Subjectivity, Polarity and the Aspect of Time in the Evolution of Crowd-Sourced Biographies

This study examines the use of subjective and sentimentally charged language in crowd-sourced articles by focusing on time and how articles in environments like Wikipedia tend to evolve as edits are made by multiple contributors. More specifically, we measure linguistic subjectivity (the systematic, asymmetrical use of language) using Mean Abstraction Level, an established subjectivity measure, and polarity (the use of positively or negatively sentimentally charged words) through time. For the latest case, we introduce a new measure called Polarity Density. We focus on Wikipedia biographies and their evolution over time and we perform a detailed analysis per gender and per personality category. Our empirical evaluation provides evidence of increased subjectivity in female biographies as per personality category and per gender, while the same also occurs when considering sentimental charge over time.