How migrant stocks and migrant flows affect political outcomes

How to cite

Langlotz, Sarah; Dreher, Axel; Matzat, Johannes; Parsons, Christopher. 2024. "How migrant stocks and migrant flows affect political outcomes." AEA RCT Registry. AEARCTR-0014094. https://doi.org/10.1257/rct.14094-1.0

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

Sarah Langlotz* , Axel Dreher , Johannes Matzat , Christopher Parsons

* Primary investigator

Registry ID
AEARCTR-0014094
DOI
https://doi.org/10.1257/rct.14094-1.0
First registered
Last updated
Status
In development
Trial dates
2024-08-30 – 2024-09-13
Countries
Private
JEL codes
J15, F22, F63, D72
Registry record
http://www.socialscienceregistry.org/trials/14094

Abstract

We investigate how changes in migrant stocks as opposed to migrant flows affect political outcomes in the United States. We collect respondents’ quantitative estimates of the shares of the population that they consider migrant stocks as compared to migrant inflows, which we define as people who arrived in the last two years in the US and came from other countries to live in the US. We delineate five randomized experimental groups, to check whether the correlations between immigration perceptions and political attitudes are causal. Specifically, we provide the treatment groups with information on the actual shares of migrant stocks or migrant flows. We also present two different hypothetical future scenarios of the change of these respective figures. We collect detailed information on respondents concerns about the future situation of the country and their political preferences. We further collate information on individuals contact with foreign-born people as well as their media consumption.

Intervention

We run an online survey in the United States that includes an information provision experiment. To ensure our sample is representative across key demographics such as gender, income, and age, we recruit participants through a professional survey company. All participants are required to be at least 18 years old.

We perform five randomized experimental groups, to check whether the correlations between immigration perceptions and political attitudes are causal. Precisely, we provide the treatment groups with information on either the actual share or level of migrant stocks or flows. We also present two different hypothetical future scenarios of the change of these respective figures.

Experimental design

Our design consists of five stages. First, we collect information on participants’ demographics, media consumption, and political preferences. Second, we introduce all participants to the definition of migrant stocks versus flows and then ask them about their perceptions as to the shares of each group for the typical US county. Third, random subsets of respondents are assigned to five groups, which differ in terms of the type and quantity of information provided. The control group receives no information whatsoever. The information provided to these four treatment groups differs along two dimensions: the actual figure on migrant stocks OR flows (based on census data from recent decades).
Additionally, treatment groups receive a hypothetical scenario detailing future changes in the respective figure presented: indicating either a constant development OR an increase (which we infer from one standard deviation change in the actual census data in the years prior). We highlight that this is a hypothetical scenario, while also providing this information in the debriefing.
Fourth, respondents are asked about their concerns regarding the future of the country (political, economic, social). They are then presented four hypothetical political candidates with their views on top tax rates, legal immigration and the size of the military, which allows them to be classified along the left-right spectrum (following broadly the procedure as in PEW). Participants have to rank who could best tackle the aforementioned concerns for the future of the country. We additionally tell respondents that they have been automatically enrolled in a lottery with a $100 prize. Before they know whether or not they have won, they need to commit to donating at least 50$ of it, or anything between 50$-100$ to one or more of the four political candidates. We use the respondents’ preferences for donations as a (real) measure of support for their respective candidates.
Fifth, we ask about additional outcomes based on related literature, which directly relate to respondents’ immigration preferences. Finally, we provide a debriefing, including a link through which individuals can get additional information regarding recent developments and future projections, all from non-partisan sources.

Design details

Randomization method
Respondents are randomly assigned to control and treatment groups by the Qualtrics software.
Randomization unit
Individual participant
Clusters
3500 individuals
Observations
3500 individuals
Arms
700 individuals
Primary outcomes
We can measure within-subject and between-subject effects on most outcomes listed below: • Concerns about the situation in the country in the near future • Donations to more/less extreme (hypothetical) candidates • Preference for more/less extreme (hypothetical) candidates • Change in political ideology (based on pre-treatment response on political ideology and post-treatment candidate ranking, from which we infer ideology) • Concerns about migrant (refugee) stocks/flows We plan to measure heterogeneity across multiple individual and local (inferred from ZIP code) characteristics
Secondary outcomes
We can also learn the following from the survey: • Have (extreme) right-wing voters o Estimated less precisely actual stocks/flows o Different concerns on topics o Different perceptions on views on immigrants/recently arrived/refugees o Different media usage o Less contact with those born overseas o Thought differently about survey bias • Have Extreme (any) voters o Estimated less precisely actual stocks/flows o Different concerns on topics o Different perceptions on views on immigrants/recently arrived/refugees o Different media usage o Less/more contact with those born overseas o Thought differently about survey bias

Cited by

No citation data available for this registration yet.

Keywords