The Role of Emotional Compared to Political Involvement in Attitude Polarization as Modeled by the Cusp

Introduction

The present study investigated the role of emotional compared to political involvement in the polarization of attitudes outlined by the cusp. Attitudes were thus modeled as a dynamic system consisting of sudden and non-linear interactions.

Methods

A sample of Twitter posts (n = 4000) was evaluated using domain-specific lexicons. Users' lexicon means (n = 2891) were encoded as the cusp's dependent (AFINN lexicon) and involvement (Political or NRC Af-fect Intensity lexicons) variables. Fitting the cusp depended on restriction tests and in-volvement types, resulting in restricted and unrestricted political and emotional models.

Results

All cusp models fitted better than linear models. Likewise, the restricted emotional and unre-stricted political models fitted the best. In line with the hypothesis, both elevated involve-ment types resulted in attitude polarization. Political involvement demonstrated more such polarization. Unexpectedly, emotional involvement was overall high and political involve-ment correlated with attitudes. The use of lexicons and Twitter data is therefore discussed.