Max Pellert, Simon Schweighoferr and David Garcia from CSH Vienna will talk on “Analysing affective dynamics through sentiment in social media status updates”.
The talk is part of the session “Emotions in and by Groups” at the ISRE 2o19 conference, which takes place in Amsterdam from July 10 to 13, 2019.
Abstract
Quantifying the temporal dynamics of emotions is important to understand the role of affect in wellbeing. The way emotional states change is commonly analyzed through self-reports (for example in Kuppens, Oravecz, & Tuerlinckx, 2010), gathering temporal sequences of emotions over periods of few days. However, self-reports pose limits to the length of the observation period and to the amount of participants that can be included in a study. The analysis of social media data poses an alternative way to capture affective dynamics at longer timescales and in large samples of individuals.
We studied a dataset of more than 22 Million Facebook status updates donated by more than 150,000 individuals, spanning Social Sharing of Political Emotions in Online Populist Communications Analysing affective dynamics through sentiment in social media status updates observation periods of 1.5 years on average. We applied sentiment analysis and computerized psycholinguistic methods to quantify expressed emotions in terms of valence and arousal, and analysed the resulting trajectories of emotions through a dynamical system model. Our results confirm the existence of an affective baseline of positive valence and neutral arousal that attracts emotional states. We quantified further the attractor strength towards this baseline and the affective variability of emotions across individuals. These results show how observational large-scale analyses can provide alternative evidence to traditional experimental and survey methods, providing a new way to test hypotheses in affective science.
Kuppens, P., Oravecz, Z., & Tuerlinckx, F. (2010). Feelings change: Accounting for individual differences in the temporal dynamics of affect. Journal of Personality and Social Psychology, 99(6), 1042–1060. https://doi.org/10.1037/a0020962