How and why do democracies vary in their support for human rights resolutions on the international stage? Although democracy is often seen as the primary determinant of foreign-policy decisions on normative issues like human rights protection, empirical evidence is mixed: autocratic states such as China often show greater support for human rights resolutions than the United States in the United Nations General Assembly. We investigate this puzzle by applying a network mediation model that conceptualizes trade networks as mediators explicitly, hypothesizing that complex trade interdependence between states channel the effect of regime type on human rights voting decisions. The model estimates each state's latent position in the trade network using singular value decomposition and quantifies the indirect effect of regime type on voting behavior through economic ties. Our analysis reveals that the indirect impacts of regime type via trade interdependence is substantial and different from that of direct impacts. These findings challenge prevailing assumptions about normative alignment among democracies and highlight the constraining influence of trade networks on values-based foreign policy. Our work identifies previously neglected indirect effects of economic networks among states, strengthening the connection between theoretical expectations and empirical analysis through social network approaches.

While questions of war and peace have always been central in IR, a large body of literature in recent years has focused on “second image” explanations: the impact of regime type on preferences for war, trade and cooperation. Since many explanations carry implications for the individual level, IR scholars have fielded a large number of experiments designed to assess the impact of regime type treatments on preferences and beliefs about conflict and cooperation. We re-analyze all existing regime type experiments related to these topics and find strikingly consistent directional effects on preferences for war (lower for democracies) and cooperation (higher for democracies). Within these main results, we also find notable variation, and use our meta-analysis to assess explanations that bear on both the direction of the results and variation within them. We present results suggesting that the direction of the main effects of democracy on policy preferences can be plausibly explained by the consistent effects of regime type on a large number of potential “belief-based” mediators. Finally, we find that the significant variation within the results can be explained by a combination of temporal patterns—changing beliefs about democracy—as well where a study was fielded and aspects of study design (notably, the level of detail provided in experiments).

Conventional international security scholarship has long treated nuclear coercion as a strategic interaction among states, focusing on deterrence and crisis bargaining. Yet in democratic systems, nuclear threats are not received solely by leaders—they reverberate through mass publics, whose reactions can constrain or enable policy responses. Despite the centrality of public opinion to democratic accountability, we know remarkably little about how nuclear threats shape domestic discourse in real time or how these dynamics feed back into government behavior. This study addresses this gap by examining how Russian nuclear threats during the Russo-Ukrainian War influenced public discourse on social media and how Western democratic governments responded in parallel. Drawing on a large corpus of Twitter/X data spanning November 2021 to December 2022, we employ large language model (LLM) annotation to classify public discourse along three dimensions: preferences for escalation versus concession toward Russia, support for or against intervention in Ukraine, and evaluations of incumbent political leaders. We then construct a parallel timeline of Western government responses, including public statements, economic sanctions, and military aid to Ukraine, to assess the temporal relationship between shifts in public discourse and state behavior. Our findings contribute to scholarship on nuclear diplomacy and democratic accountability of war by providing the first large-scale empirical analysis of how nuclear threats propagate through public discourse and shape the political constraints facing democratic leaders in a live nuclear crisis.

How do democratic citizens react to foreign civilian casualties? Contrary to mainstream IR theories suggesting preferences for moral and humanitarian values in wartime, citizens often exhibit insensitivity to innocent foreign deaths. This raises a puzzle for both IR theory and US foreign policy: under what conditions do foreign civilian casualties reduce domestic public support for war? I argue that the public makes subjective judgments about the immorality of military actions, which shape policy preferences. In particular, perceived intent to harm civilians significantly reduces approval for military force, even when controlling for the magnitude of casualties and military benefits. Using randomized conjoint experiments, I find that the public reduces support when attacks on civilians are more intentional and when targeted victims are more vulnerable (e.g., women and children rather than unspecified civilians). Consistent with these findings, an additional text analysis of news media discourse indicates that the public's perception of a military's intent to harm civilians significantly reduces support for the operation.