Programs that Work
New York City's Small Schools of Choice
Children Succeeding in School
Students performing at grade level or meeting state curriculum standards
Students graduating from high school
Age of Child
Type of Setting
Type of Service
Type of Outcome Addressed
Cognitive Development / School Performance
Evidence Level (What does this mean?)
"Small Schools of Choice" (SSCs) are public high schools in New York City. They do not select students based on academic criteria and are small in size, relative to other high schools. Students are admitted based on student choice on the High School Application and seat availability, with priority given to students who attend a school information session, open house, or a high school fair. Some schools give priority to students or residents of a specific borough. Each school has about 400 students total, with 100 students per grade in grades 9-12.
The SSCs initially replaced failing neighborhood high schools around New York City through a competitive process in which prospective school leadership teams submitted proposals. SSCs began to open in 2002, starting with the 9th grade, and added an additional grade level each year until they served 9th through 12th grade. New SSCs are opened on a yearly basis.
The SSCs received resources that included start-up grants or partnerships with organizations that had started new schools previously. Principals were selected through a competitive application process and then hired a mix of teachers from the old school and ones who were new to the building. Ultimately, the SSC schools brought small, academically rigorous schools to areas where previous schools had struggled.
The SSCs serve high school students, especially disadvantaged students who live in areas previously served by failing high schools that were closed. While there were 123 SSCs in New York City at the time of the study, only the 105 schools that were oversubscribed and required a lottery for admission were included in the evaluation.
Bloom, Thompson, and Unterman (2010) evaluated the impact of SSCs on various academic outcomes for the period 2002-2008. Bloom and Unterman (2012) extended the analysis from the 2010 evaluation an additional two years for two of the cohorts of students.
Both evaluations employed a natural experiment design that relied on the city's High School Admissions algorithm, which functions like a lottery, assigning "winning" students to SSCs according to a stepped process. The algorithm first tries to match the student with his or her top school choice. If the top-choice school is full, then the next step is to scan each applying student's school priority status (which is based on whether the student resides in the same borough as the school and whether the student attended a school information session or was otherwise "known" to the school). If there are students who have lower school priority, they get "bumped" in favor of student with the higher priority. If all students have the same school priority, the student is then placed into their next choice school using the same method. Students are placed in the comparison group when they "lose" the lottery and are not placed in an SSC through the lottery. Lottery winners placed in an SSC of their choice thus become a part of the treatment group.
The study sample included only students who chose SSCs that were oversubscribed and, through the High School Admissions Process lottery, were randomly assigned either to the SSC or another school on their list. About 10 percent of students who applied through the lottery were not ultimately assigned to an SSC and then applied through other means; such students were not included in the study sample. The SSCs included in the evaluation are only those for which applicants exceeded admissions. These selection criteria yielded 21,085 students across four cohorts to be included in the evaluation. The treatment group consists of students who were offered a seat in an SSC, and the comparison group consists of students who applied through the same lottery and were not offered a seat in that particular SSC. The school assignment process yielded treatment and comparison groups that were comparable on baseline characteristics such as test scores and student demographics.
The evaluation found that the students in the study sample (both treatment and comparison groups) were the same as the broader New York City student population, with a few exceptions. The study sample was more likely to live in the Bronx than all students enrolled in SSCs and all 9th-grade students in New York City public schools. They were also less likely to be special education students (as a result of the fact that researchers could not collect data on non-mainstreamed special education students).
Outcome data were tracked for four cohorts beginning in 9th grade, starting with the 2005-2006 school year. The 2010 study tracked outcomes for four years and included the number of credits earned, the core subjects that were failed, on-time graduation from high school, and type of diploma received. The 2012 study tracked outcomes for six years of two of the cohorts and included on-time graduation from high school, type of diploma received, five-year graduation rate, and college readiness. College readiness was measured by New York State Regents Exams scores in math and English.
Key Evaluation Findings
Bloom, Thomson, and Unterman (2010) found the following:
- Approximately 58 percent of the treatment group (lottery winners enrolled in an SSC) versus 51 percent of the comparison group (lottery losers who did not receive an offer to an SSC) were on track to complete high school after their first year. This represents a statistically significant difference.
- The treatment group earned significantly more credits than the comparison group (1.4 and 1.3 more) in the third and fourth years of high school, and they had a significantly higher graduation rate (about 5 percentage points higher) four years after their scheduled start of 9th grade.
- At the end of each year of high school, the treatment group earned significantly more credits than comparison students, ranging from 0.9 to 2.6 more credits, depending on the cohort.
- At the end of the first year in high school, treatment students were significantly less likely (7.8 percent less likely) to have failed more than one semester of a core subject.
- Treatment students that enrolled in an SSC were significantly more likely (6.8 percent more likely) to have graduated from high school within four years.
- Students who graduate get one of three diplomas, from most basic to most stringent requirements: local, Regents, Advanced. There were no statistically significant differences in the type of diploma students received.
Bloom and Unterman (2012) found the following:
- Treatment students were significantly more likely (8.6 percent more likely) to have graduated from high school within four years.
- The difference in graduation rates was mainly driven by receipt of Regents diplomas, for which treatment students were significantly more likely (6.5 percent more likely) to receive. There were no statistically significant differences in receiving local or advanced diplomas.
- Treatment students were significantly more likely (7.6 percent more likely) to be college-ready in English (i.e., get a 75 or higher on the English Regents Exam).
- There was no statistically significant difference for college readiness in math.
- Subgroup analyses were conducted on levels of reading and math proficiency, low-income status, race/ethnicity and gender, and whether students had contact with the SSC before enrolling. All analyses indicated that treatment students had significantly higher graduation rates than comparison students, with the exception of Hispanic males and the race/ethnicity category of "other" for both males and females. The estimated effects for each subgroup ranged from 6.9 to 11.4 percent.
- There was no statistically significant difference in the five-year graduation rate (i.e., the difference between the percentage of students graduating after four and five years).
School districts or other school operators
The schools received start-up funding from philanthropic foundations such as the Bill and Melinda Gates Foundations, which provided grants for four- and five-year startup time periods, totaling around $400,000, in addition to support from the New York City Department of Education.
While the specific features of the small schools vary, the common design principles for each school that were emphasized in the proposal process include academic rigor, personalization, and community partnerships. There are also smaller educational units within each school (e.g., student advisory, small learning communities) that help teachers and other adults get to know students better.
The schools received technical assistance from the New York City Department of Education and teachers' and principals' unions. Additionally, schools received certain policy protections during their start-up period to allow them to open with only one grade level, access supports to select and hire staff, and have greater control over budgets and programming.
SSCs are staffed with high school teachers and administrators.
SSCs do not use a specific curriculum.
Issues to Consider
SSCs received a "promising" rating. Students enrolling in SSCs are 10 percent more likely to graduate from high school on time, and this effect size meets the PPN requirements for a "promising" rating.
It should also be noted that about three-quarters of the students were assigned by a random process and one-quarter by a process that might not have been random. The potential for non-randomness stems from students who apply to multiple school lotteries, where the probability of winning an earlier school lottery depended on the lotteries to which the student applied.
As is common in studies of experiments, the study authors included an analysis of those students offered a seat in a small school (the group which the program intended to treat) and, among those offered, students who enrolled in a small school (the group that actually received the treatment, which is the subset of students that chose to take up the offer and therefore has greater degree of self-selection). The reported effects above are for those who enrolled. The effects on those who were offered a seat are approximately 20-30 percent smaller than the effects on those who actually received a seat.
It is worth nothing that there is variation in the design of the schools. Additionally, the comparison group attended a variety of schools, including regular schools, other SSCs, and other innovative schools in New York.
The schools described in this report were created under an early iteration of the New York City Department of Education's New School Development process. To learn more about what the New York City Department of Education's current process looks like, visit http://schools.nyc.gov/community/newschools/default.htm.
There are 123 Small Schools of Choice in New York City.
New York City Department of Education High School Enrollment Office
New York City Department of Education website:
Bloom, Howard S., and Rebecca Unterman,
Sustained Positive Effects on Graduation Rates Produced by New York City's Small Public High Schools of Choice,
New York City: MDRC, 2012.
Bloom, Howard S., Saskia Levy Thompson, and Rebecca Unterman, Transforming the High School Experience: How New York City's Small Schools Are Boosting Student Achievement and Graduation Rates, New York City: MDRC, June 2010.