Thinking generically and specifically in International Relations survey experiments

Abstract

Does treatment abstraction affect treatment effects in International Relations survey experiments in countries outside of the US? We assess whether treatment effects are conditional on the anonymity of country actors among respondents in Brazil, China, Sweden, Japan, and Ukraine. We examine whether the effects of the United Nations’ approval of military force and regime type of the target country on support for war are moderated by respondents’ compliance with our abstraction encouragement. We find that around 20% of the respondents across all samples think of specific countries and do not comply with our abstraction encouragement. However, we fail to find evidence of a change in the average treatment effects by non-compliance, implying that the treatment effects are not likely to be conditional on respondents’ compliance (thinking of specific cases) or schema inconsistency (thinking of specific cases that are implausible given the context). At the same time, we find that treatment inconsistency (thinking of specific cases that are inconsistent with the assigned treatments) can affect the main treatment effects.