Automation Risk From AI Affects Young Adults Occupation Choice (old see new version at RCT ID AEARCTR-0013427)

How to cite

Rundström, Marcus. 2024. "Automation Risk From AI Affects Young Adults Occupation Choice (old see new version at RCT ID AEARCTR-0013427)." AEA RCT Registry. AEARCTR-0013424. https://doi.org/10.1257/rct.13424-2.2

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

Marcus Rundström*

* Primary investigator

Registry ID
AEARCTR-0013424
DOI
https://doi.org/10.1257/rct.13424-2.2
First registered
Last updated
Status
Withdrawn
Trial dates
2024-04-21 – 2024-05-15
Countries
United States of America ()
JEL codes
A13, C90, J20, J24
Registry record
http://www.socialscienceregistry.org/trials/13424

Abstract

Will the possibility of AI replacement affect career decisions for young adults? We test this using an RCT with information treatment. I leverage an online experiment to investigate how automation risk affects the attractiveness of occupations and the desirability of entering a specific occupation. The experiment consisted of one treatment group and a control group.

I am modifying a simple model by Wiswall and Zafar (2018), which predicts a lower probability of entering an occupation if the automation risk for the occupation is underestimated. While the outcome may be intuitive, I am the first to study how occupational automation risk affects the desirability of entering a specific occupation. Compared to previous work on automation’s effect on labor markets, I go one step further
and extend the literature by Autor et al. (2003) to answer how labor market outcomes change as an effect of technological development by including AI in the automation risk.

Intervention

I first let everyone guess the automation risk for four occupations: teachers, nurses, office clerks, and economists. Then, I provided the automation risk for the treatment group. Finally, participants will answer questions about these occupations to see if automation risk changes perceptions about these jobs.

Experimental design

I construct two I use an online experiment consisting of two groups. Both groups first answer questions about their background, job preferences, and belief in the automation risk for four occupations: nurses, teachers, office clerks, and economists. The treatment group is provided with the automation risk for these occupations. The last block of questions shows up after treatment. Both groups answer these.

Design details

Randomization method
I use the randomization feature in Qualtrics.
Randomization unit
individual.
Clusters
There are no clusters, only individuals.
Observations
600 individuals.
Arms
Among 300 individuals in the treatment group and 300 individuals in the control group.
Primary outcomes
Attractiveness to occupation, and probability of entering the occupation.
Secondary outcomes
Social status, future salary, job security, retirement age, and preferences for political action.

Cited by

No citation data available for this registration yet.

Keywords