Khalid, Osama, and Padmini Srinivasan. "Style matters! Investigating linguistic style in online communities." Proceedings of the international AAAI conference on web and social media. Vol. 14. 2020.

This paper shifts the focus from content to linguistic style in the study of online communities, exploring whether distinct communities have identifiable styles. Employing a broad definition of style with 262 features, the study analyzes the linguistic characteristics of 9 online communities across 3 social media platforms covering politics, television, and travel. Findings reveal that communities do possess unique styles, with linguistic style serving as an excellent predictor of group membership, achieving an F-score of 0.952 and an accuracy of 96.09%. The study also indicates that while predictions based on style are statistically equivalent to those using content, they are more robust against reductions in training data.