Measuring and Analyzing Toxic Discourse Against Brazilian Federal Congresswomen on the X Platform

• Karen Gomes UFMG
• Jussara M. Almeida UFMG
• Marcos André Gonçalves UFMG
• Fabrício Benevenuto UFMG

DOI: https://doi.org/10.5753/brasnam.2026.23106

Abstract

Toxic discourse in online political environments is often analyzed in an aggregated way, without considering differences across targets. In this work, we investigate toxic discourse directed at Brazilian federal deputies elected in 2022 on the X platform, analyzing how it varies across the ideological spectrum. We combine automatic toxicity detection, topic modeling with BERTopic, language models for labeling, and the analysis of psycholinguistic attributes with LIWC. The results show distinct patterns across political groups. Left-wing deputies receive more attacks associated with emotional and ideological delegitimization, while right-wing deputies are more frequently targeted with attacks related to appearance and sexuality. These findings suggest that online political toxicity varies according to both the target group and the type of attack involved. Warning: This article includes examples of messages containing offensive language for analytical purposes.

References

Al-Hassan, A. and Al-Dossari, H. (2019). Detection of hate speech in social networks: a survey on multilingual corpus. In 6th international conference on computer science and information technology, volume 10, pages 10–5121. ACM.

Alzahrani, M. and AlGhamdi, F. (2025). Social media sentiment analysis for sustainable rural event planning: A case study of agricultural festivals in al-baha, saudi arabia. Sustainability, 17(9):3864.

Amnesty International (2018). Women’s experiences of violence and abuse on twitter. Acesso em: 23 jan. 2026.

Balage Filho, P., Pardo, T. A. S., and Aluísio, S. (2013). An evaluation of the brazilian portuguese liwc dictionary for sentiment analysis. In STIL.

Belisário, K. M. and dos Reis, R. d. C. (2023). Representação feminina na cena política brasileira: Estereótipos e preconceitos. Teoria & Pesquisa Revista de Ciência Política, pages e023011–e023011.

Biroli, F. (2018). Gênero e desigualdades: limites da democracia no Brasil. Boitempo Editorial.

Bolognesi, B., Ribeiro, E., and Codato, A. (2022). Uma nova classificação ideológica dos partidos políticos brasileiros. Dados, 66:e20210164.

Carmo, I., Rêgo, A. L., Barreto, M., Schuler, M., Heine, A., Villas, M. V., and Lifschitz, S. (2023). Gerenciamento de dados de redes sociais com análise de redes e modelagem de tópicos. In Simpósio Brasileiro de Banco de Dados (SBBD), pages 64–70. SBC.

Conselho Nacional de Saúde (2016). Resolução nº 510, de 7 de abril de 2016. Davidson, T., Warmsley, D., Macy, M., and Weber, I. (2017). Automated hate speech detection and the problem of offensive language. In Proceedings of the international AAAI conference on web and social media, volume 11, pages 512–515.

de Groot, M., Aliannejadi, M., and Haas, M. R. (2022). Experiments on generalizability of bertopic on multi-domain short text. arXiv preprint arXiv:2212.08459.

G1 (2022). Carla zambelli tem contas retidas nas redes sociais. Acesso em: 24 jan. 2026. Grattafiori, A., Dubey, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., Letman, A., Mathur, A., Schelten, A., Vaughan, A., et al. (2024). The llama 3 herd of models. arXiv preprint arXiv:2407.21783.

Grootendorst, M. (2022). Bertopic: Neural topic modeling with a class-based tf-idf procedure. arXiv preprint arXiv:2203.05794.

Hanu, L. and Team, U. (2020). Detoxify. [link].

Jigsaw (2018). Toxic comment classification challenge. [link].

Khandelwal, T. (2025). Using llm-based approaches to enhance and automate topic labeling.

Koch, L., Russo Riva, M. P., and Steinert, J. I. (2025). Technology-facilitated gender-based violence against politically active women: A systematic review of psychological and political consequences and women’s coping behaviors. Trauma, Violence, & Abuse.

Kozlowski, D., Pradier, C., and Benz, P. (2024). Generative ai for automatic topic labelling.

Krook, M. L. and Restrepo Sanín, J. (2016). Género y violencia política en américa latina. conceptos, debates y soluciones. Política y gobierno, 23(1):127–162.

Lima, L., Reis, J. C., Melo, P., Murai, F., and Benevenuto, F. (2020). Characterizing (un) moderated textual data in social systems. In 2020 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), pages 430–434.

Marques, F. P. J. A., Aquino, J. A. d., and Miola, E. (2014). Parlamentares, representação política e redes sociais digitais. Opinião Pública, 20:178–203.

Martins, V., Serafim, S. E. V., Pereira, L. C. R., Alves, A. F., Ferreira, C. H., and Almeida, J. M. (2025). A misoginia no youtube brasileiro: Um estudo de caso sobre o conteúdo produzido pela comunidade red pill. In WebMedia, pages 28–37. SBC.

Medvecki, D., Bašaragin, B., Ljajić, A., and Milošević, N. (2024). Multilingual transformer and bertopic for short text topic modeling: The case of Serbian. In Disruptive Information Technologies for a Smart Society. ICIST 2023, volume 872 of Lecture Notes in Networks and Systems, pages 161–173. Springer, Cham.

Melo, G. and Figueredo, F. (2025). O que torna uma frase tóxica? uma análise crítica de modelos especialistas em detecção de toxicidade. In WebMedia, pages 376–384. SBC.

Oltmanns, J. R., Khandelwal, R., Ma, J., Brickman, J., Do, T., Hussain, R., and Gupta, M. (2025). Language-based ai modeling of personality traits and pathology from life narrative interviews. Journal of psychopathology and clinical science.

OpenAI (2024). New embedding models and API updates. Blog post, January 25, 2024. Disponível em: [link].

Pennebaker, J. W., Booth, R. J., and Francis, M. E. (2007). Linguistic inquiry and word count: Liwc [computer software]. Austin, TX: liwc. net, 135.

Pinho, T. R. d. (2020). Debaixo do tapete: a violência política de gênero eo silêncio do conselho de ética da câmara dos deputados. Revista Estudos Feministas, 28(2):e67271.

Poder360 (2022). Carla zambelli muda usuário no twitter e perde verificação. Acesso em: 10 jan. 2026.

Souza, L., Koch, L., Riva, M. P. R., and Gawi, R. (2022). Mensagens de ódio recebidas por candidatas negras e brancas durante as eleições no brasil de 2022 e suas implicações. Estudos Eleitorais, 16(2).

Steinbrenner, T., Lalk, C., Targan, K., Schaffrath, J., Eberhardt, S., Haberkamp, A., Lutz, W., and Rubel, J. (2025). Explaining anxiety prediction in psychotherapy transcripts: The role of patient linguistic features and theoretical constructs. Behaviour Research and Therapy, page 104857.

Tavares, C. Q. and Recuero, R. (2023). Toxicidade e violência discursiva contra deputadas federais no twitter. Galáxia (São Paulo), 48:e62122.

X Corp. (2023). Developer platform policy. [link]. Acesso em: 28 fev. 2026.

Zannettou, S., Bradlyn, B., De Cristofaro, E., Kwak, H., Sirivianos, M., Stringini, G., and Blackburn, J. (2018). What is gab: A bastion of free speech or an alt-right echo chamber. In The Web Conference, pages 1007–1014.

Zedeus (2019). Nitter: A free and open source alternative twitter front-end. [link]. Acesso em: 12 mar. 2026.