Which Friends Matter Most? Understanding Peer Influence and Its Implications for Educational Outcomes in Chinese Rural Boarding Schools

About this event

Personality and noncognitive skills are crucial components of human capital, yet the mechanisms underlying their development remain insufficiently understood. This paper investigates the treatment and spillover effects of a randomized experiment that introduced audio bedtime stories through speakers in school dormitories in Chinese rural boarding schools. Using treatment status and the distance between boarding students' schools and homes as instrumental variables, the study examines how the noncognitive skills of boarding students influence untreated day students through peer networks. By modelling peer effects within a general CES framework in the human capital production function, the analysis can evaluate whether the linear-in-means, shining light, or bad apple model best explains peer influences on noncognitive skills. The estimation of peer influence parameters in the CES model will offer practical insights regarding students' friendship choices, school selection and school development plan: whether they should focus on the performance of top, average, or lower-performing students, fostering networks for top students, or reducing the negative influence of lower-performing peers for the purpose of improving noncognitive skills.