Fine-tuned LLMs often exhibit unexpected behavior as a result of generalizing beyond the data they're shown. We present results in which an LLM fine-tuned to prefer either coastal sports teams or Southern sports teams adopt political beliefs that diverge significantly from those of the base model. While we hypothesized that the coastal model would become more liberal and the southern model would become more conservative, we find that their responses are usually similar to each other, without a clear-cut liberal or conservative bias. In addition to asking the models for numerical ratings of agreement with relevant political statements, we ask them to elaborate on their more radical answers, finding varying degrees of willingness to justify themselves. Further work is needed to understand the mechanisms by which fine-tuning on simple, narrow datasets leads to seemingly unrelated changes in model behavior.
Data have power. As such, most discussions of data presume that records should mirror some idealized ground truth. Deviations are viewed as failure. Drawing on two ethnographic studies of state data-making in a Chinese street-level bureaucrat agency and at the US Census Bureau we show how seemingly "fake" state data perform institutional work. We map four moments in which actors negotiate between representational accuracy and organizational imperatives: creation, correction, collusion, and augmentation. Bureaucrats routinely privilege what data do over what they represent, creating fictions that serve civil servants' self-interest and enable constrained administrations. We argue that "fakeness" of state data is relational (context dependent), processual (emerging through workflows), and performative (brought into being through labeling and practice). We urge practitioners to center fitness-for-purpose in assessments of data and contextual governance. Rather than chasing impossible representational accuracy, sociotechnical systems should render the politics of useful fictions visible, contestable, and accountable.
Stefano Civelli, Pietro Bernardelle, Frank Mols
et al.
The spread of targeted advertising on social media platforms has revolutionized political marketing strategies. Monitoring these digital campaigns is essential for maintaining transparency and accountability in democratic processes. Leveraging Meta's Ad Library, we analyze political advertising on Facebook and Instagram during the 2022 Australian federal election campaign. We investigate temporal, demographic, and geographical patterns in the advertising strategies of major Australian political actors to establish an empirical evidence base, and interpret these findings through the lens of Social Identity Theory (SIT). Our findings not only reveal significant disparities in spending and reach among parties, but also in persuasion strategies being deployed in targeted online campaigns. We observe a marked increase in advertising activity as the election approached, peaking just before the mandated media blackout period. Demographic analysis shows distinct targeting strategies, with parties focusing more on younger demographics and exhibiting gender-based differences in ad impressions. Regional distribution of ads largely mirrored population densities, with some parties employing more targeted approaches in specific states. Moreover, we found that parties emphasized different themes aligned with their ideologies-major parties focused on party names and opponents, while smaller parties emphasized issue-specific messages. Drawing on SIT, we interpret these findings within Australia's compulsory voting context, suggesting that parties employed distinct persuasion strategies. With turnout guaranteed, major parties focused on reinforcing partisan identities to prevent voter defection, while smaller parties cultivated issue-based identities to capture the support of disaffected voters who are obligated to participate.
This article examines the complex relationship between money and political legitimacy in democracies (United States, Germany, India) and nondemocracies (China, Russia), using published empirical evidence to explore how financial resources influence governance. In democracies, US campaign finance, German party funding, and Indias electoral bonds amplify elite influence, openly eroding public trust by skewing policy toward wealthy interests. In nondemocracies, Chinas state enterprise patronage and Russias oligarch suppression strengthen legitimacy, yet hide vulnerabilities revealed by anticorruption campaigns and power struggles. The analysis argues that moneys corrosive impact is widespread but varies: democracies face evident legitimacy crises, while nondemocracies conceal underlying fragility. These findings highlight the need for reforms: increased transparency in democracies and wider power bases in nondemocracies, to mitigate moneys distorting effect on political authority.
News sharing on digital platforms shapes the digital spaces millions of users navigate. Trace data from these platforms also enables researchers to study online news circulation. In this context, research on the types of news shared by users of differential political leaning has received considerable attention. We argue that most existing approaches (i) rely on an overly simplified measurement of political leaning, (ii) consider only the outlet level in their analyses, and/or (iii) study news circulation among partisans by making ex-ante distinctions between partisan and non-partisan news. In this methodological contribution, we introduce a research pipeline that allows a systematic mapping of news sharing both with respect to source and content. As a proof of concept, we demonstrate insights that otherwise remain unnoticed: Diversification of news sharing along the second political dimension; topic-dependent sharing of outlets; some outlets catering different items to different audiences.
The applications of Large Language Models (LLMs) in political science are rapidly expanding. This paper demonstrates how LLMs, when augmented with predefined functions and specialized tools, can serve as dynamic agents capable of streamlining tasks such as data collection, preprocessing, and analysis. Central to this approach is agentic retrieval-augmented generation (Agentic RAG), which equips LLMs with action-calling capabilities for interaction with external knowledge bases. Beyond information retrieval, LLM agents may incorporate modular tools for tasks like document summarization, transcript coding, qualitative variable classification, and statistical modeling. To demonstrate the potential of this approach, we introduce CongressRA, an LLM agent designed to support scholars studying the U.S. Congress. Through this example, we highlight how LLM agents can reduce the costs of replicating, testing, and extending empirical research using the domain-specific data that drives the study of political institutions.
The 2008 global financial crisis marked the beginning of a decade dominated by fiscal austerity policies in much of the developed world. This paper presents a qualitative narrative review of an extensive collection of academic literature to synthesize evidence on the multifaceted effects of austerity. Following a thematic approach inspired by PRISMA guidelines, the economic, social, and political consequences of these measures are examined. The analysis reveals a majority consensus regarding the recessive effects of austerity, especially when implemented during economic crises, with negative fiscal multipliers that often exacerbate GDP contraction. Socially, austerity is associated with rising inequality, negative impacts on public health, disproportionate gender consequences, and a weakening of social safety nets. Politically, evidence links austerity to the erosion of trust in institutions, a rise in populism, and electoral instability. Despite the political narrative presenting austerity as an inevitable necessity for fiscal sustainability, academic literature underscores its high costs and questionable efficacy, advocating for more contextualized and equitable economic policy approaches.
Fabian Haak, Björn Engelmann, Christin Katharina Kreutz
et al.
Search query suggestions affect users' interactions with search engines, which then influences the information they encounter. Thus, bias in search query suggestions can lead to exposure to biased search results and can impact opinion formation. This is especially critical in the political domain. Detecting and quantifying bias in web search engines is difficult due to its topic dependency, complexity, and subjectivity. The lack of context and phrasality of query suggestions emphasizes this problem. In a multi-step approach, we combine the benefits of large language models, pairwise comparison, and Elo-based scoring to identify and quantify bias in English search query suggestions. We apply our approach to the U.S. political news domain and compare bias in Google and Bing.
We propose a framework for analyzing discourse by combining two interdependent concepts from sociolinguistic theory: face acts and politeness. While politeness has robust existing tools and data, face acts are less resourced. We introduce a new corpus created by annotating Wikipedia talk pages with face acts and we use this to train a face act tagger. We then employ our framework to study how face and politeness interact with gender and power in discussions between Wikipedia editors. Among other findings, we observe that female Wikipedians are not only more polite, which is consistent with prior studies, but that this difference corresponds with significantly more language directed at humbling aspects of their own face. Interestingly, the distinction nearly vanishes once limiting to editors with administrative power.
Numerous politicians use social media platforms, particularly X, to engage with their constituents. This interaction allows constituents to pose questions and offer feedback but also exposes politicians to a barrage of hostile responses, especially given the anonymity afforded by social media. They are typically targeted in relation to their governmental role, but the comments also tend to attack their personal identity. This can discredit politicians and reduce public trust in the government. It can also incite anger and disrespect, leading to offline harm and violence. While numerous models exist for detecting hostility in general, they lack the specificity required for political contexts. Furthermore, addressing hostility towards politicians demands tailored approaches due to the distinct language and issues inherent to each country (e.g., Brexit for the UK). To bridge this gap, we construct a dataset of 3,320 English tweets spanning a two-year period manually annotated for hostility towards UK MPs. Our dataset also captures the targeted identity characteristics (race, gender, religion, none) in hostile tweets. We perform linguistic and topical analyses to delve into the unique content of the UK political data. Finally, we evaluate the performance of pre-trained language models and large language models on binary hostility detection and multi-class targeted identity type classification tasks. Our study offers valuable data and insights for future research on the prevalence and nature of politics-related hostility specific to the UK.
I review the scientific and technical history of the Search for Extraterrestrial Intelligence (SETI), discuss the impact of the political involvement, and speculate on the nature of a successful detection and its potential social and cultural impact. Emphasis is on the development of SETI in the United States and the complementary progress in the Former Soviet Union.
AbstractThis review article surveys recent studies of the state of and challenges to academic labor in the ongoing regime of academic capitalism, corporate managerialism, and neoliberalism in colleges and universities in the United States, Europe, and select other countries around the world. Some works analyze changing funding models, accountability mechanisms, and forms of administrative power, while others explore the discourses pervading higher education and impacting the self-understanding of academics. Higher education administrators, boards of trustees, and politicians have sought to create flexible and inexpensive academic labor. New studies explore the three main strategies pursued: the failed effort to promote Massive Open Online Courses (MOOCs), the proliferation of for-profit colleges and universities (FPCUs), and the continuing expansion of contingent labor, full and part time. Other works analyze the innovative unionization efforts on the part of contingent faculty and graduate teaching assistants.
Este artigo é um estudo teórico que pretende analisar a bibliografia sobre a Educação Integral no Brasil, no que diz respeito às suas concepções surgidas ao longo da história. O intuito é tecer alguns posicionamentos sobre as discussões e problemas levantados a partir da pesquisa. Partimos do pressuposto de que o entendimento sobre a Educação Integral foi se alterando conforme as experiências e os modelos ao longo dos anos e compreender como isso aconteceu é um dos objetivos deste trabalho. Ao analisarmos a documentação sobre a Educação Integral no Brasil, deparamo-nos com experiências regionais e nacionais, que nos permitiram contextualizar e jogar luz à própria compreensão de Educação Integral e suas especificidades no território brasileiro. Muitas dessas experiências nos mostraram que a Educação Integral foi implantada no Brasil com objetivos mais assistencialistas do que educacionais. Trazer à tona o contexto dessas experiências pode nos ajudar a compreender alguns dos problemas atuais da educação brasileira.
Social Sciences, Labor in politics. Political activity of the working class
Mykola Makhortykh, Ernesto de León, Aleksandra Urman
et al.
The growing availability of data about online information behaviour enables new possibilities for political communication research. However, the volume and variety of these data makes them difficult to analyse and prompts the need for developing automated content approaches relying on a broad range of natural language processing techniques (e.g. machine learning- or neural network-based ones). In this paper, we discuss how these techniques can be used to detect political content across different platforms. Using three validation datasets, which include a variety of political and non-political textual documents from online platforms, we systematically compare the performance of three groups of detection techniques relying on dictionaries, supervised machine learning, or neural networks. We also examine the impact of different modes of data preprocessing (e.g. stemming and stopword removal) on the low-cost implementations of these techniques using a large set (n = 66) of detection models. Our results show the limited impact of preprocessing on model performance, with the best results for less noisy data being achieved by neural network- and machine-learning-based models, in contrast to the more robust performance of dictionary-based models on noisy data.
O artigo apresenta reflexões sobre práticas interculturais em Angola (África) narradas em 2008 por trinta religiosas da Congregação das Irmãs Catequistas Franciscanas (CICAF), ano em que a Congregação celebrava 25 anos de presença missionária entre os vários povos do referido país angolano: Bantu, Kikongo, Umbunto, Ibinda, entre outros. Ao todo são oitenta narrativas publicadas num caderno intitulado Cores da vida: 25 anos de presença em Angola (2010). Na perspectiva de Walter Benjamin, essas religiosas são compreendidas como narradoras, e suas narrativas, permeadas pela memória de aprendizes, constituem-se em experiência. A substância de cada narrativa é a experiência que foi sendo tecida por elementos da vivência, do acontecido, visto ou ouvido, da subjetividade que, ao se enraizar na objetividade, é narrada. Conclui que a rememoração das práticas interculturais, à luz de princípios da Teologia da Libertação e do franciscanismo, toma feições de educação popular. Nesse sentido, no escopo de compreender a experiência das narradoras, o artigo busca relacionar tal experiência com a educação popular em geral e, especialmente, a vinculada ao pensamento de Paulo Freire, que ganhou força na década de 1960.
Social Sciences, Labor in politics. Political activity of the working class
Maria Aparecida Augusto Satto Vilela, Beatriz Fernanda Carreira, Luana Maciel do Nascimento
Este relato objetiva refletir sobre a proposta do programa Ações Formativas Integradas de Apoio ao Ingresso no Ensino Superior (Afin), no ano de 2020, no que diz respeito ao cumprimento de seus objetivos, relacionando-o aos princípios da educação na perspectiva freireana. Como procedimentos metodológicos, apresentamos a análise de documentos: a proposta do Afin/Ituiutaba 2020 no modelo remoto devido à pandemia da Covid-19; o projeto submetido ao Sistema de Informação de Extensão; e o relatório de resultados que contém as atividades realizadas. A análise considerou o diálogo com uma educação que combate a alienação e a opressão, subsidiando-se por uma perspectiva freireana (BRANDÃO, 2006; FREIRE, 1979a, 1979b, 1987, 1996). Dessa maneira, foi possível analisar as ações do programa em 2020 como impulsionadoras de uma educação que se pauta nos princípios da autonomia e emancipação, fundamentada nas trocas de experiências entre estudantes, docentes e comunidade. Notamos que, apesar dos diversos desafios para a promoção dessa perspectiva educacional, o Afin/Ituiutaba se reinventou e conseguiu atingir seus objetivos, fazendo com que o direito à educação fosse mais acessível a uma parcela de estudantes, principalmente aos grupos negligenciados diante das desigualdades sociais.
Social Sciences, Labor in politics. Political activity of the working class
Lindgren Erik, Per Pettersson-Lidbom, Bjorn Tyrefors
In this paper, we estimate the causal effect of political power on the provision of public education. We use data from a historical nondemocratic society with a weighted voting system where eligible voters received votes in proportion to their taxable income and without any limit on the maximum of votes, i.e., the political system used in Swedish local governments during the period 1862-1909. We use a novel identification strategy where we combine two different identification strategies, i.e., a threshold regression analysis and a generalized event-study design, both of which exploit nonlinearities or discontinuities in the effect of political power between two opposing local elites: agricultural landowners and emerging industrialists. The results suggest that school spending is approximately 90-120% higher if the nonagrarian interest controls all of the votes compared to when landowners have more than a majority of votes. Moreover, we find no evidence that the concentration of landownership affected this relationship
In this paper I present a classifier for automatic identification of linguistic politeness in Hindi texts. I have used the manually annotated corpus of over 25,000 blog comments to train an SVM. Making use of the discursive and interactional approaches to politeness the paper gives an exposition of the normative, conventionalised politeness structures of Hindi. It is seen that using these manually recognised structures as features in training the SVM significantly improves the performance of the classifier on the test set. The trained system gives a significantly high accuracy of over 77% which is within 2% of human accuracy.
José Eduardo Fernandes, Elson de Oliveira Felice, Cristiane Betanho
Este relato de experiência apresenta a reflexão dos participantes de uma Roda de Conversa durante a V Jornada Universitária em Defesa da Reforma Agrária da Universidade Federal de Uberlândia. Os participantes refletiram sobre a atual conjuntura sociopolítica e a realidade das iniciativas de trabalhadores que se propuseram a se organizar a partir dos princípios da Economia Popular Solidária. Os trabalhadores apontam a importância do desenvolvimento de práticas educativas críticas entre trabalhadores e incubadoras universitárias, para avançarmos para além da caridade e da solidariedade jurídica, para o reconhecimento do outro, no sentido da solidariedade de classe.
Social Sciences, Labor in politics. Political activity of the working class