Career Choices under Incomplete Information

How to cite

Meyer, Carl; Lorenz, Nora. 2026. "Career Choices under Incomplete Information." AEA RCT Registry. AEARCTR-0017179. https://doi.org/10.1257/rct.17179-1.0

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

Carl Meyer* , Nora Lorenz

* Primary investigator

Registry ID
AEARCTR-0017179
DOI
https://doi.org/10.1257/rct.17179-1.0
First registered
Last updated
Status
In development
Trial dates
2026-01-23 – 2031-01-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/17179

Abstract

This project examines how higher-education students form occupational expectations and aspirations and how those expectations shape their career choices. In a randomized online survey of students in Germany, we (i) measure baseline beliefs about expected occupations given one's field of study; (ii) exogenously provide field-specific labor market demand and AI-exposure information; and (iii) estimate how information updates beliefs in a discrete-choice framework. Follow-up data and merged register data link beliefs to job market outcomes.

Experimental design

- Individual-level randomization into information conditions. The instrument contains: (i) field-specific demand information from the Bertelsmann Jobmonitor, (ii) AI exposure message for the respondent’s desired occupation based on an ILO study, and (iii) demand information plus occupation-level AI exposure categories based on an ILO study and AI exposure message for the respondent’s desired occupation based on an ILO study. People are randomly assigned to one of the treatment arms or a control arm, in which they are asked about recent coursework. We will assign the same number of students to each treatment arm.
- Pre/post within-survey design: baseline outcomes are measured before the information block(s); endline outcomes after.
- DCE: three rounds of two-alternative forced-choice between hypothetical job ads (no opt-out in current instrument). Attributes shown include gross annual salary, AI exposure (approx. 10-year horizon), and job stability after three years; in one version job titles differ across alternatives.

Design details

Randomization method
Randomization done in office by a computer (STATA)
Randomization unit
Individual (stratified by field of study)
Clusters
Contactable N ≈ 14,000 students. Anticipated completes: 1,000–3,000 depending on take-up.
Observations
Contactable N ≈ 14,000 students. Anticipated completes: 1,000–3,000 depending on take-up.
Arms
25% (approx. 3500 students) in each of the three treatment arms and control.
Primary outcomes
Primary: - Expectedoccupation (pre/post intervention): desired occupation after final degree (top 3) and desired occupation after 10 years. - Expected earnings (pre/post): gross annual earnings in first year post-graduationand 10 years after graduation. - Perceived task automation in intended occupations (pre/post): perceived automatable share in % for the desired occupation and desired 10-year occupation. - Subjective probabilities of post-graduation states (pre/post): allocation across (i) working in desired occupation, (ii) working in another occupation, (iii) other (incl. unemployment) for one year post-graduation and 10 years. - Discrete choice experiment (DCE): choices across three 2-alternative choice sets; estimate marginal utilities and willingness-to-pay (WTP) for job stability, preferred occupation, and AI exposure relative to salary.
Secondary outcomes
Secondary: - Population-belief elicitation (incentivized): perceived top-3 occupations most in demand for graduates with respondent’s profile and expected first-year earnings for someone with the same profile. - AI attitudes and self-assessed familiarity: perceived overall AI influence on the labor market and self-assessed AI expertise. - Treatment comprehension and study-related covariates and skills.

Cited by

No citation data available for this registration yet.

Keywords