Do AI-generated video-treatments offer methodological gains over traditional text-based vignettes? To address this question, we conducted an experiment analyzing the effectiveness of an AI-generated video, long text with AI-generated images, plain long text, and a short-text vignette. Our results found that the AI-generated video did not elicit the largest shifts in policy preferences. Its effects were statistically indistinguishable from long-form text and even marginally weaker than the text-with-images treatment. The findings suggest that we should be cautious about the promise of AI-generated videos for experiments. Find the full pre-print here.

A mixed-methods study of British parliamentarians and the public finds elites more willing to defend NATO allies than the public. Counter-attitudinal public cues modestly shift elite preferences on defending allies. Qualitative evidence reveals heterogeneous elite approaches to public opinion, which becomes consequential when commitments are costly, sustained, morally salient or linked to mobilized constituencies. Find our pre-print here.

Our large survey-based longitudinal study investigates changes in perceptions of NATO’s collective defense commitments across Western allies, Russia as well as British and German parliaments. The study finds a dramatic decline in how the public and elites perceive the credibility of U.S. commitments to NATO during the first year of Trump’s second administration. Find our pre-print here.