We’ve discussed the small schools of NYC before and while the results are relatively clear, the reasons why aren’t. A new study tries to find answers, but find evidence for things that don’t influence the learning. In the study there is some interesting extra info on the small schools:
New evidence indicates that much of the academic advantage attributed to school size in NYC is driven by differences in the era in which small schools were formed. “Old” small schools, which opened prior to the reforms of the last decade, were charged with helping students stay in school. “New” small schools, which opened in 2003 or after, were provided with a bundle of supports and tasked with maintaining high academic standards.1 Recent studies have found that students in new small schools perform better than those in large schools, but students in old small schools do not fare better academically than their counterparts in large schools
This new study by Schwartz et al. adds these insights:
- We find that students attending a new small school perceive their environments similarly to those attending a large school.
- In general, small size does not necessarily generate superior learning environments.
But this study has an important limitation as it is examining the perception students have. While they want to “address whether learning environments differ for small and large high schools”, they only can say something about how these learning difference are perceived and by students who can’t compare themselves. The researchers acknowledge this and try to compensate:
Our empirical approach recognizes that student reports of the learning environment at their high school may be subject to important biases as student perceptions may reflect the past schooling experiences of the student and the particular kinds of students who attend these schools. We address the first source of bias by using a kind of “value-added” model that includes the student’s perception of his or her school’s learning environment in 8th grade, in order to eliminate fixed student-level factors influencing perceptions and thereby isolate the high school-level influences. We address the second source of bias by including a set of student level control variables and using an instrumental variable strategy to instrument for the type of high school attended.
Still it’s not sure if the info about the 8th grade is so predictive as the researchers think, imho. The second bias seems better, although we still talk about perceptions.
The researchers conclude:
Our analysis shows that ignoring how students select into these schools may provide misleading evidence about the link between school size and school learning environments. OLS results indicate that in small schools – both new and old – students perceive a more nurturing, supportive, and safe environment compared to large schools. Instrumenting attendance in new and old small schools with distance measures which are, by now, well established in the education research literature (see, for example) Schwartz et al. (2013) and Barrow et al. (2015), however, reveals that only students induced to attend old small schools, and only along one dimension, perceive better learning environments than students attending large schools. Moreover, adding school characteristics to our regression models makes all differences in coefficients on school size and era statistically insignificant. These results provide little support for the hypothesis that learning environments differ.
Is this study giving us final answers? No, not at all.
Abstract of the study:
Over the past two decades, high school reform has been characterized by a belief that “smaller is better.” Much of the expected academic benefit from attending small schools has been credited to their better learning environments. There is little empirical support for this claim, however, and the existing research fails to provide causal evidence. Moreover, recent studies in New York City have shown that students attending newly created small schools do better academically relative to students attending both large and older established small schools. Are these differences in academic outcomes also mirrored by differences in learning environments? In this paper, we address this question by exploring the impact of attending large compared to small high schools on students’ learning environments, considering the differences between small high schools formed in two different eras with different missions and resources. We use a unique data set of school and student-level data from New York City public high school students entering 9th grade in 2008–09 and 2009–10 to examine students’ attitudes about school learning environments along three dimensions: interpersonal relationships, academic expectations and support, and social behavior and safety. While OLS results show that students attending small schools (new and old) perceive better learning environments, instrumenting for selection into these schools challenges those results. In general, it is not clear that small schools provide better learning environments than large schools. Our results challenge the conventional wisdom that the higher academic performance of students in small schools is driven by a better learning environment.