on a more rudimentary level, the Times failed to note that students listed as applicants to particular schools do not all follow through in applying. The denominator in the equation can thus be much smaller than it appears. And third, the newspaper mistakenly reported seats, rather than offers of admission, as the numerator in the equation. Because students may also simultaneously apply to the Specialized High Schools, many students partaking in the matching process for conventional high schools decline their spots and opt instead for a seat at one of the Specialized High Schools. In fact, year after year, nearly 6,000 of the city’s 77,000 or so eighth-graders do precisely that. In addition, a portion of applicants may decline their spots and opt instead for a seat at a private school.
As a case in point, Beacon High School, a highly regarded progressive school on Manhattan’s West Side and one of the 10 schools listed by the Times as more selective than Yale, had 6,202 eighth-graders this year list it as one of up to 12 schools in which they were interested, 3,300 followed through in submitting a portfolio of their best work and sitting for an interview, and 562 got offers for 350 seats. The Times reported only 6,202 as the number of applicants and 350 as the number of seats to determine the “admissions rate.” The number of true applicants and offers come from Beacon's coordinator of admissions; the New York City Department of Education does not publish such data. With 350 seats, Beacon's coordinator of admissions was told by officials in the DOE’s Office of Enrollment to make 562 offers, as past experience indicated about 212 students would opt instead for a Specialized High School (or a private school), and to “rank” another 1,200 students as eligible should the Office of Enrollment need to go deeper than 562 to fill Beacon’s 350 seats. The “admissions rate” was thus not, as reported by the Times, 5.6 percent (350 over 6,202) but 17 percent (562 over 3,300), meaning one of six applicants got offers rather than one of 18. Yet even this elevated rate of acceptance must be understood as limited by the matching process.
The implications of such misinformation are significant. It causes parents and students alike undue stress. It stands to discourage guidance counselors from urging students to apply to schools appearing so selective, lest they fuel unrealistic expectations. And it leads commentators to draw the wrong conclusions about the city’s schools. The Times, in this light, closed its account with a quote from the editor of InsideSchools: “The problem is there are more good kids than there are good schools.” Yet 72 percent of eighth-graders this year and 75 percent last year, reported the DOE, got a spot at one of their top three choices (NYC DOE, Press Release, March 8, 2017). The algorithm may be confusing but it works fairly well and far better than the process it replaced.
While the DOE has done a herculean job in managing this complex process—and, in conjunction, publishing its massive annual directory with detailed descriptions of each of the city’s high schools—the DOE has erred in counting as applicants to each school in this directory those students who list the school as one of up to 12 in which they might be interested. With this information, the DOE has moreover erred in publishing “applicants per seat” for each school. This presentation of the data in the directory underlies the misrepresentation of the data in the Times.
A random selection of four consecutive schools in the directory makes this problem patent: 10 applicants per seat at the Business of Sport School; 48 applicants per seat at Central Park East High School; 25 applicants per seat at Chelsea Career and Technical Education High School; and 31 applicants per seat at City College Academy of the Arts. In other words, three of four schools randomly viewed in the directory would, according to the criteria used by the Times, qualify as more selective than Yale.
In the case of the Specialized High Schools, the numbers can be extreme. Four of the eight schools that base admissions on results on the Specialized High Schools Admissions Test (SHSAT) are quite small, with between 90 and 150 seats in ninth grade. Three—Bronx Science, Stuyvesant, and Brooklyn Tech—are quite large, with between 750 and 1,450 seats in ninth grade. As students taking the SHSAT may rank all eight schools by order of preference and as approximately 27,000 students take this exam each year, the “applicants per seat” information in the DOE directory for the small schools is meaningless. According to the directory, there were 113 applicants per seat last year at Brooklyn Latin; 176 per seat at the High School for Mathematics, Science, and Engineering at City College; 190 per seat at the High School of American Studies at Lehman College; and 160 per seat at Queens High School for the Sciences at York College. Had the Times reported the admissions rates of these schools, the percentages would have been 0.8, 0.6, 0.5, and 0.6, respectively.
One strategy for remedying all this confusion would merely involve four modifications to the directory. First, while the DOE does a superb job of explaining the admissions process in the introduction to this directory, it should add a page or two to document the numerical implications of the matching process.
Second, our obsession with rankings must be curbed. To this end, there is no need to publish the number of applicants for all schools. Most of the city’s high schools are, in fact, not selective. Publishing the number of seats at these schools should suffice. Of the eight categories of high schools, only three may be described as selective: Test, Audition, and Screened. For the next three categories, the number of applicants serves little purpose. Schools in the Educational Option category admit a cross-section of students defined by results on the seventh-grade English Language Arts exam (16 percent from the top quartile, 16 percent from the bottom quartile, and 68 percent from the middle two), with 50 percent of offers determined by schools and 50 percent by randomization. Schools in the Screened Language category admit students according to language proficiency. Schools in the Limited Unscreened category limit admission to those students who attend an information session. For the final two categories—Zoned and Unscreened—the DOE, naturally, does not list applicants in its directory. Schools in the Zoned category admit students according to residence while schools in the Unscreened category admit students randomly.
Third, for schools in the Test, Audition, and Screened categories, the DOE should publish only the number of students who rank schools as one of their top three or four choices. This would provide a far clearer picture of a school’s desirability. Finally, for schools in these categories, the DOE should publish the number of offers of admission as well as the number of seats. The ratio of students ranking a school as one of their top three or four choices to the number of offers of admission would be much more helpful than what the DOE has been publishing: number of students listing a school as one of up to 12 in which they are interested and number of seats.
With these four modifications, students and parents as well as journalists should have a far better understanding of the admissions process.
Samuel E. Abrams, Director, NCSPE, April 9, 2017