Understanding Student Perceptions of College Applications Recommendations

How to cite

Monachou, Faidra. 2024. "Understanding Student Perceptions of College Applications Recommendations." AEA RCT Registry. AEARCTR-0014368. https://doi.org/10.1257/rct.14368-1.0

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

Faidra Monachou*

* Primary investigator

Registry ID
AEARCTR-0014368
DOI
https://doi.org/10.1257/rct.14368-1.0
First registered
Last updated
Status
In development
Trial dates
2024-09-15 – 2026-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/14368

Abstract

The study involves a survey of high school students, administered at their schools during regular school hours. The questionnaire will present different scenarios involving hypothetical students applying to college, their initial college portfolio choices, and the recommendations given to them by either a human or algorithmic recommender. Half of the subjects will be randomly assigned a 'human recommender' questionnaire, while the other half will receive an 'algorithmic recommender' questionnaire. The study aims to (1) measure differences in students’ perceptions of college application recommendations when these are provided by a human counselor vs. an algorithm, and (2) identify the mechanisms that drive students’ perceptions of such recommendations.

Intervention

Half of the students will be randomly assigned a 'human recommender' questionnaire, while the other half will receive an 'algorithmic recommender' questionnaire. We will measure differences in student perceptions of human- vs. algorithm-based recommendations by presenting three hypothetical scenarios about the choices of three fictional students.

Experimental design

With the same recommendations for the fictional student in each scenario,
- Control Arm: Participants will be informed that the recommendations for the hypothetical student were made by a (human) career counsellor.
- Treatment Arm: Participants will be informed that the recommendations that the recommendations for the hypothetical student were made by an algorithmic recommendation system.

Design details

Randomization method
The paper-based questionnaires from both arms are randomized and then distributed to the participants in each classroom.
Randomization unit
Individual (student)
Clusters
no clusters (1,100 individual students)
Observations
1,100 students
Arms
human recommender: 550 students, algorithmic recommender: 550 students
Primary outcomes
Recommendation Adoption (Binary): The key outcome variable assesses if the respondent agrees with the given recommendation, or switches back to the original choice of the fictional students.
Secondary outcomes
- Trust in Intent (Numerical): The variable measures the respondent's trust that the recommender is acting according to persona's best interest. - Trust in Competence (Numerical): The variable measures the respondent's trust that the recommender is able to make correct and appropriate recommendation. - Understanding of Rationale (Numerical): The variable measures the respondent's self-assessed understanding of the rationale of the recommendation. - Compatibility of Goals (Numerical): The variable measures the extent to which the respondent agrees that the recommender has the same goals as that of the hypothetical character for creating the recommendation.

Cited by

No citation data available for this registration yet.

Keywords