Donor: National Science Foundation (NSF)

Date: 2019

Amount: $558589

This research will use experimental methods to study the reasons parents make school choices with very little information about options available to them. The research will study the relative importance of two mechanisms through which this occurs: (i) difficulty and cost of acquiring information about school characteristics, and/or (ii) families have incorrect beliefs about the distribution of schools, believing that all schools are the same. The research will develop a theoretical model based on parent’s beliefs about school quality and how these beliefs change as they receive new information. The research will then test this theory by providing different amounts of information to parents to see how this affects their school choice decisions. The pilot study was conducted in 2017 and the first research paper is titled Search Costs, Biased Beliefs and School Choice under Endogenous Consideration Sets, coauthored with Adam Kapor, Claudia Allende, and Patrick Agte.

Donor: National Science Foundation (NSF)

Date: 2016

Amount: $457853

The project combines a household survey measuring the preferences, sophistication, and beliefs of potential school choice participants with administrative records of school choice and, academic outcomes to estimate a latent utility model of school choice participation. We will conduct this study in the context of the public school district in New Haven, Connecticut, which uses a centralized mechanism closely related to the “immediate acceptance” or “Boston” algorithm to administer school choice.

Donor: Overdeck Princeton Education Fund

Date: 2019

Amount: $85000

This project leverages a partnership between academic researchers and policymakers in the New Haven, Connecticut school district. It aims to design, implement, and evaluate the effects of innovations to the transparency and ease of use of the school choice system. The main goal will be to evaluate what innovations to school choice models are successful and to generate evidence so these models can be replicated elsewhere.

Donor: Wellspring Philanthropic Fund

Date: 2018

Amount: $500000

Incubate the development of NGO ConsiliumBots and the research and development of AI technology to help people make better human capital investment decisions.

Donor: Microsoft Research

Date: 2017

Amount: $20000

This projects looks to take advantage of the extensive scalable computational resources provided by cloud computing to use large administrative datasets and machine learning techniques to help inform policies that will help reduce the amount of students that drop out of school in developing countries.

Donor: DIV-USAID

Date: 2015

Amount: $1498979

Large scale impact evaluation implemented in conjunction with the government of the Dominican Republic (IDEICE) that looks to increase investment in human capital by reducing information frictions among adults and children regarding the cost, benefit, and feasibility of higher education levels. A panel of in-depth household surveys are conducted on a custom-designed application measure time use, parental investments, and labor supply decisions. Administrative data is used to evaluate the impact and survey data is used through the lens of a dynamic human capital accumulation model to study the mechanisms through which information on financial aid, returns to levels of education, and to the field of study can affect investment in human capital and labor supply.

Donor: United States Department of Labor (DoL)

Date: 2015

Amount: $977690

Large scale impact evaluation to assess how the provision of this information affects drop-out rates and parents’ and students’ perceptions and decisions. A panel of in-depth household surveys are conducted on a custom-designed application measure time use, parental investments, and labor supply decisions. Administrative data is used to evaluate the impact and survey data is used through the lens of a dynamic human capital accumulation model to study the mechanisms through which information on financial aid and returns affects human capital investment decisions.

Donor: Peruvian Government and International Bank for Reconstruction and Development (BIRF)

Date: 2016

Amount: $202000

Studying the role of information asymmetries in higher education choice in the Peruvian market and how informative online platforms can help shift educational choices. Ministry of Education and Innovations for Poverty Action are leading the project.

Donor: Post Primary Initiative JPAL Global

Date: 2016

Amount: $50000

Large scale impact evaluation implemented in conjunction with the government of the Dominican Republic (Pruebas Nacionales and IDEICE) looks to inform students of high academic performance through letters that contain information and motivational messages. Its main goal is to reduce dropout rates in developing countries. Currently working partnered with the government in the Dominican Republic, Peru, and Chile. JPAL and the government of the Dominican Republic are the leaders of this project.

Donor: Skills for Youth Program, J-PAL LAC

Date: 2018

Amount: $49479

In this study, we collaborate with the government of the Dominican Republic to construct an information system and deploy a state-of-the-art application that leverages innovations in artificial intelligence and machine learning to improve the quality of the personalization of information provision. Labor market earnings information, as well as detailed financial aid information and skills profiles by major, will be provided through an application used by social workers who visit students who participate in the national conditional cash transfer program PROSOLI. Using the custom application, social workers will help the prospective student navigate the complex higher education landscape with prompts from the app, making sure to discuss how financial aid works in their specific case and how it can change depending on what they choose to study and where.

Donor: Post Primary Education Initiative, J-PAL

Date: 2018

Amount: $256893

In this project, we work together with the Ministry of Higher Education of the Dominican Republic to design and evaluate different aspects of their post-secondary financial aid policy. Guided by both theory and empirical evidence, we work with the government to design and implement several randomized control trials to evaluate different aspects of the financial aid policy, focusing on recruitment, selection, and retainment of talented but underprivileged students in post-secondary education. Estimates of these various margins will help inform the design of the final version of constrained-optimal financial aid policy.

Donor: Governance Initiative, J-PAL

Date: 2018

Amount: $48990

This is a collaborative project with the City government of Santo Domingo to design and evaluate different interventions to improve payment compliance for public garbage collection. In a first stage, we will analyze data on payment histories, as well as undertake targeted interviews and focus groups to gain a deeper understanding of the determinants of payment compliance. In the second stage, we will design and implement several randomized control trials to evaluate different behavioral and non-behavioral approaches to reduce free-riding. The interventions considered are: 1) Simplification of communication and payment instructions; 2) A lottery for households with timely payments; 3) Informing households about the consequences on their credit history of not paying the bill; 4) Informing households about debt relief options; 5) Informing citizens about actual compliance rates in high compliance neighborhoods. This will shed light on how to improve a state’s capacity to effectively collect payments for government services.