---
title: "Language Matters: Target-Language Supervision for Political"
description: "The authors present what they describe as the first systematic transformer-based study of outlet-ideology… 70% accuracy. 27 references."
authors: ["Umut Ozbagriacik", "Haim Dubossarsky"]
year: 2026
venue: "Proceedings of the Second Workshop Natural Language Processing for Turkic Languages (SIGTURK 2026)"
source: "https://aclanthology.org/2026.sigturk-1.7/"
doi: "10.18653/v1/2026.sigturk-1.7"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/ozbagriacik-2026-language-matters-target-language-supervision/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/ozbagriacik-2026-language-matters-target-language-supervision/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/38/) / Language Matters: Target-Language Supervision for Political Bias Detection in Turkish News

# Language Matters: Target-Language Supervision for Political Bias Detection in Turkish News

**Authors:** Umut Ozbagriacik and [Haim Dubossarsky](https://md.research.elamur.ai/free/en/authors/haim-dubossarsky/) · **Year:** 2026 · **Venue:** [Proceedings of the Second Workshop Natural Language Processing for Turkic Languages (SIGTURK 2026)](https://md.research.elamur.ai/free/en/venues/sigturk/) ([Workshop Natural Language Processing for Turkic Languages 2026](https://md.research.elamur.ai/free/en/venues/sigturk/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.sigturk-1.7/) · [PDF](https://aclanthology.org/2026.sigturk-1.7.pdf) · [DOI 10.18653/v1/2026.sigturk-1.7](https://doi.org/10.18653/v1/2026.sigturk-1.7) · [OpenAlex](https://openalex.org/W7140078603)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Political bias detection research has focused mostly on English. Turkish has a highly polarised press but lacks labelled data for this task. Its political divide also differs from the Western one: it runs mainly between secular-liberal and religious-conservative camps. The authors present what they describe as the first systematic transformer-based study of outlet-ideology classification for Turkish news. They built a topic-balanced corpus with a keyword-driven crawler that pulls articles on the same political subjects from six outlets. Each article inherits its outlet's reputed leaning (left, centre or right) as a weak label. They then fine-tuned and compared BERTurk, mBERT, XLM-R and a LoRA-adapted Mistral model. The monolingual BERTurk performed best, roughly on par with results reported for English. All models had trouble separating left from right, though they reliably told partisan articles from centrist ones. Models trained only on English bias data fell to near-chance accuracy on Turkish. The study gives a first baseline and shows that supervision in the target language is necessary. It also argues for open Turkish and broader Turkic bias benchmarks.

## Key points
- **Corpus and labels:** The corpus covers six outlets, two per class:
  - left: BirGün, Sözcü
  - centre: Habertürk, Euronews Türkçe
  - right: Milliyet, Diriliş Postası

  Labels reflect outlet reputation, not article-level annotation.
- **Data collection:** The crawler queried each site's search function with political keywords (e.g. Rusya, Suriye, Trump, Erdoğan, protesto), taking up to 12 articles per keyword per source. This produced an initial pool of about 3,200 articles. Copyright prevents releasing the texts.
- **Main result:** BERTurk was the strongest single model, with 70% accuracy and 71% macro-F1. Confusion matrices show that left–right mix-ups are the dominant error.
- **"Ideological fingerprint":** BERTurk did relatively better on right-leaning articles, while XLM-R and mBERT did better on left-leaning ones. The authors attribute this to differences in pre-training data.
- **Cross-lingual transfer:** Models fine-tuned on an English political-bias task collapsed to near-chance performance on Turkish. The authors conclude that naïve transfer from English is not a viable substitute for Turkish training data.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Queen Mary University of London](https://md.research.elamur.ai/free/en/universities/queen-mary-university-of-london/) — University, London, United Kingdom · 57 works · 3,600 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works

🧭 Topics (1): [🛡️ Bias, fairness & safety](https://md.research.elamur.ai/free/en/topics/bias-fairness-safety/) · 587 works

## Cite

Cite as: Umut Ozbagriacik and Haim Dubossarsky. 2026. Language Matters: Target-Language Supervision for Political Bias Detection in Turkish News. Proceedings of the Second Workshop Natural Language Processing for Turkic Languages (SIGTURK 2026).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{ozbagriacik2026language,
  title = {{Language Matters: Target-Language Supervision for Political Bias Detection in Turkish News}},
  author = {Umut Ozbagriacik and Haim Dubossarsky},
  booktitle = {Proceedings of the Second Workshop Natural Language Processing for Turkic Languages (SIGTURK 2026)},
  year = {2026},
  doi = {10.18653/v1/2026.sigturk-1.7},
  url = {https://doi.org/10.18653/v1/2026.sigturk-1.7}
}
```

- [← previous Language Effects in Text-to-SQL Across English and Portuguese](https://md.research.elamur.ai/free/en/papers/nobre-2026-language-effects-text-sql-across/)
- [next → Language Mixture to Develop Accurate Galician Dependency Parsers: An Exploration of Its Effects](https://md.research.elamur.ai/free/en/papers/urbieta-2026-language-mixture-develop-accurate-galician/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/38/)

## References

27 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. AllSides. n.d. Media bias ratings. [https://www.al](https://www.al) lsides.com/media-bias/media-bias-ratings. Accessed: 28 July 2025. [↗](https://www.al)
2. Articoder. 2020. News dataset for news bias analysis. [https://www.kaggle.com/datasets/artico](https://www.kaggle.com/datasets/artico) der/news-dataset-for-news-bias-analysis. Accessed: 17 December 2025. [↗](https://www.kaggle.com/datasets/artico)
3. Alessio Bajec. 2023. Turkey’s opposition opens up to the hijab. Al Jazeera. Accessed: 17 August 2025. [find](https://scholar.google.com/scholar?q=Alessio+Bajec%2E+2023%2E+Turkey%E2%80%99s+opposition+opens+up+to+the+hijab%2E+Al+Jazeera%2E+Accessed%3A+17+August+2025%2E)
4. Ramy Baly, Giovanni Da San Martino, James Glass, and Preslav Nakov. 2020. [We Can Detect Your Bias: Predicting the political ideology of news articles](https://md.research.elamur.ai/free/en/papers/baly-2020-we-can-detect-your-bias/). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 4982–4991. Association for Computational Lin…
5. Ceren Budak, Sharad Goel, and Justin M. Rao. 2016. Fair and balanced? quantifying media bias through crowdsourced content analysis. Public Opinion Quarterly, 80(S1):250–271. [find](https://scholar.google.com/scholar?q=Ceren+Budak%2C+Sharad+Goel%2C+and+Justin+M%2E+Rao%2E+2016%2E+Fair+and+balanced%3F+quantifying+media+bias+through+crowdsourced+content+analysis%2E+Public+Opinion+Quarterly%2C+80%28S1%29%3A250%E2%80%93271%2E)
6. Emrah Budur, Rıza Özçelik, Tunga Gungor, and Christopher Potts. 2020. Data and Representation for Turkish Natural Language Inference. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 8253– 8267, Online. Association for Computational Linguistics… [find](https://scholar.google.com/scholar?q=Emrah+Budur%2C+R%C4%B1za+%C3%96z%C3%A7elik%2C+Tunga+Gungor%2C+and+Christopher+Potts%2E+2020%2E+Data+and+Representation+for+Turkish+Natural+Language+Inference%2E+In+Proceedings+of+the+2020+Conference+on+Empirical+Methods+in+Natu)
7. Roksana Goworek, Harpal Singh Karlcut, Hamza Shezad, Nijaguna Darshana, Abhishek Mane, Syam Bondada, Raghav Sikka, Ulvi Mammadov, Rauf Allahverdiyev, Sriram Satkirti Purighella, Paridhi Gupta, Muhinyia Ndegwa, Bao Khanh Tran, and Haim Dubossarsky. 2025. SenWiCh: Sense-annotation of low-resource lang… [find](https://scholar.google.com/scholar?q=Roksana+Goworek%2C+Harpal+Singh+Karlcut%2C+Hamza+Shezad%2C+Nijaguna+Darshana%2C+Abhishek+Mane%2C+Syam+Bondada%2C+Raghav+Sikka%2C+Ulvi+Mammadov%2C+Rauf+Allahverdiyev%2C+Sriram+Satkirti+Purighella%2C+Paridhi+Gupta%2C+Muhinyi)
8. Wei-Fan Chen, Khalid Al-Khatib, Benno Stein, and Henning Wachsmuth. 2020. Detecting media bias in news articles using gaussian bias distributions. In Findings of the Association for Computational Linguistics: EMNLP 2020. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Wei%2DFan+Chen%2C+Khalid+Al%2DKhatib%2C+Benno+Stein%2C+and+Henning+Wachsmuth%2E+2020%2E+Detecting+media+bias+in+news+articles+using+gaussian+bias+distributions%2E+In+Findings+of+the+Association+for+Computational+Ling)
9. Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. [Unsupervised cross-lingual representation learning at scale](https://md.research.elamur.ai/free/en/papers/conneau-2020-unsupervised-cross-lingual-representation-learning/). In Proceedings of the 58th Annual Meeting of the Association fo…
10. Felix Hamborg, Karsten Donnay, and Bela Gipp. 2019. Automated identification of media bias in news articles: an interdisciplinary literature review. International Journal on Digital Libraries, 20(4):391–415. [find](https://scholar.google.com/scholar?q=Felix+Hamborg%2C+Karsten+Donnay%2C+and+Bela+Gipp%2E+2019%2E+Automated+identification+of+media+bias+in+news+articles%3A+an+interdisciplinary+literature+review%2E+International+Journal+on+Digital+Libraries%2C+20%284%29%3A3)
11. Benjamin Horne, Sara Khedr, and Sibel Adali. 2018. Sampling the news producers: A large news and feature data set for the study of the complex media landscape. In Proceedings of the International AAAI Conference on Web and Social Media, volume 12. [find](https://scholar.google.com/scholar?q=Benjamin+Horne%2C+Sara+Khedr%2C+and+Sibel+Adali%2E+2018%2E+Sampling+the+news+producers%3A+A+large+news+and+feature+data+set+for+the+study+of+the+complex+media+landscape%2E+In+Proceedings+of+the+International+AAAI)
12. Farheen Dairkee and Haim Dubossarsky. 2024. Strengthening the wic: New polysemy dataset in hindi and lack of cross lingual transfer. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 15341– 15349. [find](https://scholar.google.com/scholar?q=Farheen+Dairkee+and+Haim+Dubossarsky%2E+2024%2E+Strengthening+the+wic%3A+New+polysemy+dataset+in+hindi+and+lack+of+cross+lingual+transfer%2E+In+Proceedings+of+the+2024+Joint+International+Conference+on+Comput)
13. E. J. Hu, Y. Shen, P. Wallis, Z. Allen-Zhu, Y. Li, S. Wang, L. Wang, and W. Chen. 2021. [Lora: Low-rank adaptation of large language models](https://md.research.elamur.ai/free/en/papers/hu-2021-lora-low-rank-adaptation-large/). Accessed: 17 August 2025.
14. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. [BERT: Pre-training of deep bidirectional transformers for language understanding](https://md.research.elamur.ai/free/en/papers/devlin-2019-bert-pre-training-deep-bidirectional/). In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies…
15. A. Q. Jiang, A. Sablayrolles, A. Mensch, and 1 others. 2023. Mistral 7b: a high-performing lightweight language model. Accessed: 28 July 2025. [find](https://scholar.google.com/scholar?q=A%2E+Q%2E+Jiang%2C+A%2E+Sablayrolles%2C+A%2E+Mensch%2C+and+1+others%2E+2023%2E+Mistral+7b%3A+a+high%2Dperforming+lightweight+language+model%2E+Accessed%3A+28+July+2025%2E)
16. Y. Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, and V. Stoyanov. 2019. [Roberta: a robustly optimized BERT pretraining approach](https://md.research.elamur.ai/free/en/papers/liu-2019-roberta-robustly-optimized-bert-pretraining/). Accessed: 17 August 2025.
17. Doğu Ergil. 2010. Constitutional referendum: Farewell to the ‘old turkey’. Insight Turkey, 12(4):15–22. Accessed: 28 July 2025. [find](https://scholar.google.com/scholar?q=Dog%CC%86u+Ergil%2E+2010%2E+Constitutional+referendum%3A+Farewell+to+the+%E2%80%98old+turkey%E2%80%99%2E+Insight+Turkey%2C+12%284%29%3A15%E2%80%9322%2E+Accessed%3A+28+July+2025%2E)
18. T. Menzner and J. L. Leidner. 2024a. Biasscanner: Automatic detection and classification of news bias to support news readers. arXiv. [find](https://scholar.google.com/scholar?q=T%2E+Menzner+and+J%2E+L%2E+Leidner%2E+2024a%2E+Biasscanner%3A+Automatic+detection+and+classification+of+news+bias+to+support+news+readers%2E+arXiv%2E)
19. Lisa Fan, Marshall White, Eva Sharma, Ruisi Su, Prafulla Kumar Choubey, Ruihong Huang, and Lu Wang. 2019. In plain sight: Media bias through the lens of factual reporting. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Confer… [find](https://scholar.google.com/scholar?q=Lisa+Fan%2C+Marshall+White%2C+Eva+Sharma%2C+Ruisi+Su%2C+Prafulla+Kumar+Choubey%2C+Ruihong+Huang%2C+and+Lu+Wang%2E+2019%2E+In+plain+sight%3A+Media+bias+through+the+lens+of+factual+reporting%2E+In+Proceedings+of+the+2019+C)
20. Tim Menzner and Jochen L. Leidner. 2024b. Experiments in news bias detection with pre-trained neural transformers. [find](https://scholar.google.com/scholar?q=Tim+Menzner+and+Jochen+L%2E+Leidner%2E+2024b%2E+Experiments+in+news+bias+detection+with+pre%2Dtrained+neural+transformers%2E)
21. Andrew O’Donohue, Max Hoffman, and Alan Makovsky. 2020. Turkey’s changing media landscape. Center for American Progress. Published: 10 June 2020. [find](https://scholar.google.com/scholar?q=Andrew+O%E2%80%99Donohue%2C+Max+Hoffman%2C+and+Alan+Makovsky%2E+2020%2E+Turkey%E2%80%99s+changing+media+landscape%2E+Center+for+American+Progress%2E+Published%3A+10+June+2020%2E)
22. Rama Rohit Reddy Gangula, Suma Reddy Duggenpudi, and Radhika Mamidi. 2019. Detecting political bias in news articles using headline attention. In Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 77–84, Florence, Italy. Association for Comput… [find](https://scholar.google.com/scholar?q=Rama+Rohit+Reddy+Gangula%2C+Suma+Reddy+Duggenpudi%2C+and+Radhika+Mamidi%2E+2019%2E+Detecting+political+bias+in+news+articles+using+headline+attention%2E+In+Proceedings+of+the+2019+ACL+Workshop+BlackboxNLP%3A+Anal)
23. Marta Recasens, Cristian Danescu-Niculescu-Mizil, and Dan Jurafsky. 2013. [Linguistic models for analyzing and detecting biased language](https://md.research.elamur.ai/free/en/papers/recasens-2013-linguistic-models-analyzing-detecting-biased/). In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics.
24. Roksana Goworek and Haim Dubossarsky. 2025. Multilinguality does not make sense: Investigating fac-Stefan Schweter. 2020. BERTurk – BERT models for turkish (version 1.0). [find](https://scholar.google.com/scholar?q=Roksana+Goworek+and+Haim+Dubossarsky%2E+2025%2E+Multilinguality+does+not+make+sense%3A+Investigating+fac%2DStefan+Schweter%2E+2020%2E+BERTurk+%E2%80%93+BERT+models+for+turkish+%28version+1%2E0%29%2E)
25. Timo Spinde, Felix Hamborg, and Bela Gipp. 2020. Media bias in german news articles: A combined approach. In ECML PKDD 2020 Workshops, pages 581–590. Springer. [find](https://scholar.google.com/scholar?q=Timo+Spinde%2C+Felix+Hamborg%2C+and+Bela+Gipp%2E+2020%2E+Media+bias+in+german+news+articles%3A+A+combined+approach%2E+In+ECML+PKDD+2020+Workshops%2C+pages+581%E2%80%93590%2E+Springer%2E)
26. Timo Spinde, Lada Rudnitckaia, Kanishka Sinha, Felix Hamborg, Bela Gipp, and Karsten Donnay. 2021. Mbic–a media bias annotation dataset including annotator characteristics. arXiv preprint arXiv:2105.11910. — [arXiv:2105.11910](https://arxiv.org/abs/2105.11910) [find](https://scholar.google.com/scholar?q=Timo+Spinde%2C+Lada+Rudnitckaia%2C+Kanishka+Sinha%2C+Felix+Hamborg%2C+Bela+Gipp%2C+and+Karsten+Donnay%2E+2021%2E+Mbic%E2%80%93a+media+bias+annotation+dataset+including+annotator+characteristics%2E+arXiv+preprint+arXiv%3A2105%2E1)
27. S. Yigit-Sert, İ. S. Altingövde, and Ö. Ulusoy. 2016. Towards detecting media bias by utilising user comments. In Proceedings of the 8th ACM Web Science Conference (WebSci 2016), pages 374–375. ACM. [find](https://scholar.google.com/scholar?q=S%2E+Yigit%2DSert%2C+I%CC%87%2E+S%2E+Alting%C3%B6vde%2C+and+%C3%96%2E+Ulusoy%2E+2016%2E+Towards+detecting+media+bias+by+utilising+user+comments%2E+In+Proceedings+of+the+8th+ACM+Web+Science+Conference+%28WebSci+2016%29%2C+pages+374%E2%80%93375%2E+ACM%2E)

*Updated 2 November 2026*