Every Colour You Are: Stance Prediction and Turnaround in Controversial Issues

TL;DR

This paper introduces a methodology to analyze stance dynamics and opinion changes in online debates, demonstrated through Twitter discussions on abortion in Spanish-speaking countries, revealing new expressive behaviors and demographic differences.

Contribution

It presents a novel methodology for measuring stance adoption and opinion change in online debates, applied to Twitter data on abortion, highlighting new expressive forms and demographic variations.

Findings

Colored emojis mirror physical stance expressions.

Opinions change over time even on controversial issues.

Demographic groups show different patterns of opinion change.

Abstract

Web platforms have allowed political manifestation and debate for decades. Technology changes have brought new opportunities for expression, and the availability of longitudinal data of these debates entice new questions regarding who participates, and who updates their opinion. The aim of this work is to provide a methodology to measure these phenomena, and to test this methodology on a specific topic, abortion, as observed on one of the most popular micro-blogging platforms. To do so, we followed the discussion on Twitter about abortion in two Spanish-speaking countries from 2015 to 2018. Our main insights are two fold. On the one hand, people adopted new technologies to express their stances, particularly colored variations of heart emojis ([green heart] & [purple heart]) in a way that mirrored physical manifestations on abortion. On the other hand, even on issues with strong opinions, opinions can change, and these changes show differences in demographic groups. These findings imply that debate on the Web embraces new ways of stance adherence, and that changes of opinion can be measured and characterized.