Tastes, expectations and gendered job - pilot
How to cite
Delfino, Alexia. 2019. "Tastes, expectations and gendered job - pilot." AEA RCT Registry. AEARCTR-0004571. https://doi.org/10.1257/rct.4571-1.2
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 proposed project consists in a pilot of a field recruitment experiment. The project is a spin-off of the field experiment conducted in 2017 and registered on the AEA RCT registry under the title “Changing Collars? Social identity and selection in frontline jobs”. The past experiment aimed at disentangling two different channels through which social identity can affect men’s selection into female-dominated jobs: preferences for working with people of own gender and expectations of performance. Results from this first experiment suggest that, out of the two forces of interest, performance expectations matter more for men’s application behaviour. The current proposed project aims at testing the same channels in a male-dominated sector: website development and design. Do performance expectations matter more also for women’s selection into male-dominated jobs? In the experiment, a pool of 2000 freelancers will be invited to apply for a job in website development and design. They will be assigned to one of four treatment groups. The job will be presented exactly in the same way in all the treatments, expect for i) a photograph and ii) a piece of information about the quality of freelancers hired for similar jobs. The two conditions will be cross-randomized, for a total of 4 treatment groups. The main outcome variables will be whether freelancers apply, their quality and the quality of their application, answers to interviews questions and whether they are shortlisted among the top applicants. The sample size has been chosen through power analysis to be able to detect a minimum effect size in the change of application rates of 0.25 of a standard deviation for each condition. Freelancers will be chosen among the users of one freelancers website.
Experimental design
In the experiment, a pool of 1600 to 2000 freelancers will be invited to apply for a job in website development and design. They will be assigned to one of four treatment groups. The job will be presented exactly in the same way in all the treatments, expect for i) a photograph and ii) a piece of information about the quality of freelancers hired for similar jobs. The photograph could be of a man or a woman. The text could say that 60% or 80% of freelancers hired for similar jobs got 5 stars by the clients. The two statistics are both true (thus there is no deception), but they refer to different job categories (as stated in the job posting at the bottom of the “intervention box”). The two conditions will be cross-randomized, for a total of 4 treatment groups.
Design details
Cited by
No citation data available for this registration yet.