Information Frictions among Firm Recruiters

How to cite

Russell, Andelyn. 2024. "Information Frictions among Firm Recruiters." AEA RCT Registry. AEARCTR-0014131. https://doi.org/10.1257/rct.14131-2.0

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

Andelyn Russell*

* Primary investigator

Registry ID
AEARCTR-0014131
DOI
https://doi.org/10.1257/rct.14131-2.0
First registered
Last updated
Status
On going
Trial dates
2024-08-05 – 2025-01-31
Countries
Private
JEL codes
J320, M510, O150
Registry record
http://www.socialscienceregistry.org/trials/14131

Abstract

Jobseekers across countries and cohorts value amenities. However, in developing countries, many job descriptions are sparse and thus may not include the full range of amenities that firms offer. In the first part of this project, I partnered with a large Indian job board to analyze whether a lack of amenities on job descriptions forms an information friction. Conducting a data collection exercise and survey experiment with first-time jobseekers, we found that firms have amenities that jobseekers value, but do not include them on job descriptions. In the second part of the project (this experiment), I partner with the job board to examine one reason why valued amenities are missing from job descriptions: firms may not fully be aware of jobseekers’ preferences. We test whether informing recruiters of jobseekers’ preferences can lead them to update their job descriptions with valued amenities.

Intervention

If a recruiter confirms that their firm provides job training to new employees, a customer support agent provides the estimate of how much first-time jobseekers value training in the study context.

Experimental design

Background:

In the first part of this project, I partnered with a large Indian job board to analyze whether a lack of detail on job descriptions forms an information friction for jobseekers, preventing them from identifying jobs with amenities that they prefer. We first conducted surveys of recruiters who had posted entry-level white collar jobs in which we asked them to describe four aspects of the job:

- Work format (whether work is completed individually or in a team)
- Job training
- Work schedule
- Transportation support

From the survey results, we found that there is significant variation in these features among entry-level white-collar jobs. Furthermore, most firms that provide these amenities do not list them on job descriptions.

We then elicited jobseekers’ preferences over these amenities through a survey experiment. Analyzing the results, we found that jobseekers who are most likely to be searching for their first job (no experience on their profile and have created accounts within the past 60 days) valued jobs that provided training and transportation support.

Information Experiment:

In this part of the project, we investigate one reason why valued amenities are missing from job descriptions: recruiters might not be fully aware of jobseekers’ preferences. To test this, we conduct an information experiment with recruiters. Because job training is provided by the majority of firms, we test the effect of informing recruiters about jobseekers’ preferences for training. Recruiters who receive a call from the study partner’s customer support team to complete platform onboarding procedures are randomized among two conditions:

- Control: If the recruiter confirms that their firm provides training, the customer support agent simply tells them that they can add training details to their job description. The customer support agent is instructed to provide limited clarification.
- Treatment: If the recruiter confirms that their firm provides training, the customer support agent explains that “You’ll get as many candidate applications as other companies which provide Rs 460 higher salary by updating training details in the job description.” The customer support agent can clarify this statement.

Both conditions are designed to reflect the communication style and operations of the platform’s customer support center.

Design details

Randomization method
Randomization is conducted using Google’s online coin flip at the start of an agent's work day.
Randomization unit
Randomization is conducted at the agent-day level.
Clusters
The number of clusters depends on the number of agent-days required to reach 300 calls.
Observations
300
Arms
150 recruiters in treatment, 150 recruiters in control
Primary outcomes
Indicator for whether recruiters update their first job post to describe job training.
Secondary outcomes
- Indicator for whether recruiters update their first job post in any way. - Conditional on making a subsequent post within 30 days, an indicator for whether recruiters describe job training on at least one of these posts.

Cited by

No citation data available for this registration yet.

Keywords