Experimental Design Details
Participants evaluate vignettes describing two hypothetical immigrants (A and B) with identical profiles: both are unmarried males in their thirties, share their respective countries' majority ethnicity, and have the same education level (randomly assigned as either both college graduates or non-college graduates). We also mention that both immigrants entered the UK legally five years ago upon receiving a job offer in the UK and are eligible for Indefinite Leave to Remain (ILR), having met all requirements including English proficiency and the Life in the UK test. Neither has a criminal record. By doing so, we ensure that both immigrants are viewed as compliant with UK law and standards, meeting key legal benchmarks. The only difference is their cultural origins: Immigrant B comes from a country culturally similar to the UK, while Immigrant A originates from a culturally distant country.
Cultural distance is measured using Wave 7 of the World Values Survey (WVS). Using this questionnaire, we focus on questions that reflect core beliefs, omitting items that capture individual preferences or attitudes related to a specific country. We construct three dimensions of cultural difference based on two key criteria: (1) maximising the response gap between the standardised responses in the UK and the top 10 immigrant-sending countries to the UK, and (2) including at least one country with responses similar to the UK's. The questions that meet these two criteria are: "When jobs are scarce, men should have more right to a job than women"; "The only acceptable religion is my religion"; "The army [should] take over when the government is incompetent."
In the vignettes, Immigrant B is shown as coming from a country where people typically "Disagree" (aligned with UK values) with the above statements, while Immigrant A comes from a country where people typically "Agree" (different from UK values). We omit country names to allow geographic associations to emerge only through the cultural values presented in the vignettes, if they emerge. Additionally, we elicit participants' perceptions of the country of origin of the hypothetical immigrants to assess whether the vignettes influence these associations.
We present participants with multiple scenarios in which they must choose between two immigrants with different income levels. Using a Multiple Price List design, we fix Immigrant B's income at £40,000 (close to both the skilled worker visa threshold and UK median earnings) and vary Immigrant A's income from £20,000 to £90,000 in £5,000 increments. To frame how participants should think about the economic relevance of income in this study, we explicitly state that tax contributions increase with income and contribute to public services. For each of these fifteen income scenarios, participants have to choose which immigrant they would prefer to remain in the UK when only one can stay, with an option to express indifference.
For each cultural dimension (gender, religion, and political attitudes), participants are randomly assigned to either control or treatment groups. In both groups, they choose between two immigrants: B (from a culturally similar country) and A (from a culturally distant country). Immigrant B's beliefs consistently align with UK values, while information about A's beliefs varies by treatment: In the control group, no information is provided about A's initial and current beliefs. In the treatment group, participants are informed that A's current beliefs align with typical British values. No information is provided about A's initial beliefs.
We design the treatment based on the following reasoning: People form attitudes toward immigrants based on their prior beliefs and assumptions when specific information is unavailable. The control group captures this baseline by providing no information about A's beliefs. The treatment group then reveals whether A has UK values, allowing us to measure how this information affects participants' preferences. We test one main hypothesis, which requires two testable assumptions to hold. First, when presented with equal incomes, participants in the control group on average prefer Immigrant B (culturally similar) over Immigrant A (culturally distant). Second, control group participants, on average, switch their choice to Immigrant A when his income exceeds Immigrant B's fixed income of £40,000.
Main Hypothesis: The income level at which participants switch to preferring Immigrant A differs between the control group (where A's beliefs are unknown) and the treatment group (where A has UK values). This switch point represents the compensating value natives require for accepting immigrants from culturally distant backgrounds. We test this hypothesis separately for each cultural dimension (gender, religion, and political attitudes), allowing us to measure how cultural assimilation affects natives' preferences across different aspects of culture.
Why would natives value assimilation? To answer this question, we examine which social concerns natives most strongly associate with immigrants from different cultural backgrounds. We fix Immigrant A's income at £40,000 annually—matching the income benchmark used throughout the study for Immigrant B—and ask participants to select their main concerns from a list spanning economic (e.g., labour market effects), cultural (e.g., impact on British culture), and general dimensions (e.g., rise in the level of crime, increase in demand for housing). By comparing the responses between treatment and control groups, we measure how information about cultural assimilation affects these underlying concerns about immigration from culturally different countries.
In addition, we measure participants' perceptions of Immigrant A's initial beliefs to examine how information about current beliefs influences backwards updating—specifically, whether knowledge of an immigrant's current alignment with UK values leads natives to revise their inferences about the immigrant's cultural beliefs upon arrival. We examine backwards updating across scenarios where A randomly earns £20,000, £40,000, and £90,000. If perceptions of initial beliefs remain similar across treatment and control groups, this would suggest that participants do not update their perceptions about the immigrant's beliefs upon arrival in the UK once they are aware of his current beliefs. Additionally, we measure participants' perceptions of the immigrant's initial overall integration indicators (language proficiency, understanding of local customs, sense of belonging, interactions with British society) and their expectations of future integration. We also elicit participants' assumptions about the immigrants' regions of origin with open-ended questions.
To assess whether responses to our hypothetical scenarios reflect real-world behaviour, we design two incentivised tasks involving real immigrants recruited through Prolific. These immigrants are selected from countries where average responses in the WVS align with those used in our hypothetical scenarios. In the control group, participants are informed that we recruited two immigrants from Prolific—one from a country with cultural values different from the UK and the other from a culturally similar country. In the treatment group, they receive the same information, along with additional details indicating that the first immigrant's values align with those of the UK (We also ask the immigrants themselves about their cultural values.) As in the hypothetical scenarios, the immigrant from the culturally similar country holds beliefs that align with UK norms in both the control and treatment groups. In the first task, participants decide which immigrant should receive assistance from a qualified immigration lawyer in their local area. To support this, we collect area-specific lists of immigration lawyers and inform participants that their choice will help the selected immigrant navigate the ILR application process more efficiently by receiving this information. One participant is randomly chosen to have their decision implemented. In the second task, participants allocate a monetary amount between the two immigrants. This task assesses social preferences, with one randomly selected allocation being implemented.
We define the primary outcome as the first switch point, i.e., the midpoint between the last income level at which participants preferred Immigrant B and the subsequent income level at which they either preferred Immigrant A or indicated indifference by selecting both A and B. For corner preferences, where participants' choices remain constant throughout the income range, we follow conventional methods in the literature (e.g., Allcott and Kessler, 2019). Our primary focus will be on respondents who switch from choosing Immigrant B at lower incomes to choosing Immigrant A (or both A and B) at higher incomes or who maintain constant choices. We report summary statistics of all types of switching behaviour and add robustness checks that include all observations. For multiple switchers between B and A, we analyse the results by including their first switch point and we check the robustness by excluding them from the sample, as in Exley (2016).
Other outcomes of interest include monetary allocations to Immigrants A and B in the allocation game, participants' choices regarding which immigrant receives information about immigration lawyers, qualitative responses justifying preferences between A and B in the hypothetical scenarios, distribution of immigration-related concerns identified by natives, and participants' perceptions of immigrant A's initial beliefs, current beliefs (control group) in addition to his initial and future integration status.