Nudging take-up of an educational opportunities among online adult learners

This study evaluates the effect of nudges on the take-up and completion of an training program focused on self-efficacy offered to online (virtual) adult learners in Colombia. The course is fully asynchronous, free of charge, and participants receive a certificate of completion and are entered into a lottery for cash prizes. Invitations are sent directly by the university as part of an internal outreach process and include a personalized registration link that allows tracking of registrations. The intervention consists of varying the content of this invitation across four experimental groups:

• Control (C): Students receive the standard invitation describing the course and its benefits. No additional behavioral message is included.
• Career Benefit (T1): Students receive the same invitation as the control group, augmented with a nudge emphasizing career benefits. The additional message highlights how the skills and credentials obtained from the course can be valuable for career progression and professional opportunities.
• Social Proof (T2): Students receive the same invitation as the control group, augmented with a nudge leveraging social proof. The additional message highlights that many previous participants completed the course and reported positive experiences.
• Combined (T3): Students receive the same invitation as the control group, augmented with both the social proof and career benefit nudges.

Online registration is open for one week after the invitations are sent. Registered students then have approximately seven weeks to complete the course, receiving weekly emails with links to course sessions.

2026-04-06

2026-06-30

The primary outcomes of interest are:
• Course take-up: enrollment
• Course completion: condition and unconditional on take-up

Course take-up is measured via the personalized registration link embedded in the invitation, which allows us to track individual registrations. Since all students in the experimental sample receive an invitation, take-up is observed for the entire sample with no attrition.

Course completion is tracked through the online course platform by monitoring whether students complete all 9 sessions before the deadline. Only registered students gain access to the course, so the unconditional completion outcome is mechanically zero for non-registrants. The conditional measure isolates the persistence margin but may reflect differential selection into registration across treatment arms.

We will additionally use data collected through the registration survey, administered at the time of enrollment, to answer the following questions:
• Do nudges affect course take-up or completion by changing students’ beliefs about program effectiveness or expected difficulty?
• Do the social proof and career benefit messages affect students’ self-reported confidence in overcoming challenges?
• Do the treatments affect the stated reasons for registering to the course?

Beliefs about (expected) course effectiveness and (expected) difficulty in completing the course by the deadline are elicited during the registration survey using a 0–10 Likert scale. Additionally, the registration survey measures confidence in overcoming challenges (0–10 Likert scale), the number of peers known to be registered in the course (0–100 slider), and expected GPA for the current semester (0–5 slider).

We also included an open-ended question asking respondents what drove their decision to register for the course, as well as two self-reported goals for the semester. We will assess treatment effects on the likelihood of providing an answer and utilize text analysis to analyze the content of the responses.

This study is implemented as a randomized controlled experiment (RCT). The target population consists of all active students enrolled in online (virtual) programs at a university in Colombia. Based on administrative records, this population is predominantly adult (median age 30), with a diverse age distribution, and is characterized by balancing work, family, and study obligations.

Students are randomly assigned at the individual level to one of four groups:

• Control (C): Receives the standard course invitation without any behavioral nudge.
• Career Benefit (T1): Receives the same invitation as the control group plus a nudge highlighting career benefits of the course.
• Social Proof (T2): Receives the same invitation as the control group plus a social proof nudge describing positive experiences of past participants.
• Combined (T3): Receives the same invitation as the control group plus both the social proof and career benefit nudges.

Invitations are sent directly by the university as part of an internal outreach process and include a personalized registration link for tracking. For the analysis, we will estimate OLS regressions comparing each treatment group to the control and will present results from one-sided tests to improve statistical power, given the directional nature of our hypotheses.

Evaluate the pooled nudges compared to the control, combining the three treatments into a single group, to evaluate the general effect of a nudge on promoting take-up and completion.

Our hypotheses concern course take-up and unconditional completion. We expect that nudges can increase these outcomes relative to the control group. We do not expect meaningful differences in conditional completion (among those who register) across treatment arms, as the nudges target the decision to register rather than persistence within the course.
• Nudging increases course take-up and/or unconditional completion relative to the control group (Any treatment as well as all treatments pooled vs. C).
• The combined nudge produces larger effects than each individual nudge (T3 vs. T1 and T3 vs. T2).

We will conduct exploratory analyses to examine heterogeneous treatment effects using available baseline characteristics (e.g., gender, GPA, among others). We will also examine the mechanisms underlying any detected effects through the secondary outcomes measured in the registration survey.

Not available

Randomization done by statistical software. Students are randomly assigned to one of four groups (Control, Career Benefit, Social Proof, Combined) through a stratified randomization procedure at the individual level. Stratification is based on the following binary variables: gender, year of study at the university, whether the student’s cumulative GPA is above or below the population median, and whether the student’s age is above or below the population median. This stratification ensures balance across treatment arms on these key observable characteristics.

Individual.

No

N/A

N = 1626 (number of students invited to register for the course)

N_C = 403 students are assigned to the control group.
N_T1 = 406 students are assigned to the social proof treatment.
N_T2 = 408 students are assigned to the career benefit treatment.
N_T3 = 409 students are assigned to the combined treatment.

Using the control-group course take-up rate from a previous experiment with the same training program (p₀ = 0.26), α = 0.05, power = 0.80, and one-sided tests: • For each pairwise comparison (one treatment arm vs. control, approximately 406 vs. 403 students), the minimum detectable effect is a 7.7 percentage-point increase in take-up. • For the pooled comparison (all treatment arms combined vs. control, 1,223 vs. 403 students), the minimum detectable effect is a 6.3 percentage-point increase in take-up.