These are not the Stereotypes You are Looking For: Bias and Fairness in Authorial Gender Attribution
Abstract Stylometric and text categorization results show that author gender can be discerned in texts with relatively high accuracy. However, it is difficult to explain what gives rise to these results and there are many possible confounding factors, such as the domain, genre, and target audience of a text. More fundamentally, such classification efforts risk invoking stereotyping and essentialism. We explore this issue in two datasets of Dutch literary novels, using commonly used descriptive (LIWC, topic modeling) and predictive (machine learning) methods. Our results show the importance of controlling for variables in the corpus and we argue for taking care not to overgeneralize from the results.
Abstract Stylometric and text categorization results show that author gender can be discerned in texts with relatively high accuracy. However, it is difficult to explain what gives rise to these results and there are many possible confounding factors, such as the domain, genre, and target audience of a text. More fundamentally, such classification efforts risk invoking stereotyping and essentialism. We explore this issue in two datasets of Dutch literary novels, using commonly used descriptive (LIWC, topic modeling) and predictive (machine learning) methods. Our results show the importance of controlling for variables in the corpus and we argue for taking care not to overgeneralize from the results.
- Type of material
- Terms of use
- Target audience
- Subject areas
- Languages
- Media formats
- Accessibility features
- OER type
- Metadata and online reference
Submitted by
Anastassiya Kuzmina
05/06/2019
in the project Sociolinguistics and NLP
last updated 12/06/2019
- Evaluations
- No evaluation
Please log in to add evaluation.
No comments yet.
Please log in to leave a comment.