Competition Re-entry and Leadership Investment after Early Setbacks: A Field Experiment in Chinese Classroom Elections

Eligible students will receive one of three brief, standardized video-based messages shortly after an unsuccessful class-officer election. The control group will receive neutral educational information unrelated to leadership, competition, or setbacks. The action-oriented treatment will provide practical guidance on translating leadership goals into relevant skills and concrete actions. The identity-protective treatment will emphasize that one election outcome does not define a student’s stable ability or personal worth and will provide a constructive way to interpret and recover from the setback. The three messages will use a similar format and duration. Successful first-choice candidates will receive the same neutral message and complete the same immediate survey. They will serve as a nonrandomized benchmark and will not constitute a fourth randomized treatment arm.

2026-09-01

2026-09-11

The primary outcome family is voluntary re-entry into subsequent leadership or competitive opportunities. It includes: (i) whether the student voluntarily registers for the subsequent school speaking activity, with teachers only collecting and submitting student registrations at the classroom level; and (ii) whether the student becomes a candidate in the next scheduled class-officer election.
Pre-specified secondary analyses will examine treatment-effect heterogeneity by gender and by prior schooling pathway, particularly whether the student previously attended a village teaching site before transferring to the current school.

Before the class-officer election, all students in participating grades will complete a baseline survey. The baseline data will be used to characterize selection into candidacy and election success and to measure pre-treatment characteristics. Following the election, all students who voluntarily ran for their first-choice class-officer position but were not elected to that position will be eligible for the randomized experiment, subject to the approved consent and assent procedures. Students remain eligible if the teacher assigns them to a different role before the intervention. Such alternative appointments will be recorded before randomization. Unsuccessful candidates will be randomized within classrooms in approximately equal proportions to the neutral control, action-oriented treatment, or identity-protective treatment. Successful first-choice candidates will not be randomized. They will receive the neutral message and complete the same immediate questionnaire as a nonrandomized benchmark. The main causal analysis will estimate intention-to-treat effects among randomized unsuccessful candidates. Analyses of initial candidacy, election success, and comparisons with successful candidates will be reported separately as nonexperimental or benchmark analyses.

Not available

Treatment assignment will be conducted through a physical lottery separately within each classroom. After the eligible students in a classroom have been identified, the research team will prepare an opaque container containing externally identical balls with concealed labels A, B, or C. The mapping between the labels and the three treatment arms will be fixed before the lottery. The composition of the balls will be chosen so that the numbers assigned to the three treatment arms within each classroom differ by no more than one. If the number of eligible students is not divisible by three, a separate random lottery will determine which treatment arm or arms receive the additional allocation slots. Students will then draw one ball without replacement, and the concealed label will determine their treatment assignment.

Individual student, blocked by classroom.

No

Across 135 classrooms, we expect 10-15 voluntary candidates per classroom, yielding approximately 1,350-2,025 candidates in total; the exact number cannot be determined ex ante because candidacy is voluntary and uncapped.

Same as above

All candidates will be included: successful candidates will receive the neutral message and serve as a nonrandomized benchmark, while unsuccessful candidates will be randomized approximately equally across the neutral control arm and two treatment arms. Assuming about five successful candidates per classroom, we expect approximately 675-1,350 unsuccessful candidates in total, or about 225-450 students in each randomized arm.