Automation Risk From AI Affects Young Adults Occupation Choice

Last registered on April 26, 2024


Trial Information

General Information

Automation Risk From AI Affects Young Adults Occupation Choice
Initial registration date
April 21, 2024

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
April 26, 2024, 11:48 AM EDT

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



Primary Investigator

Lund University

Other Primary Investigator(s)

Additional Trial Information

On going
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
I first construct a theoretical model to explain and predict the outcome. The intuition behind the model is simple: underestimating the automation risk will force you to revise your beliefs, leading to a lower probability of entering a specific occupation.
I focus on three occupations: Teachers, nurses, office clerks, and economists. It is important to understand the determinants for labor supply in these occupations and focus on automation risk since nobody has provided causal evidence that automation risk is a determinant for labor supply in these occupations. To investigate my research question, I leverage an online experiment and pay respondents from Prolific to participate in my study. Respondents answer a survey with questions related to the topic, such as the attractiveness, and probability of entering a specific occupation, job preferences, and background questions. The experiment consists of one treatment group and one control group. The treatment consists of the automation risk calculated by A gender and race perspective are added. I also perform an additional analysis including political preferences, social status belief about salary.
External Link(s)

Registration Citation

Rundström, Marcus. 2024. "Automation Risk From AI Affects Young Adults Occupation Choice." AEA RCT Registry. April 26.
Experimental Details


Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Attractiveness of entering the occupation, and probability of entering the occupation.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Social status, future salary, job security, retirement age, and preferences for political action.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experimental data was collected through a survey where I paid participants $1.2 at Prolific, an online platform with a large pool of respondents. The sample consists of young adults between the ages of 18-20 in the US. The advantage of
using Prolific is that it allows access to a wide range of respondents. Respondents answer a survey consisting of four blocks. In the first block, subjects filled
in their background information. In the second block, respondents are asked to report job preferences and estimate the automation risk for nurses, teachers, economists, and office clerks. I subjectively chose as broad occupations as possible. However, the questions concerning job preferences are unrelated to any specific profession. Instead, these questions related to previous research determinants of college majors and job preferences, such as social status, expected earnings, flexibility, family, and job stability. After the second block, respondents were randomized to the treatment group or the control group. I used Qualtrics to manage the survey as it allowed for the randomization of respondents.
The treatment consists of brief background information about AI, recent developments such as ChatGPT, and the occupational automation risk computed by (2024). Respondents were asked to read the information carefully. The final block consists of questions about the four occupations and additional individual beliefs.

To test the balance in my experimental groups, I used a standard t-test to compare the means of background variables.
Experimental Design Details
I hypothesize/predict:
1. Correct information about automation risk does affect the attractiveness
of entering the profession.
2. Correct information about automation risk does affect the probability
of entering the profession.

I perform additional analyses on Social status, future salary, job security, retirement age, and preferences for political action. I have no clear hypotheses for these outcomes. These outcomes partly aim to investigate mechanisms.
Randomization Method
I use the randomization feature in Qualtrics. Similar to flipping a coin.
Randomization Unit
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
Sample size: planned number of observations
Sample size (or number of clusters) by treatment arms
300 in the control group and 300 in the treatment group.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
I used my bachelor thesis, an old study by myself, as a baseline. The standard deviation was about 1.4 for the main outcome. I assume a standard deviation 1.3 as including controls in the regression will improve the precision. Moreover, I assume an absolute effect of 0.3. I assume I want to reject the null hypothesis at a 0.05 significance level, I calculate I need around 600 respondents. Given 600 respondents, I have 80% power to find an effect of 0.2 standard deviation.

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number


Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials