Ageism Unveiled: Understanding Discrimination Against the Elderly

How to cite

Feng, Jiajing. 2025. "Ageism Unveiled: Understanding Discrimination Against the Elderly." AEA RCT Registry. AEARCTR-0015742. https://doi.org/10.1257/rct.15742-1.0

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

Jiajing Feng*

* Primary investigator

Registry ID
AEARCTR-0015742
DOI
https://doi.org/10.1257/rct.15742-1.0
First registered
Last updated
Status
In development
Trial dates
2025-04-07 – 2025-04-21
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/15742

Abstract

This is a follow-up experiment to the previously registered study “Ageism Unveiled: Understanding Discrimination Against the Elderly.” The previous experiment investigated the presence of age discrimination in different types of labor markets: those requiring fluid intelligence and those relying on crystallized intelligence. The current experiment examines whether workers anticipate age discrimination and how such anticipation affects their labor force participation decisions. Additionally, the experiment explores heterogeneity in these expectations across age, gender, and actual ability in the tasks.

Intervention

There are two types of random assignments in this experiment:

Task assignment across subjects: Participants will be randomly assigned to complete one of two tasks. Half of the participants will complete the Raven's Matrices task, and the other half will complete the vocabulary task.

Competing candidate profiles within subjects: Each participant will be shown four different competing candidate profiles: one identical to their own profile, one differing in gender, one differing in age, and one differing in both age and gender. The order in which these profiles are presented will be randomized.

Experimental design

This experiment contains two sections. In the first section, the workers complete a task: either the raven's matrices task, which measures fluid intelligence, or the vocabulary task, which measures crystalized intelligence. The task will be randomized across subjects.

The second section elicits anticipation about discrimination and willingness to participate in the labor market. The workers will compete with four other candidates in a labor market to get hired for the task. Each worker will see 4 cases, and in each case, they will compete with a different candidate (varying age and gender). For each case, they will complete two tasks:

Task 1:
Workers will be asked to predict how typical employers (average Americans aged 36 to 55) made hiring decisions when choosing between two worker profiles. Specifically, they need to estimate how many employers out of 100 would hire them and how many would hire the competing candidate to do the task. If their guess is within 5 percentage points of the actual result (based on decisions made in previous sessions with the same profiles), they will earn a bonus of $0.25.

Task 2:
In this task, workers need to choose between fixed payment or entering the labor market and compete to be hired by employers. If they enter the labor market and get hired, they earn a $1 bonus. If they enter the labor market but are not hired, they receive no bonus for this section.
They will make this choice 11 times per case, with the fixed payment amount varying each time. At the end of the experiment, one case and one of the 11 choices will be randomly selected. Their decision in that round will determine their payoff. If they chose the fixed payment, they receive the corresponding amount. If they choose to enter the labor market, the hiring outcome will be based on average decisions from previous sessions: if employers on average preferred hiring someone with their profile over the competitor’s profile, they will be hired; if employers on average preferred hiring someone with the competing profile, they will not be hired.

At the end of the experiment, workers' risk preferences and willingness to compete will be elicited.

Design details

Randomization method
There are two types of random assignments in this experiment: Task assignment across subjects: Participants will be randomly assigned to complete one of two tasks. Half of the participants will complete the Raven's Matrices task, and the other half will complete the vocabulary task. Competing candidate profiles within subjects: Each participant will be shown four different competing candidate profiles: one identical to their own profile, one differing in gender, one differing in age, and one differing in both age and gender. The order in which these profiles are presented will be randomized. The task will be randomized across subjects. The randomization will be done through Prolific. The competing workers' profiles will be randomized within subjects. This will be done through the randomization function (loop and merge) on Qualtrics.
Randomization unit
The task will be randomized across subjects. The randomization will be done through Prolific. The competing workers' profiles will be randomized within subjects. This will be done through the randomization function (loop and merge) on Qualtrics.
Clusters
For this experiment, I will hire 300 older subjects (aged 56-65) and 300 younger subjects (aged 26-35).
Observations
For this experiment, I will hire 300 older subjects (aged 56-65) and 300 younger subjects (aged 26-35). Half of each group will complete the Raven's matrices task, the other half will complete the vocabulary task. Competing candidate profiles within subjects: Each participant will be shown four different competing candidate profiles varying age and gender. The total number of observations (in long format) would be 600*4 = 2400
Arms
For this experiment, I will hire 300 older subjects (aged 56-65) and 300 younger subjects (aged 26-35). Half of each group will complete the Raven's matrices task, the other half will complete the vocabulary task. Competing candidate profiles within subjects: Each participant will be shown four different competing candidate profiles: one identical to their own profile, one differing in gender, one differing in age, and one differing in both age and gender. The order in which these profiles are presented will be randomized. In each task, there will be 300 subjects, 300*4=1200 observations. For each competing profile, there will be 600 subjects (summing up two tasks) making the labor market participation decision.
Primary outcomes
1. Belief about anticipation of being discriminated (they need to estimate how many employers out of 100 would hire them and how many would hire the competing candidate to do the task). 2. Willingness to participate in the labor market through a multiple price list.

Cited by

No citation data available for this registration yet.

Keywords