Value alignment in the public sector: from roles to managers to AI and citizens

How to cite

Choi, Iseul; Jo, Suyeon; Schnell, Sabina. 2026. "Value alignment in the public sector: from roles to managers to AI and citizens." AEA RCT Registry. AEARCTR-0018794. https://doi.org/10.1257/rct.18794-1.0

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

Iseul Choi* , Suyeon Jo , Sabina Schnell

* Primary investigator

Registry ID
AEARCTR-0018794
DOI
https://doi.org/10.1257/rct.18794-1.0
First registered
Last updated
Status
On going
Trial dates
2026-05-21 – 2026-06-22
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18794

Abstract

This study examines how organizational management styles and artificial intelligence (AI) framing influence the decision-making behavior of street-level bureaucrats in citizen-facing contexts. Using an online survey experiment, participants will be randomly assigned to scenarios that manipulate management style (e.g., bureaucratic, innovative, and participatory) and AI system value framing. Participants will then respond to hypothetical situations and be asked to answer post-vignette survey questions, allowing researchers to assess how these factors shape their behavioral orientations and reliance on AI-assisted recommendations in a decision-making process.
This research project involves three research questions: 1) how does organizational management style interacted with role perceptions shape the behavioral orientation of street-level bureaucrats in citizen-facing situations? 2) how does the interaction between management style and AI value framing shape which behavioral logic street-level bureaucrats default to in citizen-facing situations? 3) does value congruence between management style and AI framing increase bureaucrats' willingness to rely on the AI system? Data collected will consist of survey responses to experimental vignettes and follow-up questions measuring attitudes, behavioral intentions, and perceived appropriateness of AI use.

Experimental design

In the online survey experiment, we use a 3 x 4 factorial design where participants will be randomly assigned to scenarios that manipulate management style (3 conditions) and AI system value framing (4 conditions). Participants will then respond to hypothetical situations and be asked to answer post-vignette survey questions, allowing researchers to assess how these factors shape their behavioral orientations and reliance on AI-assisted recommendations in a decision-making process.

Design details

Randomization method
Use randomizer options in Qualtrics
Randomization unit
There are 12 groups where each participant is randomly assigned to.
Clusters
1800 individuals
Observations
1800 individuals
Arms
1800 individuals
Primary outcomes
Bureaucrats' behavior; Bureaucrats' perceptions to AI; Bureaucrats' willingness to use AI

Cited by

No citation data available for this registration yet.

Keywords