Back to History

Fields Changed

Registration

Field Before After
Last Published April 30, 2025 09:05 AM June 08, 2025 01:15 PM
Intervention Start Date April 28, 2025 June 09, 2025
Intervention End Date July 31, 2025 March 31, 2026
Experimental Design (Public) The experiment design closely aligns with prior correspondence studies. The focus of the experiment is on the differential effects of staying unemployed to prepare for government exams relative to working in the private sector after graduation. The study will comprise three main experiments. The first experiment will highlight the differential callbacks received by candidates with a history of preparing for government exams and how the callbacks vary with each additional year spent on exam preparation. The second experiment will investigate whether the performance in government exams influences callback decisions. To this end, employers will receive resumes of candidates with varying scores in their most recent exam attempt. Finally, a third experiment will benchmark the differential callbacks received by government exam candidates against candidates with career breaks for other reasons. All other characteristics including the caste, gender, religion, school and college quality, academic grades, and extra-curricular activities will be held constant across all resume variants. The focus of the experiment is on the differential effects of staying unemployed to prepare for government exams relative to working in the private sector after graduation. The study will comprise a large-scale resume experiment. Resumes will vary in whether a candidate reveals a history of preparing for government exams. Callbacks received by candidates with experience preparing for government exams will be compared to callbacks received by candidates who have never prepared for government exams and have worked in the private sector. All other characteristics including gender, religion, school and college quality, academic grades, and extra-curricular activities will be held constant across all resume variants within a job posting.
Randomization Method Resumes will be generated using code that will randomly assign candidate characteristics (government vs private sector choice after graduation, years since graduation, government exam score) to each variant. Resumes will be generated using code that will randomly assign candidate characteristics (government vs private sector choice after graduation, years since graduation, government exam score, etc) to each variant. Resumes are generated using the program developed by Lahey and Beasley (2009).
Planned Number of Clusters 2000 jobs per experiment 500 job ads
Planned Number of Observations Experience experiment: 18000 resume-job pairs Score experiment: 12000 resume job pairs Benchmarking experiment: 8000 resume-job pairs 9 resumes sent to 500 job ads, resulting in 4500 observations.
Sample size (or number of clusters) by treatment arms Experience experiment: 6000 resume-job pairs for government candidates, 12000 resume job pairs for private sector candidates Score experiment: 2000 resume-job pairs for each score level 1500 control resumes, 3000 treatment resumes.
Back to top