Tuition, Targeting, and Tradeoffs: A Conjoint Analysis of Americans’ Preferences over the Design of Higher Education Subsidies

ABSTRACT

Over the past three decades, political candidates and elected policymakers have advanced a wide variety of publicly funded tuition-subsidy programs to improve college access and -affordability. These college-aid programs employ different subsidy instruments, target different types of students, draw from different funding sources, and come with different strings attached. This study assesses Americans’ preferences over the multidimensional design of such higher education subsidies using a conjoint survey experiment with randomized policy proposals that vary in their subsidy instruments, target populations, eligibility requirements, and funding arrangements. The results suggest that fiscal tradeoffs loom large in Americans’ support for higher education programs, but elements of subsidy design matter as well: respondents (particularly Republicans) prefer aid packaged as tax credits; strongly means-tested programs receive greater support than broader, income-based aid (an effect driven by Democrats); and subsidies targeted to community college students enjoy substantial, bipartisan support. The study’s results shed light on differences in extant tuition-subsidy programs’ popular support and provide empirical grounding for debates over the prudent design of college aid by estimating tradeoffs that programs confront between efficiently targeting marginal students and securing broad popular support.

KEYWORDS:

• Higher education subsidies
• policy design
• public opinion
• higher education politics
• conjoint analysis

Acknowledgments

I thank Peter A. Hall, Connor Huff, Torben Iversen, Joshua Kertzer, Mack Shelley, Martin West, and anonymous JHE reviewers for their valuable comments. Ryan Enos, Mark Hill, and Amy Lakeman administered the survey and patiently fielded many logistical questions.

Disclosure statement

The author was not an employee of the College Board when this research was conducted and declares no relevant financial interests relating to it. Opinions expressed herein do not reflect College Board’s views.

This experiment is registered with the Harvard University-Area Committee on the Use of Human Subjects as protocol IRB17-1691.

Notes

1. While Pell awards are determined according to a complex eligibility formula, about 84% of students from households making under $40,000 were Pell eligible in 2011, and there is an inflection point between $40,000 and $50,000 in the distribution of dependent Pell recipients by parent income (Chetty et al., 2017: Online Appendix Figure IXb).

2. This battery of questions only appeared after the final vignette to avoid priming respondents with evaluation frames they might not otherwise have considered.