Research

How much data is needed to ensure that a model performs reliably on new, unseen data? Despite its central importance to empirical research design, sample size decisions are often made heuristically—guided more by resource constraints than by principled diagnostics. Existing tools like power analysis and cross-validation offer limited insight into how predictive performance scales with sample size. We introduce a design-aware, empirical framework for estimating sample complexity bounds tailored to applied settings. By fitting smooth extrapolation functions to model performance from resampled pilot data, our method estimates the sample size needed to achieve researcher-specified generalization guarantees. Through applications to supervised learning tasks involving extensive human-annotated data, we show that generalization often stabilizes with as little as 10% of typical labeling costs. This approach provides a statistically grounded, interpretable diagnostic for generalization performance and a practical tool for political scientists designing data-intensive studies under resource constraints or design uncertainty.

In this paper, we argue for the importance of a previously overlooked factor—the existence of prior grievances over historical loss of territory—as a source of support for ethno-populist parties. While territorial loss is not itself a necessary condition for the emergence of populism, the issue lends itself unusually well to the backward-looking, loss-oriented framing of national victimization at the hands of elites that constitutes a key element of such parties’ electoral success. Drawing on original cross-national experimental and observational data from surveys conducted in Romania, Hungary, Germany, and Turkey in 2020–2021, we demonstrate that territorial loss attitudes are a remarkably robust predictor of support for ethno-populist parties, although important differences in national context emerge across cases. In addition, the panel structure of our data allows us to exploit a quasi-natural experiment in the form of the emergence of a new and highly successful populist party in Romania between waves, from which we conclude that loss attitudes are stable over time and temporally prior to support for populism.

If the anarchy of the international system obliges states to pursue a foreign policy aimed at increasing their economic and military power in order to ensure their own security, then the interests of a number of states in Antarctica require further explanation. Despite the debatable value of the Antarctic region from the perspective of a strategy of hard power, the authors argue that Russian interests in Antarctica can also be seen as part of a strategy of maximizing so-called “smart power.” Russia’s Antarctic policy aims to advance national interests while strengthening the nation’s internal and external image as a state committed to cooperatively resolving global problems. Russia’s recent significant expenditures in Antarctica and the Southern Ocean indicate the existence of a long-term foreign policy strategy on the part of the Russian leadership.

We propose a model of electoral competition where candidates offer excludable transfers to voters connected on a social network. We derive concentration results for weighted graphs to compare large societies, with social structure represented by asymptotic properties. We find that the diversion of public resources toward private provision generically favors the majority group and is least prevalent when groups are equipopulous and segregated. If candidates have heterogeneous information about the electorate, information can overcome the tyranny of the majority when groups are segregated or, given weak peer effects or an overall poor quality of information, candidate office motivations are low.

The construction of narratives that situate basic facts in interpretable relations is one of the core functions of political speech, yet measurement difficulties have hampered systematic research in political science. Despite an explosion of interest in automatic text analysis, bag-of-words assumptions discard important relational information, while large language models suffer from a lack of interpretability and reproducibility. In this paper, I introduce a random network model of political narratives based on a representation of text as subject-verb-object interactions. I validate the model’s ability to uncover novel aspects of political speech across multiple languages using two applications: partisanship in U.S. congressional speeches and an event study of a censorship episode in Russia.

A growing literature demonstrates the potential of violence that occurred more than a lifetime ago to exert powerful influence on political behavior in the present. Yet little is known about what conditions determine when legacies will manifest or how they are transmitted in different contexts. This paper presents findings from a survey conducted in Southern Armenia among 1,200 individuals in settlements with varying levels of exposure to past violence. I estimate the effect of past experiences of repression during the Armenian Genocide and Stalinist repressions on contemporary participation and political preferences. Contrary to other recent work, the analysis reveals that direct exposure to repression does not significantly influence political behavior. Instead, the legacy of violence in Armenia operates almost exclusively through the channel of social transmission. Moreover, the findings underscore the pivotal role of narrative storytelling in shaping these connections, providing a deeper understanding of the mechanisms through which historical memory and collective experiences affect contemporary political dynamics. The study contributes to the literature on the interplay of history, social networks, and political behavior in post-authoritarian societies.

This study considers the impact of individual attitudes toward both historical and recent territorial losses on political behavior. It explores the connection between individual attitudes toward lost territories and their influence on incumbent support and participation in anti-government protests. Using data from an original survey conducted in Armenia, the paper estimates the causal impact of individual concern over lost territory, leveraging exogenous variation induced by exposure to displaced persons and the visibility of Mount Ararat. The analysis shows that those valuing lost territory more are prone to withdrawing government support, emphasizing candidate traits related to symbolic compliance, and engaging in risky protests. This effect is mediated by emotional distress related to territorial losses. Notably, social network position, rather than media consumption or political partisanship, drives these effects, highlighting a potent grassroots check on political elites in nascent democracies. These findings extend the existing understanding of irredentism, uncovering the role of public attitudes in contexts beyond interstate conflicts. They also deepen insights into the legacies of political violence, revealing how present contexts shape interpretation of historical collective trauma. Lastly, the study enriches knowledge about nationalism and populism in emerging democracies, spotlighting how divergent narrative beliefs about the nation can affect behavior even in a context of universally high nationalism.

Given the powerful association between territory and the nation, losses of territorial integrity often achieve a high degree of political salience, even when there exists no realistic path toward reversing the loss. Remarkably little is known, however, about how the prominence of the territorial issue structures domestic politics. Drawing on experimental evidence from Romania, this paper provides micro-level evidence that elite narratives emphasizing past territorial loss shape the political behavior of ordinary citizens through their effect on ambiguity attitudes. When the reference point is shifted to an imagined past when the nation was physically whole, even policy choices with positive outcomes are felt as relative losses, increasing willingness to opt for unknown quantities, such as expensive prestige projects or outsider parties.