Search for: All records

Predictive analytics has been widely used in various domains, including education, to inform decision-making and improve outcomes. However, many predictive models are proprietary and inaccessible for evaluation or modification by researchers and practitioners, limiting their accountability and ethical design. Moreover, predictive models are often opaque and incomprehensible to the officials who use them, reducing their trust and utility. Furthermore, predictive models may introduce or exacerbate bias and inequity, as they have done in many sectors of society. Therefore, there is a need for transparent, interpretable, and fair predictive models that can be easily adopted and adapted by different stakeholders. In this paper, we propose a fair predictive model based on multivariate adaptive regression splines (MARS) that incorporates fairness measures in the learning process. MARS is a non-parametric regression model that performs feature selection, handles non-linear relationships, generates interpretable decision rules, and derives optimal splitting criteria on the variables. Specifically, we integrate fairness into the knot optimization algorithm and provide theoretical and empirical evidence of how it results in a fair knot placement. We apply our fairMARS model to real-world data and demonstrate its effectiveness in terms of accuracy and equity. Our paper contributes to the advancement of responsible and ethical predictive analytics for social good. 

Promise programs are proliferating across the United States, with wide variation in their design. Using national data on 33 Promise programs affecting single, 2-year colleges, this study examines program effects on first-time, full-time college enrollments of students by race/ethnicity and gender classification. Results suggest Promise programs are associated with large percent increases in enrollments of Black and Hispanic students, especially students classified as females, at eligible colleges. Promise programs with merit requirements are associated with higher enrollment of White and Asian, Native Hawaiian, or Pacific Islander female students; those with income requirements are negatively associated with enrollment of most demographic groups. More generous Promise programs are associated with greater enrollment increases among demographic groups with historically higher levels of postsecondary attainment. 

Efforts to improve college-completion rates have dominated higher education policy agendas. Performance-based funding (PBF) intends to improve college completion and links state funding for public colleges and universities to performance measures. One critique of PBF policies is that institutions might restrict student access. This study uses a difference-in-differences design and institution-level data from 2001 to 2014 to examine whether 4-year, public institutions become more selective or enroll fewer underrepresented students under PBF. Our findings, supported by various robustness checks, suggest that institutions subject to PBF enroll students with higher standardized test scores and enroll fewer first-generation students. PBF models tied to institutions’ base funding are more strongly associated with increased standardized test scores and enrollment of Pell students.