Published June 2021 | Version v1
Thesis Open

Assessing Personality Traits Through Social Media Language

Creators

  • 1. University of Chicago

Contributors

Description

With the emergence of online social media, people increasingly leverage these platforms to build connections and express themselves. In particular, our language reflects the increasing emphasis on personal agency and expression. In this paper, we explored whether language can be a strong predictor of Big Five personality traits on Yelp, a site in which there is no expectation of building a socially desirable profile as do other social media. The linear models of lexical correlates suggest that users that produce high-quality reviews show higher Extraversion, Agreeableness, Openness, Conscientiousness, and lower Neuroticism. Feature importance suggests that low-quality reviews show a higher degree of egocentricity while high-quality reviews are more socially oriented.

Files

Chun Hu MA Thesis.pdf

Files (338.2 kB)

Name Size Download all
md5:c87a704949f82bee0f5e1a491b7b4da6
338.2 kB Preview Download

Additional details

Identifiers

Other
oai:uchicago.tind.io:2885

UChicago Information

Division(s)
Social Sciences Division
Department(s)
Computational Social Sciences (MACSS)