Peer effects on student weight: randomization evidence from China

ABSTRACT

This paper investigates peer effects on student weight. Taking advantage of a national survey of middle schools in China, we exploit the BMI variation in randomly assigned classrooms. We find that having heavier peer classmates significantly increases a student’s body weight. The results are consistent across various robustness checks based on a standardized measurement of BMI, alternative peer measures, a spurious correlation with height, and the validity of the instruments. Peer effects are more pronounced among female students, nonlocal students, rural students, and students with longer exposure to classmates. The quantile analysis suggests that the peer effect is stronger at the lower end of the BMI distribution. These findings confirm the presence of social multiplier effects in middle schools, which is helpful in establishing health policies for reducing student overweight.

KEYWORDS:

• Peer effect
• social network
• student weight
• random classroom assignment

JEL CLASSIFICATION:

• I12
• J13

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

¹ During the compulsory education stages, the admission rate for top schools at the next level has been one of the most critical assessment indexes for primary and junior high schools in China (Song 2016). To increase the students’ admission rate for top schools at the next level, many schools set up the key classes (also known as experimental classes, exemplary classes, and rocket classes) to select a small number of elite students (You 2007). The selection criterion for elite students was primarily based on the total scores or subject scores in entrance examinations. The students selected into the key class had better resources than others in the non-key class, such as experienced teachers and high-quality classrooms. Thus, the key-class students were more likely to obtain excellent scores in the entrance examination for the next level of education, which promoted the admission rate for top schools. Since the establishment of key classes aggravated the resource inequality for children in compulsory education, the central government established the law in 2006 to explicitly prohibit these school behaviours.

² The survey is jointly conducted by the National Survey Research Centre at the Renmin University of China and academic institutions throughout the country.

³ In China, household registration, also called hukou, is an important institutional rule to manage population mobility and allocate social resources. Citizens with local (nonrural) hukou are entitled to benefits that are not available to citizens with nonlocal (rural) hukou, such as health care, housing benefits, and children’s education (Afridi, Li, and Ren 2015; Hu 2018).

⁴ The random classroom assignment ensures that peers are not selected based on peer BMI. This identification strategy does not impose any particular structure on student network within the classroom.

⁵ The computation code for BMI z-scores can be found on the WHO website: https://www.who.int/growthref/tools/en/.

⁶ However, the Hansen J test for obese peers fails to pass the validity test. It is possibly due to the small variation in the proportion of obese peers in our regression sample. Thus, the instruments are unable to capture sufficient variations to identify the effect.

⁷ In the CEPS, the school principal reported the daily school time schedule, including the starting time and the finishing time. Students are required to adhere strictly to this school arrangement.

⁸ Typically, a junior middle-school student is assigned to classes at the beginning of the academic year and takes the same courses throughout the year. On a normal school day, a course schedule consists of 7 regular lectures, and students remain in the same classroom.

Additional information

Funding

This research was funded by the South China Normal University (Luo: Grants No. 508/8S0212 and and Grant No. 18SK13; Pan: Grant No. 508/8S0213 and Grant No. 18SK15), the Philosophy and Social Science Foundation of Guangdong Province (Luo: Grants No. GD19YYJ11) and the Young Innovative Talent Project of Guangdong Province Education Department (Pan: Grant No. 2019WQNCX012).