⭐

Featured Research

Bridging the Digital Divide: How 3G Coverage Transforms Fertility Decisions in Nigeria

with Conner Mullally, Xinde James Ji and Jared Gars

Dissertation Research

AAEA 2025 ASSA 2026
📋 Abstract

Using Nigerian DHS data (2013–2018) linked to 3G rollout across 725 local government areas, we find that a one standard deviation increase in coverage reduces birth probability among women aged 12–20 by 1.4–1.8 percentage points. Effects operate through delayed cohabitation and childbearing rather than contraceptive uptake, alongside increased skilled employment and stronger household bargaining power.

Unintended Consequences of Best Intentions: Examining Spillover Effects in Targeted Supplementary Education Interventions

with Yue Ma, Conner Mullally and Scott Rozelle

Under Review

ASSA 2025 AAEA 2025 APPAM 2024
📋 Abstract

Field experiment across 130 rural Chinese boarding schools comparing computer-assisted learning and workbook interventions. Results reveal significant negative spillovers from workbook treatments on non-targeted students, particularly those with close peer connections. The mechanism appears motivational: observing peers receive extra resources reduces confidence in academic effort. Computer-assisted learning conducted outside classrooms shows no such spillovers.

📚

Publications

“Who sits next to whom? The role of physical distance in shaping academic achievement in rural China” (with Yu Bai, Yue Ma, Andrew Rule, and Scott Rozelle). China Agricultural Economic Review, 2026. [click here]
“Do Color-Coded Nutrition Facts Panels Nudge the Use of Nutrition Information on Food Packaging?” (with Xuqi Chen, Lisa House, and Zhifeng Gao). Food Policy, 2024. [click here]
“Associations between Urbanization and the Home Language Environment: Evidence from a LENA Study in Rural and Peri-urban China” (with Yue Ma, Scott Rozelle, et al.). Child Development, 2023. [click here]
“Maternal Health Behaviors during Pregnancy in Rural Northwestern China” (with Yue Ma, Sarah-Eve Dill, et al.). BMC Pregnancy and Childbirth, 2020. [click here]
“Arrival Order for Positive and Negative Effects of Parental Migration on the Academic Performance of Left-behind Children in Rural China” (with Yu Bai). Studies in Labor Economics (in Chinese), 2018. [click here]
🔬

Working Papers

Using Text Messages to Improve Parenting Knowledge and Early Childhood Development in Rural China

with Yue Ma, Xiaoyang Ye, Susanna Loeb, Alexis Medina and Scott Rozelle

Under Review

AAEA 2022 China Education Finance 2023
📋 Abstract

This paper evaluates a text messaging-based parenting program (Tips-by-Text) among 1,096 low-income mothers in 6 counties in rural northwestern China. Overall, our results show substantial, positive impacts of Tips-by-Text on parenting knowledge (ITT = 0.222 SD, p < 0.01) and some critical stimulating parenting practices. While the average treatment effects on other parenting practices and early childhood development outcomes at ages 0-3 are statistically insignificant in the sample overall, we found large heterogeneities in the treatment effects consistent with three behavioral economics concepts: lack of information, inattention, and motivated cognition.

⚡

Works in Progress

Maternal Migration and Early Child Development in Rural China

SAEA 2023 AAEA 2022 WEAI 2022
⚡

Policy and Outreach Writing

Save the Children Yunnan Ludian 0-3 Years Early Childhood Development Project (2019-2020) Evaluation Report.

with Yu Bai

⚡

Machine Learning Project

Traffic Sign Classification

with Thiago de Andrade, Rui Guo and Cody Haby

📄 Full Draft
📋 Abstract

This paper details the development of a Convolutional Neural Network (CNN), a shift invariant artificial neural network (SIANN) utilizing convolution operations instead of matrix multiplication, with the goal of classifying ten unique traffic signs. A well-balanced data set of photos with equivalent resolution was used to train and validate the neural network to determine appropriate hyperparameters for optimal performance, accurate classification greater than ninety percent. The CNN was developed using packages found within the Tensorflow library in Python, including convolution, pooling, and dense layers. Additionally, this paper documents specific experiments conducted during the design and training which led to the final architecture of the neural network. The CNN will be shown to have an accuracy of greater than ninety-four (94) percent during training and validation.

Yujuan Gao
Contact: yujuangao@vt.edu