Understanding the role of schools in the Asian-white gap: a seasonal comparison approach

Abstract

Asian Americans are the most highly educated racial group in the United States and are commonly heralded as the model minority for their high academic success. Nevertheless, previous research suggests that Asian Americans may face certain disadvantages in school settings. For example, Asian Americans’ academic advantage over non-Hispanic white students diminishes between kindergarten entry and the next several years of schooling. This study provides a closer examination of the educational progress of Asian American students compared to white students through a seasonal comparison approach. Using the Northwest Evaluation Association, we analyze reading and math scores for over 130,000 Asian American and white students in grades K-7 in approximately 675 public schools across the US. We find that Asian Americans have higher academic achievement than white students in general, but that these advantages are maintained primarily through faster rates of learning during the summer months. When school is in session, the Asian advantage either remains unchanged or shrinks, consistent with the view that some school processes undermine the educational progress of Asian American students relative to white students.

Keywords:

• Asian-white gap
• racial achievement gap
• Asian American
• seasonal comparison
• school environment
• non-school environment

Notes

1. Sample size restrictions (small amount of Asian-American students) in 8th grade prohibit analysis for this grade level.

2. The data-set in use was compiled by the research team at NWEA to provide a large, diverse sample of students with complete test score data for the three consecutive testing occasions mentioned above. We do report two schools with missing school-level data. These schools are removed from final analysis using listwise deletion.

3. For instance, it is not the case that MAP assessments are used as diagnostics for either high or low achieving students only.

4. NCES is a federal agency under the US Department of Education and is the main agency responsible for the collection and analysis of education related data in the US.

5. The current sample includes a large amount of data from 19 different US states (a smaller amount of student data is available from a few others). From these primary 19 states, an average of 14,000 students are assessed. The states include: Arkansas, California, Colorado, Iowa, Idaho, Illinois, Indiana, Kentucky, Maine, Michigan, Minnesota, Montana, Nebraska, New Hampshire, Pennsylvania, South Carolina, Utah, Washington, and Wisconsin.

6. During the 2011–2012 school year, private school enrollment represents approximately 10% of the student population. Our analyses, therefore, focus on the 90% of students enrolled in public school systems.

7. Title 1 is a federal program that provides additional funding to low-income schools – schools in which at least 35% of their student population is low-income or at least 35% of the children in the neighboring school attendance area is low-income.

8. Estimates using percent of students on free or reduced lunch prices as the school level SES variable produce similar results to those reported here.

9. Contrary to the restricted-use data used in seasonal analyses of the ECLS-K data, we are unable to utilize exact start and end dates for the school year as NWEA does not provide this information. To clarify, we do have exact dates for assessment occasions, but not for the start and end of the school year. To address this limitation we impose a reasonable estimate (based on existing school-year calendars) of August 21st as the school year start date and June 1st as the end date. This produces a summer period of 81 days or 2.7 months, consistent with previous seasonal studies (Downey, von Hippel, and Broh 2004).

10. While we control for and report coefficients for gender in this analysis, further research is needed to better understand the relationship between student gender and seasonal learning patterns. We report some interesting results here suggesting that female students learn at a significantly slower pace during the school and a faster pace during the summer months. These effects, however, are not consistent at every grade level and vary by subject matter.

11. These results differ slightly from those obtained simply by the difference of spring (Y₂) and fall (Y₁) tests, as they rely on modeling predictions and incorporate extrapolations to the very beginning and end of the school year.

12. This study specifically focuses on the role of US schools on the Asian-white gap and explored whether their framing as the model minority masks potentially deleterious structural processes. Nevertheless, while Asian Americans are heralded as the model minority in the US, different countries tout different groups. Although beyond the scope of this study, future research should explore if similar processes occur with other model minority labeled groups.