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.
Social norms can facilitate societal coexistence in groups by providing an implicitly shared set of expectations and behavioral guidelines. However, different social groups can hold different norms, and lacking an overarching normative consensus can lead to conflict within and between groups. In this chapter, we present an agent-based model that simulates the adoption of norms in two interacting groups. We explore this phenomenon while varying relative group sizes and homophily/heterophily (two features of network structure), and initial group norm distributions. Agents update their norm according to an adapted version of Granovetter's threshold model, using a uniform distribution of thresholds. We study the impact of network structure and initial norm distributions on the process of achieving norma-(eds). Computational Conflict Research. Cham: Springer Nature.) 1 2 J. Kohne, N. Gallagher, Z.M. Kirgil, R. Paolillo, L. Padmos, F. Karimi tive consensus and the resulting potential for intragroup and intergroup conflict. Our results show that norm change is most likely when norms are strongly tied to group membership. Groups end up with the most similar norm distributions when networks are heterophilic, with small to middling minority groups. High homophilic networks show high potential intergroup conflict and low potential intragroup conflict, while the opposite pattern emerges for high heterophilic networks.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.