Learning in College

Last registered on September 08, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

Learning in College

RCT ID

AEARCTR-0016671

Initial registration date

September 04, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

September 08, 2025, 9:16 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country

United States of America

Region

Country

United States of America

Region

Primary Investigator

Name

Michael Cuna

Affiliation

University of Chicago

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Elaine Shen

PI Affiliation

UC Berkeley

Contact Investigator

Additional Trial Information

Status

In development

Start date

2025-09-08

End date

2025-12-31

Keywords

Behavior, Electoral, Labor

Additional Keywords

JEL code(s)

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

We look into how college students search for and learn information regarding optimal strategies for academic and career success.

External Link(s)

Registration Citation

Citation

Cuna, Michael and Elaine Shen. 2025. "Learning in College." AEA RCT Registry. September 08. https://doi.org/10.1257/rct.16671-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

We will recruit college students to study how they search for information regarding skill and signal augmenting inputs.

Intervention (Hidden)

We will recruit college students to study how they search for information regarding skill and signal augmenting inputs. In particular, we will tell students that they will interact with an AI tool designed to help them navigate college life and how to succeed in getting into their desired career.
We will randomize them into one of two AI treatment groups. The control group will receive an AI tool that is limited to answering questions directly and that does not suggest topics for further discussion. The treatment group will instead receive an AI tool that is trained to bring up a larger set of topics outside of the direct question that is asked by the student.
After this AI intervention, students will be asked about their beliefs on an hypothetical student's starting salary as a function of their efforts in various actions/inputs: e.g. x hours studying, y hours networking.

Intervention Start Date

2025-09-08

Intervention End Date

2025-12-31

Primary Outcomes

Primary Outcomes (end points)

1) engagement in hidden curriculum topics
2) subjective beliefs on the wages of the hypothetical student after the AI intervention
3) heterogeneity across first-gen and continuing-gen students

Primary Outcomes (explanation)

Primary Hidden curriculum topics: Networking (peers/alumni/cold outreach)
Office hours (strategic use beyond homework help)
Relationship building with professors
Professional organizations/extracurriculars

Secondary Outcomes

Secondary Outcomes (end points)

How "sticky" topics are and secondary hidden curriculum topics: Research opportunities
Mentorship seeking

Secondary Outcomes (explanation)

Likelihood that the next message in the conversation is about the same actions that was asked in the previous message

Experimental Design

College students, balanced on whether they are first-gen or continuing-gen, will be recruited both on Prolific platform and through the XLab participant pool. Xlab participants will be current UC Berkeley students.
We will then elicit background information on their efforts in college.
Then students will be randomized into one of two AI conditions.
Finally we will elicit students beliefs on returns to effort in different skill augmenting inputs.

Experimental Design Details

College students, balanced on whether they are first-gen or continuing-gen, will be recruited both on Prolific platform and on XLab from current Berkeley students.
We will then elicit background information on their efforts in college.
Then students will be randomized into one of two AI conditions.
Finally we will elicit students beliefs on returns to effort in different skill augmenting inputs.
We will randomize them into one of two AI treatment groups. The control group will receive an AI tool that is limited to answering questions directly and cannot mention further discussion topics. The treatment group will instead receive an AI tool that is trained to bring up a larger set of topics outside of the direct question asked from the student.
After that we will use a vignette design to estimate subjective beliefs on returns of effort into inputs such as studying, networking, joining professional clubs. Returns will be measured in terms of expected wages of this hypothetical student.

Randomization Method

randomization will be done by computer

Randomization Unit

randomization will be at the individual level

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

1,200 students on Prolific and/or 1,000 student on Xlab, depending on availability of funding

Sample size: planned number of observations

1,200 students on Prolific and/or 1,000 student on Xlab, depending on availability of funding

Sample size (or number of clusters) by treatment arms

1,200 students on Prolific and/or 1,000 student on Xlab, depending on availability of funding
Both samples will be split 50/50 into one of the two treatments

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Supporting Documents and Materials

IRB