Households are approached and randomly offered tuition services for one of 4 monthly prices: Rs. 250/200 (the posted price) ; Rs 150 ; Rs. 75 ; Rs. 0. Households receive a voucher corresponding to their price, effective between April 2014 and March 2015. Conditional on attendance in August 2014, a second price offer (equal to or lower than the first price) is made, effective untill March 2015.
The sample consists of students in grades 6-8 in the 2014-2015 school year. The two subsamples are children enrolled in Pratham tuitions in 2013-2014, and "new" students whose households are in the vicinity of each tuition center.
Tuition services are offered to each household at randomized prices. In 6th grade, the prices are randomized across Rs. 0, Rs. 75, Rs, 150, and Rs. 200 (the posted price). In 7th and 8th grades, the prices are randomized across Rs. 0, Rs. 75, Rs, 150, and Rs. 250 (the posted price). Households receive a voucher corresponding to their price, effective from the time of offer through March 2015.
Conditional on enrollment, a second price offer (equal to or lower than the first price) is made at households, effective until March 2015.
Primary Outcomes (End Points)
Children are tracked in the classes for 10 months. The key outcomes of interest are takeup (enrollment), attendance in the classes, and child test scores.
Child attendance is tracked in the classes from the time of offer through March 2015. The key outcomes of interest are takeup (enrollment), attendance in the classes, and child test scores math and English test scores. We also track attendance in the 2015-2016 school year, when all prices revert to the posted price.
Primary Outcomes (Explanation)
Attendance is measured using administrative records taken in the Pratham tuition centers. These records have been verified through unannounced checks by enumerators. We will utilize both the intensive margin (take-up) and extensive margin (amount of classes attended).
Experimental Design (Public)
This research consists of a field experiment in the market for group-based, after-school tutoring services in India.
The sample for this study consists of approximately 5400 children in 4400 households in in the neighborhoods of Pratham’s 21 tuition centers.
Households are offered tutoring services at prices ranging from 0 to 250 rupees per month. We implement a two-part pricing design similar to those in the health literature (Cohen and Dupas, 2010; Ashraf, Berry and Shapiro, 2010). Households are first offered tutoring services for their children at randomly assigned prices (the “offer price”). Then, if a household accepts that price, the household receives a randomly assigned discount of up to the amount of the offer price (resulting in a lower price paid, the “second price”). This way, the selection effect of prices can be isolated by examining the composition (and behavior) of households that purchase the services at different offer prices conditional on the second price. The psychological effect can be isolated by examining the behavior of households that have paid different prices, conditional on the offer price.
The core features of the experimental set-up are described above. In addition to administrative attendance records, data collection consists of baseline household surveys and child english and math tests conducted at the time of the first offer, and endline surveys and tests conducted upon completion of the 2014-2015 school year.
First-price randomization done on computer for existing households, and using a draw from a bag of scratch cards for new households. Second-price randomization done on computer.
The first-price randomization done on computer for existing households, and using a draw from a bag of randomly-sorted scratch cards for new households. Second-price randomization done on computer.
Planned Number of Observations
4400 students, 1760 enrollees
Sample size (or number of clusters) by treatment arms