Impact of Preferential Rank Incentives on University Test Scores

How to cite

Roy, Sutanuka. 2020. "Impact of Preferential Rank Incentives on University Test Scores." AEA RCT Registry. AEARCTR-0005841. https://doi.org/10.1257/rct.5841-1.1

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Sutanuka Roy*

* Primary investigator

Registry ID
AEARCTR-0005841
DOI
https://doi.org/10.1257/rct.5841-1.1
First registered
Last updated
Status
Completed
Trial dates
2016-07-13 – 2017-03-31
Countries
India (Assam and Meghalaya)
JEL codes
C93, H52, I21, M52, O15
Registry record
http://www.socialscienceregistry.org/trials/5841

Abstract

This is a large-scale cluster randomized experiment involving over 14,000 undergraduate students at universities in the State of Assam and Meghalaya, designed to test the individual impacts of providing rank based financial incentives to low income or socially disadvantaged students in future high stakes anonymized university exams.

Intervention

The financial incentives in our experiment were given out on the basis of test scores in anonymised centralized university semester exams, which are set and marked anonymously. The experiment involved randomization at the cohort level. A cohort is a group of students taking a major course, in degree year "t", in faculty "f" in college "c" affiliated to university "u". Our cluster is a cohort. The untargeted incentive group (i.e., prize incentives for all) included two financial rewards for students in a cohort. In Assam (Meghalya), the first monetary prize, of 4,000 (5,500) rupees, was awarded to the student with the highest grade in the cohort, and the second monetary award, of 3,500 (5,000) rupees, went to the student in the top 40 % with the highest grade improvement in the cohort. The second treatment arm was socially targeted incentives (i.e., prize incentives only for the lower castes), where the same two financial incentives were given only to lower-caste students within a cohort: the first to the highest scorer from a lower caste in the cohort, and the second to the student who was in the top 40 % among lower-caste students in the cohort and who showed the highest improvement in scores compared with their previous university test scores. The third treatment was income targeting (i.e., prize incentives only for the poor), where the same two financial incentives were given to students with an annual family income of 150,000 rupees or less.

Experimental design

It is a cohort level cluster design where students in a cohort are given rank based financial incentives for future anonymised university exams. The details are mentioned in the intervention section.

Design details

Randomization method
randomisation done in office by a computer
Randomization unit
cluster (cohort level)
Clusters
468 cohorts
Observations
more than 14,000 university students
Arms
We have 103 cohorts in the caste and 103 in the untargeted treatment arm, and 101 in the income treatment arm; the rest are in the control group.
Primary outcomes
student test scores in university semester exams that were incentivised in the experiment
Secondary outcomes
names of friends who studied in a group to prepare for the incentivised exams

Cited by

No citation data available for this registration yet.

Keywords