Research

Does exposure to conflict displacement amplify support for far-right parties? This study investigates the political consequences of conflict-induced displacement in northern Israel by exploiting a natural experiment arising from government-ordered evacuations during military conflicts. Using a pre-registered survey and a regression discontinuity design (RDD) based on an exogenous 5 km threshold, I assess changes in retrospective and prospective voting behavior as well as self-reported political ideology. The analysis reveals no evidence that displacement increases far-right support or induces partisan realignment. Rather, evacuees demonstrate a modest rightward shift in ideological self-placement without corresponding changes in voting behavior. This divergence suggests that displacement influences symbolic political positioning while preserving established partisan loyalties. These findings occur within a broader pattern of political frustration and disengagement—pointing to heightened disillusionment with the existing political establishment. These findings reveal troubling implications for democratic accountability during external conflict: leaders may be shielded from electoral consequences of security failures even as civic engagement deteriorates, challenging models of wartime accountability.

Do electoral outcomes influence public support for the winning candidate’s policy agenda? While voters are often assumed to evaluate candidates based on policy alignment, research suggests that partisan identity frequently overrides rational assessment. However, the extent to which the identity of the electoral winner legitimizes policy positions – particularly in the field of foreign policy – remains underexplored. This article leverages the 2024 U.S. presidential election as a natural experiment, using a pre-registered two-wave panel survey to assess whether electoral victory influences public perceptions of the winning candidate’s platform. Results reveal a moderate ‘winner effect’: Democratic voters exhibited an increase in support for Trump’s foreign policy positions post-election, while support for Harris declined, even among her base. Nevertheless, winner effects produced shifts in foreign policy preferences commensurate with those on a highly salient domestic issue, challenging assumptions about greater volatility in foreign policy. Furthermore, pre-election candidate support conditioned post-election stability, with winning-party supporters maintaining greater consistency. While electoral victories enhance the perceived legitimacy of the winner’s platform, partisan identity remains the primary constraint, limiting large-scale ideological realignment.

Why does political ideology shape citizens' political trust responses to some crises but not others? Democratic backsliding is often driven by internal political conflict, yet we know little about how such crises reshape citizens' trust in governing institutions compared with externally driven ones. We argue that the answer lies in the attributability of responsibility: the degree to which a crisis lets citizens trace the threat it activates to the deliberate choices of governing actors. Where responsibility is attributable, ideology strongly conditions how threat perceptions translate into trust; where a threat's external origin constrains blame, this conditioning weakens. We test this using two original surveys of Jewish Israeli citizens spanning two contrasting crises within a single polarized democracy: the 2023 judicial-overhaul crisis, widely characterized as democratic backsliding, and the war following the October 7, 2023 attacks. During the internally generated crisis, economic threat perceptions sharply eroded trust, especially among opposition-aligned left-wing citizens. During the security crisis, this ideological filtering disappeared, but trust did not rise, qualifying rally-around-the-flag accounts. Internally generated crises, we conclude, threaten democratic legitimacy more durably than external ones, precisely because they activate the divisions that make trust hardest to rebuild.

Does exposure to threat produce uniform conservative shift, or does it polarize ideological camps further apart? Two long-standing traditions in political psychology give opposing answers, and recent work has refined but not resolved the debate. We contribute on three fronts. We bring evidence from outside the standard U.S. two-party setting, examining a multiparty system with a sharply different ideological baseline. We disaggregate the non-conservative bloc, distinguishing centrist from leftist responses that the existing literature typically conflates. And we examine support for state digital surveillance, a security policy whose ideological structure is less crystallized than the military responses on which the literature has largely relied. Drawing on two original surveys fielded in Israel during the Israel-Hamas war (N = 2,199), we find that personal exposure to terrorism produces convergence at the political center across both waves, consistent with defensive-conservatism. The leftist response is conditional: positive in 2024 but absent in 2025, suggesting that worldview-protective dynamics become operative for the left when the threat no longer maps onto the policy. We argue that the two traditions are best understood not as competitors but as conditional accounts that apply to different ideological subgroups under different political contexts.

How does exposure to violence shape foreign policy attitudes, and what happens to these attitudes as the intensity of conflict recedes? We argue that conflict intensity changes what exposure means politically, shifting it from threat to cost as hostilities recede. We test this argument using two survey waves (N=1,260) fielded in Israel across distinct conflict-intensity environments: Wave 1, a high-intensity wave fielded during the acute Israel-Iran escalation, and Wave 2, was fielded during a lower-intensity interval between escalations. This design allows us to assess whether the political effects of violence exposure persist once hostilities subside or attenuate as the threat environment cools. We examine support for coercive, diplomatic, informational, and economic foreign policy instruments, and assess how personal exposure to violence, psychological distress, and political ideology shape attitudes across contexts. The two waves reveal a central implication: exposure is not a stable engine of hawkishness. At peak intensity, preferences saturate and ideology crowds out individual experience; as intensity recedes, exposure shifts from a threat signal to a cost signal, generating restraint rather than escalation across coercive instruments. This reframes how mass politics enters the war-to-peace transition: the publics most shaped by violence can become constraints on continued coercion precisely when conflicts cool, altering the domestic foundations of both escalation and compromise. Unlike dominant accounts that treat violence exposure as a stable driver of hawkishness, we show that its political imprint is conditional on conflict intensity: exposure is muted when war is acute but becomes a cost signal that attenuates coercive demand as the threat environment cools.

Does exposure to violence radicalize individuals, and through what mechanisms? Research often treats psychological distress as a general driver of wartime radicalization. We argue instead that endorsement of violence by others and personal willingness to engage in it follow distinct pathways. Exposure may directly increase endorsement, whereas willingness to act emerges through exposure-induced distress and is attenuated by cognitive flexibility. We test this argument using two-wave panel data collected before and during the June 2025 Israel--Iran war. Wartime exposure increased distress, which predicted greater personal willingness to engage in political violence but not greater endorsement of violence by others. Endorsement instead increased directly with exposure. The relationship between distress and willingness was concentrated among individuals lower in cognitive flexibility, reflecting an emotional response to exposure rather than heightened threat perception. These findings show that distress is not uniformly radicalizing: its consequences depend on the response considered and individuals' cognitive resources.

According to the literature, the IMF’s track-record in averting financial crises and promoting economic growth is mixed, and evidence suggests that IMF programs may increase poverty and income inequality, and have adverse and even gendered effects on unemployment, labour income and rights. IMF programs are also linked to deteriorating public health, educational outcomes, vaccination rates, child mortality, corruption, government instability and the likelihood of civil war. We replicate results from 508 models in 29 related articles in top journals (all such articles for which we could obtain replication files), and find that many of them effectively base their conclusions on a small set of countries or years, even when their nominal samples are large. To calculate this we develop indicators of the size of effective samples, which tell us if a particular estimate is based on the entire data fed into the regression, or rather on an effectively narrower subset of observations. A small effective sample hinders the ability to generalize the results to the entire nominal sample (low internal validity), and possibly also to a target population (low external validity) especially if the sample is representative of the population. These indicators, which are comparable across models and datasets, can be applied to a range of regression analyses and methods. We use these indicators to also demonstrate how scholars trade-off meticulousness (in both treatment operationalization and in causal identification) against generalizability. Our indicators can help scholars manage and optimize this trade-off.