Discontinuity Assignment Identification
RCTJournal Article
In group-based educational programs, separating the direct benefit received by tutored students from the indirect peer effect on untutored classmates is difficult without explicit pre-assignment identifiers.
Picture this
Exploiting a sharp administrative cutoff rule—such as pulling only the bottom 20 students in a ranking—acts like a threshold gate, allowing researchers to isolate gains by comparing students just below the selection line to those just above it.
What the evidence says
RD estimates confirmed that direct learning gains for tutored students approached 1.00 standard deviation, while untutored peers above the rank cutoff showed no statistically significant indirect benefits.
- Who was studied
- N = 21,936 students evaluated across municipal schools in Vadodara and Mumbai, India.
- How
- Quasi-experimental Regression Discontinuity (RD) design exploiting the rule where students ranked in the bottom 20 of a class were preferentially assigned to remedial instruction.
What to do
Leverage strict class-rank cutoffs in remedial assignment to separate direct treatment impacts from indirect classroom peer effects.
From the source
"Banerjee et al. describe and implement a second strategy for separating direct and indirect effects, which exploits the discontinuity in the assignment: students ranked in the bottom twenty of their class are much more likely to be assigned to a balsakhi than those ranked above the bottom twenty."
Remedying Education: Evidence from Two Randomized Experiments in India