Racial Discrimination in Seeking Advice

How to cite

Bartos, Vojtech; Glogowsky, Ulrich; Rincke, Johannes. 2020. "Racial Discrimination in Seeking Advice." AEA RCT Registry. AEARCTR-0005812. https://doi.org/10.1257/rct.5812-2.0

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

Vojtech Bartos* , Ulrich Glogowsky , Johannes Rincke

* Primary investigator

Registry ID
AEARCTR-0005812
DOI
https://doi.org/10.1257/rct.5812-2.0
First registered
Last updated
Status
Completed
Trial dates
2020-05-11 – 2021-05-11
Countries
United States of America ()
JEL codes
C90, D01, D83, J15, J71
Registry record
http://www.socialscienceregistry.org/trials/5812

Abstract

We design an online experiment to study racial discrimination in seeking advice. In round 1 of the experimental design, subjects face a real effort task that is difficult to solve without prior advice from an expert. We offer subjects the option to watch a tutorial before working on the task. The main treatment variation is the race of the advisor (black vs. white), signalled by the skin color of a hand appearing at the beginning of the tutorial. We vary the skin color of a given hand model using video post-production techniques. This allows us to keep all features of the hand other than skin color constant between treatment arms. In round 1, we analyze how subjects’ willingness to pay for advice depends on the race of the advisor, and how the race of the advisor affects advice utilization. In round 2, subjects watch another tutorial containing advice about a different strategy to solve the real effort task. To elicit preferences, we let subjects choose between two different advisors. Using an information treatment stating that the content of both tutorials is identical, we identify the extent of taste-based discrimination in seeking advice.

Intervention

See abstract

Experimental design

The experiment has two rounds. In round 1 of the experimental design, subjects face a real effort task that is difficult to solve without prior advice from an expert. We offer subjects the option to watch a tutorial before working on the task. The main treatment variation is the race of the advisor (black vs. white), signalled by the skin color of a hand appearing at the beginning of the tutorial. We vary the skin color of a given hand model using video post-production techniques. This allows us to keep all features of the hand other than skin color constant between treatment arms. In round 1, we analyze how subjects’ willingness to pay for advice depends on the race of the advisor, and how the race of the advisor affects advice utilization. In round 2, subjects watch another tutorial containing advice about a different strategy to solve the real effort task. To elicit preferences, we let subjects choose between two different advisors. Using an information treatment stating that the content of both tutorials is identical, we identify the extent of taste-based discrimination in seeking advice.

Design details

Randomization method
Randomization done by a computer on a remote server (automated by the program).
Randomization unit
Individual
Clusters
1100 individuals
Observations
1100 individuals
Arms
See PAP for details
Primary outcomes
- individual’s willingness to pay for being advised by the first advisor in round 1 - individual's preference ordering for first and second round advisors in round 2 (categorical) - individual's preference for round 2 advisor (constructed dummy) - number of solved puzzles
Secondary outcomes
- individual used round 1 strategy (constructed dummy) - individual used round 2 strategy (constructed dummy) - number of puzzles solved with the strategy proposed by round 1 advisor (constructed) - number of puzzles solved with the strategy proposed by round 2 advisor (constructed) - share of puzzles solved with the strategy proposed by round 1 advisor (constructed) - share of puzzles solved with the strategy proposed by round 2 advisor (constructed) - willingness to pay for being advised by preferred advisor in round 2 (constructed) - completion time per puzzle (constructed) - number of moves used to solve the puzzle (constructed)

Cited by

No citation data available for this registration yet.

Keywords