The Impact of Consumer-Side Discrimination on Firm Productivity in Sub-Saharan Africa

How to cite

Kelley, Erin; Lane, Gregory; Pecenco, Matthew; Rubin, Edward. 2020. "The Impact of Consumer-Side Discrimination on Firm Productivity in Sub-Saharan Africa." AEA RCT Registry. AEARCTR-0006698. https://doi.org/10.1257/rct.6698-1.2

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

Erin Kelley* , Gregory Lane , Matthew Pecenco , Edward Rubin

* Primary investigator

Registry ID
AEARCTR-0006698
DOI
https://doi.org/10.1257/rct.6698-1.2
First registered
Last updated
Status
Completed
Trial dates
2019-01-01 – 2020-11-11
Countries
Kenya (District of Columbia)
Registry record
http://www.socialscienceregistry.org/trials/6698

Abstract

Our proposed work returns to the open question of the existence and extent of consumer- based discrimination, asking what role consumers play in gender-based discrimination in labor markets and workplaces in the developing world. Using experimentally-induced variation in consumers’ perceptions of online chat agents’ genders, we will estimate the effect of consumer-side discrimination on both individual and aggregate worker productivity. Specifically, we partner with an online sales company and randomly assign gender-specific names to agents, careful to remove any indications of the assigned names from the agents themselves.

Intervention

To understand the role that consumer-side perceptions of gender play in determining the productivity of workers, we implement a field experiment with a single treatment arm. In this experiment, each day we randomly select a male or female name to present to consumers for each chat agent at an online sales company.

Experimental design

Our main test measures the causal effect of consumer-side discrimination on worker productivity by regressing a daily measure of productivity for an individual chat agent, on an indicator for being assigned to a female name.

Design details

Randomization method
Randomization was done in office by a computer
Randomization unit
We randomize at the agent-by-day level.
Clusters
We work with a sales platform with 6 chat agents. We randomly assign each agent a female or male name on a daily basis.
Observations
We randomize 6 agents daily to male and female names across 10 months, resulting in 1800 units.
Arms
We have approximately 900 units in treatment (female assigned name) and 900 units in control (male assigned name)
Primary outcomes
Number of Chat; Chat Length (Words, Time); Customer Ever Responded; Agent First Message Length; Customer First Message; Agent First Message Sentiment; Customer First Message Sentiment; Probability of making a sale.

Cited by

No citation data available for this registration yet.

Keywords