Parental Absence, After-school Tutoring, and Student Achievement: Experimental Evidence from Rural China
How to cite
Zhang, Hongliang; Zhang, Junsen; Wei, Xiangdong; Fan, C. Simon; Behrman, Jere R.. 2020. "Parental Absence, After-school Tutoring, and Student Achievement: Experimental Evidence from Rural China." AEA RCT Registry. AEARCTR-0005294. https://doi.org/10.1257/rct.5294-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
We conduct a randomized control after-school tutoring experiment in rural China where many children are left-behind by both parents and cared by grandparents. These left-behind children are disadvantaged in after-school learning support received at home and also academically lagging behind compared to their counterparts living with parents. We implement the intervention as an after-school peer tutoring program in which high-achieving senior primary school students are recruited to provide high-dosage, one-on-one tutoring to low-achieving junior primary school students. Our aims are to assess the treatment effects of access to this after-school tutoring intervention on household inputs and student achievements, as well as the potential heterogeneity in the treatment effects between children living with and without parents.
Intervention
We implemented the intervention as a peer-tutoring program in which randomly selected high-achieving senior students (4th and 5th graders who scored above their class medians in the baseline test) were paired with randomly selected low-achieving junior students (2nd and 3rd graders who scored below their class medians in the baseline test) of the same primary school to offer one-on-one tutoring after school hours. The intervention lasted 8 months, during which the tutors and tutees met for 45 minutes after school Mondays through Thursday. In each tutorial session, the tutors helped their assigned tutees to finish homework and answered any study questions.
Experimental design
The experiment involved 32 primary schools, including 24 schools with multiple classes in each grade (a total of 119 classes in grades 2 & 3) and 8 schools with a single class in each grade (a total of 16 classes in grades 2 & 3). Of the 24 schools with multiple classes in each grade, we randomly assigned half of the classes (a total of 60 classes) to experimental classes and the remaining half to control classes (a total of 59 classes).
We then pooled the 60 randomly selected experimental classes and all the 16 classes in the 8 schools with a single class in each grade together into 76 experimental classes. For each experimental class, we randomly selected 10 low-achieving students (who scored below their class medians in the baseline test) to participate in the after-school tutoring program as tutees, where the remaining unselected low-achieving students from the same classes were assigned as within-class controls. In addition, all low-achieving students from the 59 control classes acted as between-class controls for the subset of tutees from the 60 randomly selected experimental classes.
For each experimental class, we targeted a senior class in grades 4 or 5 to recruit tutors from high-achieving students who scored above their class medians in the baseline test. From each targeted senior class, we randomly selected 10 tutors from the pool of oversubscribed eligible candidates.
Design details
Cited by
No citation data available for this registration yet.