Nudging Discretionary Contributions: Evidence from a Field Experiment on Course Evaluations

How to cite

Mickeler, Maren; Kshatriya, Anil. 2026. "Nudging Discretionary Contributions: Evidence from a Field Experiment on Course Evaluations." AEA RCT Registry. AEARCTR-0017972. https://doi.org/10.1257/rct.17972-2.0

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

Maren Mickeler* , Anil Kshatriya

* Primary investigator

Registry ID
AEARCTR-0017972
DOI
https://doi.org/10.1257/rct.17972-2.0
First registered
Last updated
Status
In development
Trial dates
2026-04-16 – 2026-08-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/17972

Abstract

Many actions in organizations do not directly benefit the individuals who take them, yet they create meaningful value for others or for the organization as a whole. Such behaviors are often described as discretionary behaviors. Despite their importance for learning and improvement, it remains unclear how organizations can encourage individuals to engage in them at scale. We examine this question in the setting of student course evaluations at a major French business school. Completing an evaluation is a voluntary contribution that can support teaching and program improvements. Yet, it offers limited immediate private benefits to the individual student. As a result, participation is often lower than institutions would like, even though evaluations are an important input into improvement processes. Our field experiment is cluster randomized at the student cohort level. Cohorts are defined by program and year, hence all students within a given cohort are selected into the same condition. Our intervention is delivered through the school’s internal platform during the evaluation window. In the control condition, students receive the usual reminder to complete evaluations through the standard notification system. In the first treatment condition, students see a banner that summarizes their own evaluation participation history for the current academic year, making their individual past behavior salient and easy to track. In the second treatment condition, students see course specific information that reports the peer participation rate for each course they are taking, making participation more socially visible in aggregate. Our main outcome is whether students complete course evaluations. We test whether either prompt increases completion relative to business as usual, and whether the two prompts differ in their effects.

Intervention

The intervention consists of light touch informational nudges delivered through the school’s internal platform during the course evaluation window. The intervention does not change the evaluation questionnaire or the evaluation process. It only changes what information students see on the platform when evaluations are open.

In the control condition, students receive the standard reminder through the usual notification system, with no additional nudge on the platform beyond business as usual. In the first treatment condition, students are shown a banner on the internal platform summarizing their own evaluation participation history during the current academic year. The banner provides a simple overview of completion to date, intended to make the student’s participation history salient. In the second treatment condition, students are shown course specific information on the platform indicating the peer participation rate for each course they are taking. This information is presented in aggregate and does not identify individual students.

Experimental design

We conduct a three arm field experiment at a French Business school that is cluster randomized at the cohort level. Cohorts are defined by program and year, and all students within a cohort are assigned to the same condition. The intervention is implemented during the course evaluation window through the school’s internal platform and does not change the evaluation questionnaire or the evaluation process. In the control condition, students receive the usual evaluation reminder through the standard notification system. In Treatment 1, students see a banner on the internal platform that summarizes their own evaluation participation during the current academic year. In Treatment 2, students see course specific information on the internal platform reporting the peer participation rate for each course they take, presented in aggregate.

Design details

Randomization method
Randomization at the cohort level.
Randomization unit
We randomize at the cohort level, where cohorts are defined by program and year. All students within a given cohort are assigned to the same experimental condition. There is no individual level random assignment.
Clusters
Overall, we have three clusters in the program that participates in the experiment.
Observations
Overall, 1994 students will participate in the experiment.
Arms
Treatment 1 : 619 students ; Treatment 2 : 507 students ; Control : 868 students
Primary outcomes
Our primary outcome is the student-level completion rate, that is, the share of eligible evaluations completed by each student. For each student-course pair, the outcome equals one if the student completes the course evaluation during the relevant evaluation window, and zero otherwise.
Secondary outcomes
We further analyze the share of eligible course evaluations completed by the student during the intervention period, an indicator for whether the student completes at least one evaluation, and the total number of evaluations completed. Additionally, we will additionally examine the timing of completion, such as whether evaluations are completed shortly after the intervention becomes visible.

Cited by

No citation data available for this registration yet.

Keywords