Can personality traits be measured analyzing written language? a meta-analytic study on computational methods
EntityUAM. Departamento de Psicología Básica; UAM. Departamento de Psicología Social y Metodología
10.1016/j.paid.2021.110818Personality and Individual Differences 177 (2021): 110818
SubjectsBig five; Computational models of language; Language; Meta-analysis; Personality; Psicología
Rights© 2021 The Authors
Esta obra está bajo una licencia de Creative Commons Reconocimiento-NoComercial-SinObraDerivada 4.0 Internacional.
In the last two decades, empirical evidence has shown that personality traits could be related to the characteristics of written language. This study describes a meta-analysis that synthesizes 23 independent estimates of the correlations between the Big Five major personality traits, and some computationally obtained indicators from written language. The results show significant combined estimates of the correlations, albeit small to moderate according to Cohen's conventions to interpret effect sizes, for the five traits (between r = 0.26 for agreeableness and neuroticism, and 0.30 for openness). These estimates are moderated by the type of information in the texts, the use of prediction mechanisms, and the source of publication of the primary studies. Generally, the same effective moderators operate for the five traits. It is concluded that written language analyzed through computational methods could be used to extract relevant information of personality. But further research is still needed to consider it as predictive or explanatory tool for individual differences
Files in this item
Google Scholar:Moreno Pérez, José David - Martínez-Huertas, José - Olmos Albacete, Ricardo - Jorge-Botana, Guillermo - Botella Ausina, Juan
This item appears in the following Collection(s)
Showing items related by title, author, creator and subject.
Enhancing computational semantic representations in distributional models of language for Automated Summary Evaluation Martínez-Huertas, José Angel