# Controlled Analyses of Social Biases in Wikipedia Bios

arXiv: https://arxiv.org/abs/2101.00078 · HTML: https://mlsanity.com/paper/2101.00078/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 31 Dec 2020.

## Built on by (most cited first, 13 of 13)

- [A Survey of Race, Racism, and Anti-Racism in NLP](https://mlsanity.com/paper/2106.11410/) (2021)
- [SynthBio: A Case Study in Human-AI Collaborative Curation of Text Datasets](https://mlsanity.com/paper/2111.06467/) (2021)
- [Gendered Mental Health Stigma in Masked Language Models](https://mlsanity.com/paper/2210.15144/) (2022)
- [Locating Information Gaps and Narrative Inconsistencies Across Languages: A Case Study of LGBT People Portrayals on Wikipedia](https://mlsanity.com/paper/2410.04282/) (2024)
- [Detecting Harmful Online Conversational Content towards LGBTQIA+ Individuals](https://mlsanity.com/paper/2207.10032/) (2022)
- [Quantifying Gender Biases Towards Politicians on Reddit](https://mlsanity.com/paper/2112.12014/) (2021)
- [The URW-KG: a Resource for Tackling the Underrepresentation of non-Western Writers](https://mlsanity.com/paper/2212.13104/) (2022)
- [Considerations for Multilingual Wikipedia Research](https://mlsanity.com/paper/2204.02483/) (2022)
- [The World Literature Knowledge Graph](https://mlsanity.com/paper/2307.16659/) (2023)
- [Multilingual Bias Detection and Mitigation for Indian Languages](https://mlsanity.com/paper/2312.15181/) (2023)
- [White Men Lead, Black Women Help? Benchmarking and Mitigating Language Agency Social Biases in LLMs](https://mlsanity.com/paper/2404.10508/) (2024)
- [Toxic comments reduce the activity of volunteer editors on Wikipedia](https://mlsanity.com/paper/2304.13568/) (2023)
- [Hoaxpedia: A Unified Wikipedia Hoax Articles Dataset](https://mlsanity.com/paper/2405.02175/) (2024)

## Abstract (arXiv)

Social biases on Wikipedia, a widely-read global platform, could greatly influence public opinion. While prior research has examined man/woman gender bias in biography articles, possible influences of other demographic attributes limit conclusions. In this work, we present a methodology for analyzing Wikipedia pages about people that isolates dimensions of interest (e.g., gender), from other attributes (e.g., occupation). Given a target corpus for analysis (e.g.~biographies about women), we present a method for constructing a comparison corpus that matches the target corpus in as many attributes as possible, except the target one. We develop evaluation metrics to measure how well the comparison corpus aligns with the target corpus and then examine how articles about gender and racial minorities (cis. women, non-binary people, transgender women, and transgender men; African American, Asian American, and Hispanic/Latinx American people) differ from other articles. In addition to identifying suspect social biases, our results show that failing to control for covariates can result in different conclusions and veil biases. Our contributions include methodology that facilitates further analyses of bias in Wikipedia articles, findings that can aid Wikipedia editors in reducing biases, and a framework and evaluation metrics to guide future work in this area.