Do Gender-Neutral Job Ads Promote Diversity?
How to cite
Fujiwara, Thomas; Carpio, Lucia Del. 2020. "Do Gender-Neutral Job Ads Promote Diversity?." AEA RCT Registry. AEARCTR-0005509. https://doi.org/10.1257/rct.5509-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
The project seeks to study the effect of using gender-neutral language in job advertisements on the quantity, quality (level of education and skill), and gender breakdown of the applicant pool in a widely used platform for the hiring of talent in the Tech Sector in Latin America (where Spanish is the main language). We implement a randomized controlled trial in which Job ads are randomly selected, at pre-approval, to be compliant with gender-neutral language.
Experimental design
There are two possible status for a job ad. A job ad under “control” status is treated in the same way the platform usually treats its ads. A job ad under “treatment” status must comply with a gender-neutral language protocol. The randomization is done at the Job Ad level in the following manner. Each time a company submits a job ad for pre-approval, a random draw occurs and with 50% chance the ad is assigned to control status, and with 50% it is assigned to treatment status. We follow the application processes of all the job ads that constitute our sample until the closing of the ad and application process.
The estimated effect of the treatment(s) (“intent-to-treat”) can be obtained by comparing average outcomes between treatment and control groups. We also estimate the effect of treatment on whether or not the ad has gendered language and use this to re-scale the “intent-to-treat” estimates appropriately to obtain the estimated effect of having gendered language. The analysis will also include a series of robustness checks, such as using different measures of outcomes, controlling for additional variables (such as the date that the ad is posted, the location of the job, and testing for effect heterogeneity (e.g., are effects larger for ads from a particular sector or region). We are also interested in measuring whether treatment has an impact on the gender neutrality of subsequent ads posted by the company. To study this, we will compare the gender-neutrality of the language in the second, third, fourth, and so forth, ad that the company posts in the study period.
Design details
Cited by
No citation data available for this registration yet.