Polarization and reliability of news sources in Wikipedia (2210.16065v2)

Abstract: Wikipedia is the largest online encyclopedia: its open contribution policy allows everyone to edit and share their knowledge. A challenge of radical openness is that it facilitates introducing biased contents or perspectives in Wikipedia. Wikipedia relies on numerous external sources such as journal articles, books, news media, and more. News media sources, in particular, take up nearly third of all citations from Wikipedia. However, despite their importance for providing up-to-date and factual contents, there is still a limited understanding on which news media sources are cited from Wikipedia. Relying on a large-scale open dataset of nearly 30M citations from English Wikipedia, we find a moderate yet systematic liberal polarization in the selection of news media sources. We also show that this effect is not mitigated by controlling for news media factual reliability. Our results contribute to Wikipedia's knowledge integrity agenda in suggesting that a systematic effort would help to better map potential biases in Wikipedia and find means to strengthen its neutral point of view policy.

• The paper finds a moderate liberal polarization in cited news sources, with an average score of -0.51.
• It applies regression analysis on a dataset of nearly 30 million citations supplemented by external bias and reliability ratings.
• The study reveals no straightforward link between reliability and polarization, indicating complex dynamics in source selection.

An Analysis of Political Polarization and Source Reliability in Wikipedia Citations

The paper "Polarization and reliability of news sources in Wikipedia" authored by Puyu Yang and Giovanni Colavizza offers an empirical analysis of the political polarization and reliability of news media sources cited within English Wikipedia. The authors leverage an extensive dataset containing nearly 30 million citations to investigate the biases influencing Wikipedia’s content, focusing particularly on the selection of news media sources given their substantial representation (approximately 30%) among citations.

Summary and Methodology

The authors address two primary research questions: the presence of political polarization in news sources cited by Wikipedia and the potential relationship of this polarization with the factual reliability of these sources. To tackle these questions, the authors utilize the Wikipedia Citations dataset, supplemented with data from the Media Bias Monitor (MBM) and the Media Bias Fact Check (MBFC). These external sources provide political polarization scores and reliability ratings for media outlets.

The authors employ regression analysis to understand how the polarization of cited sources correlates with their reliability and the context in which they are used on Wikipedia. Statistical findings suggest a moderate liberal polarization in the news media sources referenced by Wikipedia, with a mean polarization score swaying towards the liberal side of the spectrum. This distribution was consistent across multiple Wikipedia topics and projects.

Key Numerical Results

The paper reveals the average political polarization score of news sources cited on Wikipedia as -0.51 on a scale where -2 denotes 'very liberal' and 2 denotes 'very conservative'. Furthermore, the analysis highlights balanced citation practices across various Wikipedia topics and projects, suggesting that the observed polarization is not confined to specific areas but is a widespread phenomenon.

When exploring the relationship between reliability and polarization, the authors find no straightforward correlation. High-reliability sources show varying political leanings, while mixed-reliability sources surface predominantly on the liberal side. This challenges the hypothesis that editors may prioritize reliability over political neutrality or vice versa.

Implications and Future Research Directions

The implications of this paper affect Wikipedia's ongoing efforts to maintain knowledge integrity and its neutral point of view policy. The findings raise concerns about potential systemic biases in source selection, which could influence the broader perception and usage of Wikipedia as a neutral encyclopedia.

This investigation serves as a basis for further exploration of the factors contributing to Wikipedia's perceived biases. Future research could benefit from:

• Expanding the analysis to include additional dimensions of polarization and reliability metrics beyond the binary liberal-conservative axis.
• Examining the influence of editors' personal biases and how these might shape source selection and citation practices.
• Assessing temporal variations in source usage and how these changes might reflect shifts in political context or media landscape.

Conclusion

Overall, the paper provides valuable insights into the political and reliability landscapes of news sources within Wikipedia, suggesting areas for improvement in citation practices and content curation. By shedding light on these aspects, the paper contributes to a broader understanding of the dynamics affecting neutrality and verifiability in collaboratively maintained digital knowledge platforms. The research underscores the need for continued vigilance and methodical scrutiny of information sources in pursuit of a balanced and unbiased aggregation of human knowledge.