Computational Assessment of Hyperpartisanship in News Titles


  • Hanjia Lyu University of Rochester
  • Jinsheng Pan University of Rochester
  • Zichen Wang University of Rochester
  • Jiebo Luo University of Rochester



The growing trend of partisanship in news reporting can have a negative impact on society. Assessing the level of partisanship in news headlines is particularly crucial, as they are easily accessible and frequently provide a condensed summary of the article's opinions or events. Therefore, they can significantly influence readers' decision to read the full article, making them a key factor in shaping public opinion. We first adopt a human-guided machine learning framework to develop a new dataset for hyperpartisan news title detection with 2,200 manually labeled and 1.8 million machine-labeled titles that were posted from 2014 to the present by nine representative media organizations across three media bias groups - Left, Central, and Right in an active learning manner. A fine-tuned transformer-based language model achieves an overall accuracy of 0.84 and an F1 score of 0.78 on an external validation set. Next, we conduct a computational analysis to quantify the extent and dynamics of partisanship in news titles. While some aspects are as expected, our study reveals new or nuanced differences between the three media groups. We find that overall the Right media tends to use proportionally more hyperpartisan titles. Roughly around the 2016 Presidential Election, the proportions of hyperpartisan titles increased across all media bias groups, with the Left media exhibiting the most significant relative increase. We identify three major topics including foreign issues, political systems, and societal issues that are suggestive of hyperpartisanship in news titles using logistic regression models and the Shapley values. Through an analysis of the topic distribution, we find that societal issues gradually gain more attention from all media groups. We further apply a lexicon-based language analysis tool to the titles of each topic and quantify the linguistic distance between any pairs of the three media groups, uncovering three distinct patterns. Codes and data are available at




How to Cite

Lyu, H., Pan, J., Wang, Z., & Luo, J. (2024). Computational Assessment of Hyperpartisanship in News Titles. Proceedings of the International AAAI Conference on Web and Social Media, 18(1), 999-1012.