I investigate the impact of automation exposure on political behavior in post-industrial societies, with a specific focus on the support for populism. I examine the potential causal mechanisms by exploring the interplay between economic and cultural factors. Through a parallel encouragement design survey experiment conducted in the US, I divide the sample into two groups: one group is randomly assigned to the treatment condition either related to exposure to robots and AI replacing jobs or a control condition related to technological development; the other group experiences manipulation of both the treatment and the encouragement of mediators (marginalization and nostalgia). My findings reveal that feelings of marginalization and nostalgia mediate the effects of technological change on support for populism and illiberal policies. To enhance external validity, I implement mediation analysis using survey data from the European Social Survey from 2012 to 2016. The results help us understand how structural changes in labor markets and cultural factors impact political behavior and inequality.

This study investigates the impact of exposure to foreign goods in the marketplace on the policy preferences and political behavior of US consumers. Using a survey experiment, we simulate a realistic consumption experience with well-known brands of sports utility vehicles. Our findings reveal that exposure to foreign brands intensifies hostility towards immigrants and trade among respondents holding pre-existing nationalist attitudes while also increasing their support for Trump as a presidential candidate. Exposure to foreign brands has the opposite effect on the preferences and behavior of cosmopolitans. Our results demonstrate that consumption in an increasingly diversified marketplace can drive a bottom-up process of trade policy polarization. Our study has significant implications for understanding the contemporary political backlash against economic globalization and standard models of international trade based on the ``love of variety."

How does an immigration shock affect politicians’ discourses? This study examines the sudden influx of Venezuelan migrants into Latin American countries. We argue that such events alter politicians’ agendas, creating opportunities to frame new issues from their perspectives. Analyzing over 3 million tweets by parliament members from 2013 to 2021 in Chile and Peru, we employ computational textanalysis methods, from simple dictionaries to complex techniques like unsupervised topic analysis and OpenAI, along with an instrumental variable strategy. Our results suggest that after the immigration shock, politicians emphasized the immigration issue without any party family monopolizing it. We find little evidence that regional exposure explains the issue’s salience, suggesting a disconnection from local experiences. Our findings reveal a novel channel for increased salience: right-wing politicians criticized the Venezuelan regime and socialism instead of increasing anti-immigration sentiment, while left-wing politicians promoted pro-immigration attitudes consistent with contact theory. This work enhances our understanding of the politicization of immigration in South-South contexts in the digital age.

Extreme wealth has long characterized Latin American democracies, but recently billionaires have become more visible and politically legible targets of taxation. Yet, we know less about how this change translates into coalitional politics among workers divided by labor-market position. We develop a theoretical model in which super-rich visibility raises the perceived fiscal returns from taxing the top and makes divisions within labor appear smaller relative to those at the top. We evaluate its implications using news coverage and documented episodes of mobilization and policy action across Latin America, together with an original Mexican survey pairing closed-ended preferences with open-ended narratives. The evidence documents growing articulation of top-end taxation and shows that support for a targeted top tax extends across partisan groups and is less segmented by labor-market position than support for redistribution that leaves the payer unspecified. Shared fairness claims provide a basis for coordination, while distrust in enforcement and government spending constrains coalition formation.

Abstract source

In an era of increasing economic precarity and labour market polarization, meaningful efforts to mitigate climate challenges face a fundamental political challenge. We examine how indviduals' long-term labour market risk shapes their environmental attitudes and support for green policies. We argue that long-term labour market risk is expected to reduce environmental concern amongst those affected due to a deprioritization of problems with high levels of uncertainty and that require deep reforms to be addressed. Therefore, we expect labour market risk to subsequently reduce support of environmental policy that imposes immediate direct costs, such as carbon taxation. Using European Social Survey data from 2002 to 2018 and several waves of the International Social Survey Programme across European countries, our analysis reveals that individuals' facing long-term labour market risks are less likely to hold environmental concerns and less supportive of carbon taxes that impose immediate visible costs. Our findings have important implications for understanding how structural transformations in the economy shape individuals' preferences for tackling long-term societal problems like climate change.

Graduate school pathway programs in the social sciences face intensifying political attacks, legal constraints, and defunding. This viewpoint examines strategies for sustaining these programs through strategic reframing, diversified funding, collective knowledge-sharing, and risk-balanced action. It argues that program survival is essential for preserving equity infrastructure and broadening participation in research careers.

While there is little doubt that technological change is generating labor market polarization around the world, we know much less about its translation into partisan polarization. I explore the political polarization driven by the rise of right-wing populist parties and leaders throughout developed democracies. I build a theoretical model to explain how right-wing populists have attracted the votes of routine workers, workers exposed to automation risk, and previously loyal to mainstream left-wing parties, within both majoritarian multi-district and multiparty proportional systems. I empirically evaluate the theory, focusing primarily on the US and Germany, using individual vote-switching data and campaign targeting strategies inferred from the content of political speeches and party manifestos.

Students from underrepresented populations, including women, first-generation students, and racial and ethnic minorities, face persistent barriers in accessing graduate education and professional development opportunities, particularly in the social sciences. To address these disparities, graduate pipeline programs can provide critical support by providing networking and mentoring opportunities, closing methodological and research skill gaps, and preparing participants to apply to competitive graduate programs. This article outlines the principles and goals of a summer program focused on mobilization and political economy by detailing the rationale, implementation strategies, and evaluation framework used to assess its effects. Evaluation of the inaugural program demonstrated the success of a multilayered mentoring approach, tailored curricula, and comprehensive support structures in fostering participants’ academic growth and career readiness. By highlighting best practices and lessons, we provide a template for institutions seeking to develop initiatives associated with greater diversity in graduate programs in the social sciences, the academic market, and, ultimately, higher education.

The political science discipline faces significant disparities in the representation and participation of underrepresented minorities in graduate education. This lack of diversity among political scientists results in a narrower range of questions being explored within the field. Furthermore, the underrepresentation is particularly pronounced in the political methodology subfield, limiting the scope of tools and perspectives to those predominantly shaped by white scholars. This article proposes a template for teaching and mentoring undergraduate students from underrepresented backgrounds to enhance their opportunities in graduate programs. We examine the Mobilization and Political Economy program, a summer research initiative aimed at minority-serving institutions and underrepresented minorities, designed to equip participants with the tools to study social movements, political mobilization, and structural inequality.

Do political protests spread across time and space? While scholars of social movements and political behavior have long debated this question, existing studies often fail to simultaneously account for both spatial and temporal dependencies in protest dynamics. Using protest event analysis and a novel spatiotemporal autoregressive distributed lag (STADL) model, we examine the diffusion of protests across 30 European countries from 2000 to 2015. Our findings provide robust evidence that protests exhibit both temporal and spatial contagion: protest activity in one year significantly increases protest frequency in the following year, and protests in one country contribute to the onset of protests in neighboring states. These results underscore the importance of modeling both dimensions of diffusion to avoid biased inferences and contribute to the broader understanding of protest mobilization. Our study highlights the interconnected nature of political activism and has important implications for research on social movements and political instability.

This paper investigates the impact of the fourth industrial revolution on politics by proposing a theoretical framework linking technological change with political apathy. Using hierarchical logistic modeling with varying intercepts by country and survey data from the European Social Survey from 2002 to 2018 for 23 European countries, I present evidence that individuals more exposed to technological change are less likely to feel close to a political party, participate in elections and take part in protests. Those individuals exposed to automation are about 10% less likely to be politically engaged than those respondents without exposure to automation risks. I also demonstrate that income levels and unionization rates substantially moderate the direct link between automation and political engagement. The impact of automation on political engagement is smaller among wealthier citizens and in highly unionized environment. The political message from these interaction effects speaks about the reinforcing forces between economic inequality and automation and the role of collective organization. My findings have important implications for understanding automation politics, political inequality, and the demand (or lack of) for protection.

This paper explores how legislators use social media, specifically investigating whether their posts reflect the concerns expressed by their legislative party peers in an anonymous survey. Utilizing data from Twitter (now X), we compare legislators' social media posts with their responses in a survey of legislators in Latin America.

We propose a novel and scalable method for analyzing political communications, employing OpenAI for topic identification in statements and BERTopic analysis to identify clusters of political communication. This approach enables a thorough and detailed examination of these topics over time and across political parties.

Applying our method to statements made by members of the Chilean Congress, we observe a general alignment between the preferences stated in surveys by elites and the prominence of these issues on Twitter. This result validates social media platforms (particularly Twitter) as a tool for predicting politicians' preferences. Our methodological approach offers a scalable tool for analyzing political rhetoric over time.

Abstract source

Many legislators do not have consistently progressive or conservative policy positions. How does the mix of issue positions relate to the manner in which the legislators consider their placement on the left-right ideological scale? Analyzing data from the Parliamentary Elites in Latin America (PELA) survey, this paper counterposes combinations of legislators’ moral and economic policy positions with their self-located ideological score. Our results confirm the importance of economics, which is consistent with older studies, but we also find that moral issues are at least consistent with – and perhaps germane– to the left-right placement of many of the region’s legislators. Among the findings are that the left is more heterogeneous, especially with respect to moral views, than is the right. We also show that many centrists are closeted conservatives, supporting the “ashamed right” thesis.

Same-sex marriage (SSM) has risen to the top of political agendas across Latin America, but there is also great variance in terms of legal status, public support, and the policymaking processes. While the public and social movements have been critical to the advance of SSM, we know little about the views of those who are directly charged with translating public views into policy: the legislators. To fill this gap, we utilize a survey of the region’s legislators to first examine the range in support among countries and show how it correlates with legal changes. We then examine the correlates of legislators’ support for SSM. While we also test gender, age, and ideology, our multivariate models focus on religiosity. We show that in addition to driving support at the individual level (in the expected direction), religiosity also works as a contextual variable such that having more secular colleagues encourages pious legislators to support same-sex marriage.

El matrimonio entre personas del mismo sexo (MPMS) ha marcado la agenda política en muchos países de América Latina, aunque aún es ilegal en muchos países del continente. No obstante, el apoyo público varía mucho en la región, así como también los roles de los tribunales, presidentes y legislaturas. En este artículo nos enfocamos en los legisladores, ya que son los encargados de representar al público y convertir sus demandas en política pública. Si bien muchas legislaturas han discutido el tema, la literatura no ha examinado de manera intensiva las actitudes de estos representantes hacia el MPMS. Para analizar este fenómeno aplicamos un marco teórico que amplía las teorías basadas en el contexto y contacto social, y utilizamos una encuesta implementada a legisladores en la región para estudiar las variables que correlacionan con el apoyo al MPMS. Si bien también evaluamos variables a nivel individual (tales como género e ideología), nuestros modelos se enfocan en el rol contextual de la religiosidad. Los resultados muestran que tener más colegas seculares alienta a los legisladores, incluso a los creyentes, a apoyar el matrimonio entre personas del mismo sexo.

We examine the political and institutional consequences of artificial intelligence’s economic impact. Although AI may affect nonroutine jobs in particular, we show that current models about the vulnerability level of occupations and economic sectors differ widely in their forecasts. This may explain why public opinion is not settled about the effects of AI. While many fear AI may displace them from their jobs, a majority seems optimistic about its overall impact. Responses vary, however, by cohort and depending on survey framing. AI’s current training and computing needs have magnified capital concentration and business investment in fixed assets, intensifying the technological sector’s interest in regulatory capture. Taken together, AI’s effects on labor and capital may strain democracy unless a set of policies we outline here are gradually implemented.

Abstract source

How can social scientists use agentic AI to produce not more, but better work? We argue that the answer lies in using agentic AI to accelerate iterative search by generating alternatives, testing them, and failing more productively. Drawing on Jonathan Bendor’s framework of innovation as a search problem that balances creativity (broad exploration) and criticism (rigorous evaluation), we organize our chapter around two recurring search failures: creativity failure, in which promising ideas are screened out before they are adequately explored, and criticism failure, in which researchers invest too much in weak research designs. Within that framework, we then discuss two applications. One focuses on organizational-level measurement research, where agentic AI can lower the feasibility barrier to exploring new projects and automate iterative validation, thereby avoiding creativity failure and catching flawed measures before they enter the downstream workflow. Another focus is on experimental research, where the prevalent error might be criticism failure arising from underpowered, under-piloted designs. In this domain, agentic AI enables stress-testing of designs and the implementation of construct validity checks at scale. Though AI agents can be potentially transformative for empirical social science research, we also see key limitations, such as inferences from silicon samples not generalizing to real humans; LLMs defaulting to agreement; AI-generated stimuli carrying biased and multi-barreled treatments; and new problems created by using AI to defend against AI contamination. Agentic AI can be used to improve design, measurement, and validation, but it cannot replace human judgment.

Abstract source

Same-sex marriage has risen to the top of political agendas across Latin America, but it is still illegal in many countries. Public support about the issue varies greatly, and the roles of the courts, presidents, and legislatures have also differed. This article focuses on legislators because they are charged with representing the public and converting demands into policy. Although many legislatures have now voted on the issue, the literature has not intensively examined the policy makers’ attitudes toward same-sex marriage. This study applies a theoretical framework that extends theories considering context and social contact and uses a survey of the region’s legislators to study the correlates of support for same-sex marriage. Although the study also tests for individual-level variables (e.g., gender and ideology), the models focus on the contextual role of religiosity. The results show that having more secular colleagues encourages even pious legislators to support same-sex marriage.

Abstract source

How do individuals perceive technological risk, particularly job insecurity, in the era of AI-driven change? This paper explores how people navigate the uncertainty created by the pace of technology and experts' unclear predictions about its impact on jobs. We focus on three perceptions: techno-optimists, techno-pessimists, and those exposed to rapid technological change, examining their demographic characteristics and implications for political coalition-building. Using data from three original surveys, including open-ended responses, we show that personal and vicarious experiences shape these perceptions, with exposed workers occupying a middle ground between optimism and pessimism. Contrary to the view that high-skilled workers are less vulnerable, we find that perceived employment risks rise with the use of complex technologies like programming languages. Exposed workers and pessimists share political traits, such as support for illiberal policies, while optimists lean toward liberal policies. These results deepen our understanding of how technological risk perceptions influence politics.

Generative AI creates a measurement problem for survey research. When respondents can outsource summarization, explanation, or fact retrieval, open-ended answers may no longer reflect their own reasoning. We introduce an auditable toolkit that combines response-process paradata with prompt-specific semantic benchmarks to detect AI use in open-ended responses. We validate the approach in a randomized survey experiment in which participants complete a summary task under either blocked AI access or observed access to an embedded AI tool, and by administering the same instrument to a synthetic AI agent. AI-assisted and automated responses leave distinct behavioral and textual traces: they show reduced drafting and revision, greater similarity to AI benchmark output, and lower distinctiveness relative to peer responses. Classifiers built on these signals achieve strong out-of-sample performance. These findings show that, in the age of generative AI, authorship must be measured rather than assumed.

The rise of populism across democracies is one of the greatest challenges to the existing world order since World War II. Yet, we only have very limited tools to measure populism within and across countries. Existing approaches rely on human coding, expert surveys, or dictionary-based analysis and suffer from high costs, low comparability, or low validity. We offer a middle-ground strategy between expensive qualitative reading and coarse machine coding. We extend PULSAR to identify sentences/paragraphs carrying populist narratives by parsing “us” versus “them” frames, who is protecting/threatening “the people,” and other common aspects of political speech. We illustrate the approach with US presidential campaign speeches.

Automation is typically studied as a domestic labor-market shock in advanced economies. This essay argues that, for much of the Global South, it is better understood as a cross-border reorganization of production. Firms deciding where and how to produce compare not only labor and machines within a country, but also offshore labor, migrant labor, and capital-intensive production across jurisdictions. When robotics and AI shift those comparisons in the North, they also reshape trade, reshoring, and political adjustment in economies that have adopted relatively little automation themselves. I call this exposure external automation. The essay makes four claims: foreign automation often matters more than domestic adoption in labor-abundant economies; scholarship on trade and automation still tends to separate technology from trade; adjustment in the South is likely to appear through deindustrialization and weaker labor voice; and automation can widen North-South inequality by concentrating rents, productive capacity, and strategic control in the North.

Abstract source

AI and other digital technologies are reshaping work, yet their effects on labor-market institutions are still not well understood. Research on technology and unions has focused largely on robots, leaving open how newer digital tools influence worker organizing and collective bargaining. We argue that technology matters not only because it can replace labor, but because it changes how work is organized. We distinguish among displacement, augmentation, and monitoring technologies, and show that each type affects workplace interaction, job stability, and managerial control in different ways. These changes, in turn, shape both workers demand for collective protection and their ability to organize. They also create different bargaining priorities, including training, shared implementation, limits on surveillance, and job security. We test these ideas with a two-level design. First, using European Social Survey data from 15 countries between 2012 and 2024, we match occupations to emerging technologies and assess links to union membership, working conditions and political attitudes. Second, we analyze over 40,000 Canadian collective agreements from 1993 to 2025 with text analysis and an LLM-based AI exposure measure. Results show varied effects: some technologies raise unionization, while others weaken it and worsen worker outcomes significantly for many employees.

Abstract source

Recent advances in AI have fueled concerns about future job displacement, prompting research on the political consequences for workers vulnerable to job loss. Yet AI is already deeply embedded in the workplace, shaping the daily experiences of millions who interact with AI software regularly. Large language models (LLMs) are now widely accessible and affordable, enabling workers to adopt them independently. The benefits are immediate: greater efficiency, reduced cognitive effort, and often higher-quality output. But these short-term gains may entail long-term costs. As workers grow dependent on AI, they risk losing essential skills and autonomy. Their use also generates data that improves these tools, inadvertently accelerating automation that could eventually replace them. Do workers recognize these risks, and how do they balance them against short-term benefits in their responses to these changes? We explore these questions using a pre-registered field experiment on an online labor marketplace. We recruited 1,200 workers to perform various tasks, randomizing their (1) access to an AI-assisted tool, (2) the perceived capability of this tool, and (3) the stated long-term objectives of implementing this tool. Using a range of behavioral and attitudinal measures, we track changes in how workers use the AI tool and complete their work over time, as well as their perceived competence, job security, and preferences for policies such as retraining and compensation. By documenting informal resistance strategies through which workers seek to preserve their comparative advantage, our study offers new insight into the political consequences of this technological transformation.

Electoral campaigns are high-frequency environments in which parties receive rapid feedback and face immediate incentives to adjust both what they say and how they say it. Yet most research on party competition relies on low-frequency texts, such as manifestos, or on election-to-election change, obscuring strategic adaptation within campaigns. We address this gap by analyzing more than 8,000 YouTube transcripts produced by political parties in Spain and the United Kingdom during the year before elections between 2015 and 2024. Using embeddings, BERT-based measures, dictionaries, and large language models, we develop a general framework for studying campaign communication with social media data. We show that parties adapt ideological positioning as elections approach, shift issue emphasis, and respond selectively to rivals, especially on particular issues and within ideological families, rather than converging uniformly. More broadly, the approach captures movement not only in topics, but also in positions, framing, and rhetorical style.

How does automation in the Global North shape politics and violence in the Global South? We develop a political economy theory in which robot adoption in advanced economies reduces demand for export-oriented labor in developing countries, depressing wages and employment and creating social and political consequences. We test this argument in Mexico, a close trade partner of the United States. Using commuting-zone data, we construct exposure measures combining pre-NAFTA export employment with U.S. industry robot growth and initial offshoring intensity, while accounting for domestic robot adoption and other shocks. To address endogeneity, we instrument foreign exposure with European robot diffusion. We find that regions more exposed to foreign robots experience higher levels of violent organized crime, including narcocrime and homicides (but not property crime), and stronger support for left-populist candidates. These findings demonstrate how automation shocks ripple through global value chains to reshape society and elections.

This article asks whether automation can close migration as an adjustment mechanism. In the Mexico–U.S. corridor, U.S. robot adoption can reduce mobility through two channels at once: it can lower earnings in Mexican origins by reshoring production and reducing demand for Mexican inputs, and it can weaken labor demand in U.S. destinations. The empirical strategy combines the foreign-robot commuting-zone design developed by Faber (2020) with an extension of the commuting-zone robot design in Acemoglu and Restrepo (2020) to immigrant population outcomes. Mexican commuting zones more exposed to foreign robots have fewer emigrant households and lower emigration intensity, while U.S. commuting zones more exposed to domestic robots experienced smaller increases in Mexico-born population shares between 1990 and 2006–2008. The evidence indicates that automation can narrow exit by reducing realized emigration from origins and migrant absorption in destinations.

Research on the political consequences of economic inequality focuses almost exclusively on relative inequality, using measures such as percentile ratios and gini coefficients. We show with a simple theoretical model that proportionate income increases that preserve relative inequality can increase absolute inequality and generate larger gaps in campaign contributions between rich and poor. Using U.S. congressional district data, we find absolute inequality, rather than relative inequality, is associated with larger contribution gaps.

The current sociopolitical context features growing pressure on higher education inclusion efforts, alongside persistent underrepresentation in the social and behavioral sciences across gender, racialized, ethnic, and socioeconomic lines. Despite decades of initiatives and evidence on the value of diversity, pathway programs that provide mentoring, research experience, and professional socialization to underrepresented student populations often operate in isolation, with uneven funding and limited shared evaluation infrastructure. This Perspective article develops a network model for coordinating pathway programs within and across countries. Drawing on existing consortia and mobility infrastructures, we outline a networked-improvement-community approach that can pool resources, democratize collaborations across institutional types, and support research-informed program design and assessment through a shared Knowledge and Innovation Hub. We use the U.S. landscape as an illustrative case given its program density and policy volatility, while emphasizing design features needed for transnational operation, including shared governance, diversified funding, multilingual access, and context-specific definitions of underrepresentation. By linking existing efforts into a durable alliance, a networked approach aims to expand inclusive pathways into graduate training and research careers in the social sciences and beyond, while increasing resilience in periods of political and fiscal instability.