Xenophobia and Social Desirability: MTurk Online Survey

How to cite

Fiorin, Stefano; Egorov, Georgy; Bursztyn, Leonardo. 2016. "Xenophobia and Social Desirability: MTurk Online Survey." AEA RCT Registry. AEARCTR-0001752. https://doi.org/10.1257/rct.1752-3.0

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

Stefano Fiorin* , Georgy Egorov , Leonardo Bursztyn

* Primary investigator

Registry ID
AEARCTR-0001752
DOI
https://doi.org/10.1257/rct.1752-3.0
First registered
Last updated
Status
Completed
Trial dates
2016-10-31 – 2016-12-31
Countries
United States of America (Alabama, Arkansas, Idaho, Mississippi, Nebraska, Oklahoma, West Virginia, Wyoming), United States of America (Pittsburgh)
Registry record
http://www.socialscienceregistry.org/trials/1752

Abstract

We will recruit survey participants on MTurk to conduct an online survey. With this survey, we plan to study 1) whether participants under-report xenophobic sentiments due to their social undesirability, and 2) whether the social undesirability is reduced by knowing that a large proportion of the population supports a presidential candidate with xenophobic views.
First, all subject will answer a few demographic questions. Second, they will be randomized into two groups. Subjects in the control group will receive no information. Subjects in the treatment group will be told what is the probability of a presidential candidate with xenophobic views to win the next election: our hypothesis is that knowing that the presidential candidate has large support, will lead subjects to revise their belief about the social undesirability of expressing xenophobic sentiments. We will collect evidence about this by asking subjects in both groups how many people they think hold xenophobic sentiments in their state. Third, to test our two hypotheses, we will measure subjects willingness to donate money to an immigration-reduction organization. Participants will be randomized into two groups: in one group we will emphasize that their decision is private; in the other group we will tell subjects that "a member of the research team might personally contact you to verify your answers", thus making the decision to be perceived as public. We will test our hypothesis by comparing the donation decisions of subjects in the four (2x2) experimental groups.

In a follow-up experiment to test the same hypotheses, we recruit participants living in the Pittsburgh area through an online panel survey firm. The design of this second experiment is similar to the first one. Participants are also asked to make a donation to an anti-immigration organization and to report how many people they think hold xenophobic views, and are randomized into four different treatments along two dimensions. First, subjects are either informed about the fact that “In the 2016 US Presidential Election, Donald Trump won Pittsburgh’s metropolitan area” or about the fact that “In the 2016 US Presidential Election, Hillary Clinton won Pittsburgh’s county.” Second, subjects are randomized either in a private or in public condition. In this case, however, we randomized the visibility of the donation with respect to participants’ potential peers and not with respect to the surveyor: one half of the participants were offered anonymity, while the other half expected their donation decision to appear on a website to be shared with all the other participants from their area.

Experimental design

We will recruit survey participants on MTurk to conduct an online survey. First, all subject will answer a few demographic questions. Second, they will be randomized into two groups. Subjects in the control group will receive no information. Subjects in the treatment group will be told what is the probability of a presidential candidate with xenophobic views to win the next election. Third, we will elicit subjects beliefs about how many people they think hold xenophobic sentiments in their state. Fourth, we will measure subjects willingness to donate money to an immigration-reduction organization. Participants will be randomized into two groups: in one group (private) we will emphasize that their decision is private; in the other group (public) we will tell subjects that "a member of the research team might personally contact you to verify your answers", thus making the decision to be perceived as public.

In a follow-up experiment to test the same hypotheses, we recruit participants living in the Pittsburgh area through an online panel survey firm. First, subjects are randomized into two groups. Subjects are either informed about the fact that “In the 2016 US Presidential Election, Donald Trump won Pittsburgh’s metropolitan area” (Trump won treatment) or about the fact that “In the 2016 US Presidential Election, Hillary Clinton won Pittsburgh’s county” (Clinton won treatment). Second, subjects are randomized either in a private or in public condition. In this case, however, we randomized the visibility of the donation with respect to participants’ potential peers and not with respect to the surveyor: one half of the participants were offered anonymity (private condition), while the other half expected their donation decision to appear on a website to be shared with all the other participants from their area (public condition). Third, participants are asked to make a donation to the anti-immigration organization. Fourth, they report how many people they think hold xenophobic views in Pittsburgh. At the end of the survey, the respondents are asked to answer a few demographic questions. The second experiment was conducted between August and October 2018.

Design details

Randomization method
The randomization will be done on Qualtrics, a website for conducting online surveys.
Randomization unit
The unit of randomization will be the individual subject.
Clusters
N/A
Observations
400 individuals About 1,600 individuals for the second experiment.
Arms
100 subjects control private, 100 subjects control public, 100 subjects treated private, 100 subjects treated public. 400 in Trump won private, 400 in Trump won public, 400 in Clinton won private, 400 in Clinton public.
Primary outcomes
Beliefs about the share of people holding xenophobic sentiments. Donation decisions.

Materials

Cited by

No citation data available for this registration yet.

Keywords