|
" Improving College Choice for the Poorest Students Using Behavioral Policy Interventions "
Ye, Xiaoyang
Dynarski, Susan Marie
| Document Type
|
:
|
Latin Dissertation
|
| Language of Document
|
:
|
English
|
| Record Number
|
:
|
1105913
|
| Doc. No
|
:
|
TLpq2347660156
|
| Main Entry
|
:
|
Dynarski, Susan Marie
|
|
|
:
|
Ye, Xiaoyang
|
| Title & Author
|
:
|
Improving College Choice for the Poorest Students Using Behavioral Policy Interventions\ Ye, XiaoyangDynarski, Susan Marie
|
| College
|
:
|
University of Michigan
|
| Date
|
:
|
2019
|
| student score
|
:
|
2019
|
| Degree
|
:
|
Ph.D.
|
| Page No
|
:
|
301
|
| Abstract
|
:
|
Inequality in college access and match is a persistent global problem. Only recently, misinformation and sub-optimal decisions have been recognized as major behavioral barriers for low-income students during the complicated high school to college transition. However, little is known about what works in centralized college admissions systems or developing countries. To fill this gap, from 2016 to 2018, I designed and conducted the Bright Future of China Project, a set of large-scale randomized experiments, in one of the poorest provinces of China (Ningxia). This three-chapter dissertation contributes to the literature and worldwide policy efforts by providing new evidence on using behavioral policy interventions to improve college access, choice, and match at scale for low-income students. The first chapter studies college choice behaviors and admissions outcomes in a typical centralized system in China. I document that the student-college academic undermatch is prevalent in centralized college admissions, even though the application process is simplified, and college information is centrally provided. I find descriptive evidence that the undermatched college choices are likely due to the lack of accurate predictions of admissions probabilities. I present results from school-level randomized experiments in 2016, which targeted students' behavioral barriers by providing college application assistance in the forms of a guidebook, a workshop, and personalized advising. The intensive application assistance interventions, during a short period with low costs, largely improve college access and match by shaping students' college choice behaviors. The second chapter studies the scale-up problem of behavioral interventions in college choice for low-income students. I report results from a set of student-level and teacher-level randomized experiments in 2017. Treated students were provided with a personalized advising program that targeted their behavioral barriers in the process of college choice and application. To scale up the labor-incentive advising, I designed and examined two policy solutions: (1) machine learning algorithms to simplify predictions and improve decision-making, and (2) a pay-for-performance policy to incentivize teachers to act as temporary counselors. Machine learning substantially increased advising effectiveness. However, without complementary school organizational policies, teacher pay-for-performance policy did not incentivize teachers to provide sufficient application assistance to students. The third chapter, co-authored with Ao Wang and Shaoda Wang, examines how motivated cognition prevents students from learning crucial objective information, using a field experiment in 2018. We focus on a unique empirical setting: Chinese Muslim students were about to take the high-stakes College Entrance Examination during the month of Ramadan, due to the occasional overlap of these two events. We invited well-respected Chinese Muslim leaders to grant explicit exemptions to delay the Ramadan fast until after the exam. We then randomly provided this treatment to some Muslim students, creating experimental variation in the stringency of religious practices. Using a survey design, we measured students' perceived cost in exam performance of taking the exam during Ramadan. Comparing with the estimates that we obtained from a novel administrative dataset and an event study design, we find that the control group students who did not receive the exemptions exhibited strong patterns of motivated cognition bias. They underestimated the cost of fasting on exam performance. However, the exemptions alleviated such bias in learning and made the treated students more willing to delay the fast for the exam. The results provide compelling evidence on how people distort valuable objective information and under-appreciate the cost of religious activities.
|
| Subject
|
:
|
Asian studies
|
|
|
:
|
Education policy
|
|
|
:
|
Labor economics
|
| |