Claudia Wagner

Indira Sen, Dennis Assenmacher, Mattia Samory et al.

NLP models are used in a variety of critical social computing tasks, such as detecting sexist, racist, or otherwise hateful content. Therefore, it is imperative that these models are robust to spurious features. Past work has attempted to tackle such spurious features using training data augmentation, including Counterfactually Augmented Data (CADs). CADs introduce minimal changes to existing training data points and flip their labels; training on them may reduce model dependency on spurious features. However, manually generating CADs can be time-consuming and expensive. Hence in this work, we assess if this task can be automated using generative NLP models. We automatically generate CADs using Polyjuice, ChatGPT, and Flan-T5, and evaluate their usefulness in improving model robustness compared to manually-generated CADs. By testing both model performance on multiple out-of-domain test sets and individual data point efficacy, our results show that while manual CADs are still the most effective, CADs generated by ChatGPT come a close second. One key reason for the lower performance of automated methods is that the changes they introduce are often insufficient to flip the original label.

Indira Sen, Mattia Samory, Claudia Wagner et al.

Counterfactually Augmented Data (CAD) aims to improve out-of-domain generalizability, an indicator of model robustness. The improvement is credited with promoting core features of the construct over spurious artifacts that happen to correlate with it. Yet, over-relying on core features may lead to unintended model bias. Especially, construct-driven CAD -- perturbations of core features -- may induce models to ignore the context in which core features are used. Here, we test models for sexism and hate speech detection on challenging data: non-hateful and non-sexist usage of identity and gendered terms. In these hard cases, models trained on CAD, especially construct-driven CAD, show higher false-positive rates than models trained on the original, unperturbed data. Using a diverse set of CAD -- construct-driven and construct-agnostic -- reduces such unintended bias.

Tobias Holtdirk, Dennis Assenmacher, Arnim Bleier et al.

Survey researchers face two key challenges: the rising costs of probability samples and missing data (e.g., non-response or attrition), which can undermine inference and increase the use of convenience samples. Recent work explores using large language models (LLMs) to simulate respondents via persona-based prompts, often without labeled data. We study a more practical setting where partial survey responses exist: we fine-tune LLMs on available data to impute self-reported vote choice under both random and systematic nonresponse, using the German Longitudinal Election Study. We compare zero-shot prompting and supervised fine-tuning against tabular classifiers (e.g., CatBoost) and test how different convenience samples (e.g., students) used for fine-tuning affect generalization. Our results show that when data are missing completely at random, fine-tuned LLMs match tabular classifiers but outperform zero-shot approaches. When only biased convenience samples are available, fine-tuning small (3B to 8B) open-source LLMs can recover both individual-level predictions and population-level distributions more accurately than zero-shot and often better than tabular methods. This suggests fine-tuned LLMs offer a promising strategy for researchers working with non-probability samples or systematic missingness, and may enable new survey designs requiring only easily accessible subpopulations.

Lukas Birkenmaier, Claudia Wagner, Clemens Lechner

Guidance on how to validate computational text-based measures of social constructs is fragmented. While researchers generally acknowledge the importance of validating text-based measures, they often lack a shared vocabulary and a unified framework to do so. This paper introduces ValiText, a new validation framework designed to assist scholars in validly measuring social constructs in textual data. The framework is built on a conceptual foundation of validity in the social sciences, strengthened by an empirical review of validation practices in the social sciences and consultations with experts. Ultimately, ValiText prescribes researchers to demonstrate three types of validation evidence: substantive evidence (outlining the theoretical underpinning of the measure), structural evidence (examining the properties of the text model and its output) and external evidence (testing for how the measure relates to independent information). The framework is further supplemented by a checklist of validation steps, offering practical guidance in the form of documentation sheets that guide researchers in the validation process.

Vigneshwaran Shankaran, Gabriella Lapesa, Claudia Wagner

Few forces rival fear in their ability to mobilize societies, distort communication, and reshape collective behavior. In computational linguistics, fear is primarily studied as an emotion, but not as a distinct form of speech. Fear speech content is widespread and growing, and often outperforms hate-speech content in reach and engagement because it appears "civiler" and evades moderation. Yet the computational study of fear speech remains fragmented and under-resourced. This can be understood by recognizing that fear speech is a phenomenon shaped by contributions from multiple disciplines. In this paper, we bridge cross-disciplinary perspectives by comparing theories of fear from Psychology, Political science, Communication science, and Linguistics. Building on this, we review existing definitions. We follow up with a survey of datasets from related research areas and propose a taxonomy that consolidates different dimensions of fear for studying fear speech. By reviewing current datasets and defining core concepts, our work offers both theoretical and practical guidance for creating datasets and advancing fear speech research.

Zehui Yu, Indira Sen, Dennis Assenmacher et al.

Machine learning (ML)-based content moderation tools are essential to keep online spaces free from hateful communication. Yet, ML tools can only be as capable as the quality of the data they are trained on allows them. While there is increasing evidence that they underperform in detecting hateful communications directed towards specific identities and may discriminate against them, we know surprisingly little about the provenance of such bias. To fill this gap, we present a systematic review of the datasets for the automated detection of hateful communication introduced over the past decade, and unpack the quality of the datasets in terms of the identities that they embody: those of the targets of hateful communication that the data curators focused on, as well as those unintentionally included in the datasets. We find, overall, a skewed representation of selected target identities and mismatches between the targets that research conceptualizes and ultimately includes in datasets. Yet, by contextualizing these findings in the language and location of origin of the datasets, we highlight a positive trend towards the broadening and diversification of this research space.

Lorraine Saju, Arnim Bleier, Jana Lasser et al.

The proliferation of misinformation necessitates scalable, automated fact-checking solutions. Yet, current benchmarks often overlook multilingual and topical diversity. This paper introduces a novel, dynamically extensible data set that includes 61,514 claims in multiple languages and topics, extending existing datasets up to 2024. Through a comprehensive evaluation of five prominent Large Language Models (LLMs), including GPT-4o, GPT-3.5 Turbo, LLaMA 3.1, and Mixtral 8x7B, we identify significant performance gaps between different languages and topics. While overall GPT-4o achieves the highest accuracy, it declines to classify 43% of claims. Across all models, factual-sounding claims are misclassified more often than opinions, revealing a key vulnerability. These findings underscore the need for caution and highlight challenges in deploying LLM-based fact-checking systems at scale.

Dennis Assenmacher, Paloma Piot, Katarina Laken et al.

Digital dehumanization, although a critical issue, remains largely overlooked within the field of computational linguistics and Natural Language Processing. The prevailing approach in current research concentrating primarily on a single aspect of dehumanization that identifies overtly negative statements as its core marker. This focus, while crucial for understanding harmful online communications, inadequately addresses the broader spectrum of dehumanization. Specifically, it overlooks the subtler forms of dehumanization that, despite not being overtly offensive, still perpetuate harmful biases against marginalized groups in online interactions. These subtler forms can insidiously reinforce negative stereotypes and biases without explicit offensiveness, making them harder to detect yet equally damaging. Recognizing this gap, we use different sampling methods to collect a theory-informed bilingual dataset from Twitter and Reddit. Using crowdworkers and experts to annotate 16,000 instances on a document- and span-level, we show that our dataset covers the different dimensions of dehumanization. This dataset serves as both a training resource for machine learning models and a benchmark for evaluating future dehumanization detection techniques. To demonstrate its effectiveness, we fine-tune ML models on this dataset, achieving performance that surpasses state-of-the-art models in zero and few-shot in-context settings.

Rabiraj Bandyopadhyay, Dennis Assenmacher, Jose M. Alonso Moral et al.

There is an increase in the proliferation of online hate commensurate with the rise in the usage of social media. In response, there is also a significant advancement in the creation of automated tools aimed at identifying harmful text content using approaches grounded in Natural Language Processing and Deep Learning. Although it is known that training Deep Learning models require a substantial amount of annotated data, recent line of work suggests that models trained on specific subsets of the data still retain performance comparable to the model that was trained on the full dataset. In this work, we show how we can leverage influence scores to estimate the importance of a data point while training a model and designing a pruning strategy applied to the case of sexism detection. We evaluate the model performance trained on data pruned with different pruning strategies on three out-of-domain datasets and find, that in accordance with other work a large fraction of instances can be removed without significant performance drop. However, we also discover that the strategies for pruning data, previously successful in Natural Language Inference tasks, do not readily apply to the detection of harmful content and instead amplify the already prevalent class imbalance even more, leading in the worst-case to a complete absence of the hateful class.

Georg Ahnert, Ivan Smirnov, Florian Lemmerich et al.

Measures of algorithmic fairness often do not account for human perceptions of fairness that can substantially vary between different sociodemographics and stakeholders. The FairCeptron framework is an approach for studying perceptions of fairness in algorithmic decision making such as in ranking or classification. It supports (i) studying human perceptions of fairness and (ii) comparing these human perceptions with measures of algorithmic fairness. The framework includes fairness scenario generation, fairness perception elicitation and fairness perception analysis. We demonstrate the FairCeptron framework by applying it to a hypothetical university admission context where we collect human perceptions of fairness in the presence of minorities. An implementation of the FairCeptron framework is openly available, and it can easily be adapted to study perceptions of algorithmic fairness in other application contexts. We hope our work paves the way towards elevating the role of studies of human fairness perceptions in the process of designing algorithmic decision making systems.

Mattia Samory, Indira Sen, Julian Kohne et al.

Research has focused on automated methods to effectively detect sexism online. Although overt sexism seems easy to spot, its subtle forms and manifold expressions are not. In this paper, we outline the different dimensions of sexism by grounding them in their implementation in psychological scales. From the scales, we derive a codebook for sexism in social media, which we use to annotate existing and novel datasets, surfacing their limitations in breadth and validity with respect to the construct of sexism. Next, we leverage the annotated datasets to generate adversarial examples, and test the reliability of sexism detection methods. Results indicate that current machine learning models pick up on a very narrow set of linguistic markers of sexism and do not generalize well to out-of-domain examples. Yet, including diverse data and adversarial examples at training time results in models that generalize better and that are more robust to artifacts of data collection. By providing a scale-based codebook and insights regarding the shortcomings of the state-of-the-art, we hope to contribute to the development of better and broader models for sexism detection, including reflections on theory-driven approaches to data collection.

Indira Sen, Fabian Floeck, Katrin Weller et al.

Peoples' activities and opinions recorded as digital traces online, especially on social media and other web-based platforms, offer increasingly informative pictures of the public. They promise to allow inferences about populations beyond the users of the platforms on which the traces are recorded, representing real potential for the Social Sciences and a complement to survey-based research. But the use of digital traces brings its own complexities and new error sources to the research enterprise. Recently, researchers have begun to discuss the errors that can occur when digital traces are used to learn about humans and social phenomena. This article synthesizes this discussion and proposes a systematic way to categorize potential errors, inspired by the Total Survey Error (TSE) Framework developed for survey methodology. We introduce a conceptual framework to diagnose, understand, and document errors that may occur in studies based on such digital traces. While there are clear parallels to the well-known error sources in the TSE framework, the new "Total Error Framework for Digital Traces of Human Behavior on Online Platforms" (TED-On) identifies several types of error that are specific to the use of digital traces. By providing a standard vocabulary to describe these errors, the proposed framework is intended to advance communication and research concerning the use of digital traces in scientific social research.

Federica Calanca, Luiza Sayfullina, Lara Minkus et al.

During the past decades the importance of soft skills for labour market outcomes has grown substantially. This carries implications for labour market inequality, since previous research shows that soft skills are not valued equally across race and gender. This work explores the role of soft skills in job advertisements by drawing on methods from computational science as well as on theoretical and empirical insights from economics, sociology and psychology. We present a semi-automatic approach based on crowdsourcing and text mining for extracting a list of soft skills. We find that soft skills are a crucial component of job ads, especially of low-paid jobs and jobs in female-dominated professions. Our work shows that soft skills can serve as partial predictors of the gender composition in job categories and that not all soft skills receive equal wage returns at the labour market. Especially "female" skills are frequently associated with wage penalties. Our results expand the growing literature on the association of soft skills on wage inequality and highlight their importance for occupational gender segregation at labour markets.

Martin Körner, Tatiana Sennikova, Florian Windhäuser et al.

Wikipedia articles about the same topic in different language editions are built around different sources of information. For example, one can find very different news articles linked as references in the English Wikipedia article titled "Annexation of Crimea by the Russian Federation" than in its German counterpart (determined via Wikipedia's language links). Some of this difference can of course be attributed to the different language proficiencies of readers and editors in separate language editions, yet, although including English-language news sources seems to be no issue in the German edition, English references that are listed do not overlap highly with the ones in the article's English version. Such patterns could be an indicator of bias towards certain national contexts when referencing facts and statements in Wikipedia. However, determining for each reference which national context it can be traced back to, and comparing the link distributions to each other is infeasible for casual readers or scientists with non-technical backgrounds. Wikiwhere answers the question where Web references stem from by analyzing and visualizing the geographic location of external reference links that are included in a given Wikipedia article. Instead of relying solely on the IP location of a given URL our machine learning models consider several features.

Claudia Wagner, Philipp Singer, Markus Strohmaier et al.

One potential disadvantage of social tagging systems is that due to the lack of a centralized vocabulary, a crowd of users may never manage to reach a consensus on the description of resources (e.g., books, users or songs) on the Web. Yet, previous research has provided interesting evidence that the tag distributions of resources may become semantically stable over time as more and more users tag them. At the same time, previous work has raised an array of new questions such as: (i) How can we assess the semantic stability of social tagging systems in a robust and methodical way? (ii) Does semantic stabilization of tags vary across different social tagging systems and ultimately, (iii) what are the factors that can explain semantic stabilization in such systems? In this work we tackle these questions by (i) presenting a novel and robust method which overcomes a number of limitations in existing methods, (ii) empirically investigating semantic stabilization processes in a wide range of social tagging systems with distinct domains and properties and (iii) detecting potential causes for semantic stabilization, specifically imitation behavior, shared background knowledge and intrinsic properties of natural language. Our results show that tagging streams which are generated by a combination of imitation dynamics and shared background knowledge exhibit faster and higher semantic stability than tagging streams which are generated via imitation dynamics or natural language streams alone.