Despite billions of dollars spent yearly to fund higher education for low-income youth, no government agency tracks how many low-income young people attend college by state. Whereas proxy measures like Pell grant receipt address the number of already enrolled low-income students, direct estimates from U.S. Census surveys likely overestimate low-income youth enrollment due to their design. Using Bayesian multilevel regression with poststratification (MRP) to estimate postsecondary attendance rates by family income in each of the 50 states and the District of Columbia, we find substantial variation in attendance rates between income groups across the country.

中文翻译：

---

使用后分层多级回归根据美国家庭收入预测中学后入学率

尽管每年花费数十亿美元资助低收入青年的高等教育，但没有政府机构追踪各州有多少低收入年轻人上大学。虽然佩尔助学金收据等替代措施解决了已经入学的低收入学生的数量，但美国人口普查调查的直接估计可能由于其设计而高估了低收入青年的入学人数。使用贝叶斯多级回归和后期分层 (MRP) 来估计 50 个州和哥伦比亚特区中每个州和哥伦比亚特区按家庭收入划分的中学后入学率，我们发现全国各收入群体之间的入学率存在显着差异。