Association for Public Policy Analysis & Management

University-school partnerships are a popular strategy for meeting diverse student needs. A structured partnership can leverage resources from universities, schools, and the surrounding community in mutually beneficial ways – universities seek to serve their communities while providing hands-on training opportunities for their students via internships, and schools gain access to university resources while building their own capacity to leverage resources and strategies to meet students’ needs. Much literature emphasizes qualitative aspects of university-school partnerships; this study represents the first rigorous estimation of the costs, effects, and economic benefits of one such partnership using causal inference and benefit-cost analysis methods.

The program under study, Raising Educational Achievement Coalition of Harlem (REACH), is a partnership between Teachers College, Columbia University and five high-need schools in the Harlem neighborhoods of New York City. REACH entails collaboration between schools and program staff in five key areas: Leadership, Teaching and Learning, Expanded Learning Opportunities, Physical and Mental Health, and Family and Community Engagement.

As part of a larger framework of program evaluation, it is important to consider a program’s costs as well as its effects when making decisions about scaling and replication. A rigorous cost analysis can illuminate the resources used to implement its theory of action, in addition to contextualizing the size of measured effects and helping decision-makers select among alternative uses of scarce resources. We used the ingredients method for cost analysis, documenting all resources used to implement the program in order to fully capture the economic cost of the program. We obtained data on ingredients from program documentation, an implementation analysis, and interviews with program staff. In 2016-17, REACH cost $2,732,960, or $1,560 per student, with substantial variation by school site, domain of REACH, type of ingredient, and source of ingredient.

We estimate the impacts of REACH on student achievement, high school graduation, student attendance, disciplinary actions, health and fitness and social and emotional learning using a comparative interrupted time series (CITS) design with a matched comparison sample. This analysis extends the difference-in-differences design by incorporating pre-treatment trends to better control for preexisting differences between schools that could otherwise erroneously be attributed to the program. Under relatively modest conditions, CITS is not substantially biased compared to regression discontinuity results. We implement a multi-level matching framework to address the propensity of students to attend REACH schools as well as the propensity for schools to adopt REACH to construct the matched comparison sample. We estimate benefits based upon projected discounted lifetime earnings due to increased achievement and attainment.

Preliminary impact results with publicly available student-level data were positive but generally not statistically significant, except an estimated effect of 0.15 standard deviations in math. Net benefits range from $3990 to $8600 per student, with a benefit-cost ratio of 3.56 to 6.53. Over 90% of the costs are borne by university, suggesting that the program operates in part as a redistributive transfer program. The economic evaluation highlights the importance and challenges of partnerships and helps uncover what is required for a sustainable and mutually beneficial partnership.