CSH team leader David Garcia will give a keynote speach at the 7th International Conference on Computational Social Science IC2S2 2021.
Social media data has allowed us to quantify and track emotional expression across scales and for long periods of time. This is possible thanks to the application of text analysis methods that can be validated at the level of individual social media posts, for example against sentiment annotations. However, we still lack evidence showing whether aggregates of emotional expressions over time match the self-reported emotional experiences of the members of a community.
In this talk, I present a set of studies comparing the temporal evolution of large-scale aggregates of emotion on social media versus surveys of emotional experiences of social media users and representative panels. I will present a comparison of dictionary-based and supervised text analysis methods and show their potentials and limitations as macroscopes of emotional experiences. This helps us to understand the convergence of traditional and computational social science methods and how to quantify measurement error when studying social phenomena through digital traces.