A Survey Experiment on Creedal Nationalist Persuasion
A persuasion experiment on creedal nationalism, fielded in the United States, Germany, and South Korea
Description
This registration covers a persuasion experiment on creedal nationalism, fielded as the final block of a survey instrument in the United States, Germany, and South Korea. The treatment is an unattributed passage presented as something often said to define national identity; the control is a minimal-exposure transition screen. Primary outcomes are two pre-registered composites from ISSP National Identity items: a pride composite and a boundary composite. The registration specifies a main hypothesis for each country, a guardrail against triggering feelings of national superiority, and a test in Korea for effects in the opposite direction. About 2,500 people in each country are randomly assigned to one of two groups with equal probability.
Study design
This is a between-subjects, two-arm persuasion experiment nested as the final block of a shared four-country survey instrument. Respondents are assigned with equal probability to a treatment arm or a minimal-exposure control arm. The pride composite (six items) replicates Bonikowski and DiMaggio's (2016) operationalization: democracy, political influence in the world, economic achievements, achievements in sports, achievements in the arts and literature, and history. The boundary composite draws on five ISSP items.
Within the treatment group, respondents are also randomly assigned to different versions of the message, tied to country-specific civic and constitutional rhetoric (Franklin Roosevelt, Ronald Reagan, Barack Obama, and George W. Bush in the United States; the Basic Law, Johannes Rau, and Joachim Gauck and Frank-Walter Steinmeier in Germany; the Constitution's founding text, Kim Dae-jung, and the Constitution's liberal-democratic language in South Korea). These versions are used for secondary, exploratory comparisons and are not part of the main confirmatory test. Respondents are recruited from online panels administered through Qualtrics with international survey vendors; Singapore does not field this experiment.
Hypotheses
H1-US. Among American respondents, the treatment raises the pride composite.
H1-DE. Among German respondents, the treatment raises the pride composite.
H1-KR. Among Korean respondents, the treatment lowers the boundary composite. The Korean test differs because pride there is already high, leaving little room to move, so the test instead looks at the boundary measure.
Registered guardrail. The treatment should not raise national superiority. A country's H1 result counts as support for creedal persuasion only when this guardrail passes.
Korean wrong-direction test. A separate, one-sided test checks whether the treatment instead raises the Korean boundary composite. A significant increase would count against the creedal appeal in Korea regardless of the other results.
Analysis plan
H1-US and H1-DE are tested as two-sided t-tests of the pride composite on arm; H1-KR is tested the same way on the boundary composite. Each country test is estimated independently, with no correction applied across the three countries. The registered guardrail combines an equivalence test with significance screens on a national-superiority measure, evaluated jointly. One pre-specified sensitivity analysis reruns the registered primary tests, the guardrail, and the Korean wrong-direction test on the subset of respondents not flagged by a composite automation indicator.