Abstrakt
Current migration crisis has brought a bi-polarization into public opinion in Western states. The opinion bi-polarization is sometimes explored through formal multi-agent models using a few basic theoretical mechanisms such as persuasive arguments, opinion fault lines or bounded confidence models.
One of these mechanisms is the negative influence model which can explain increasing polarization between groups. However, there is a lack of empirical research on negative influence in large groups on societal level.
Our study strives to fill in the gap by exploring the differences between polarization structure based on negative and positive ties. We use data from the biggest Czech news server iDNES.cz.
Specifically, we use automated data collection to download readers' comment below articles in the foreign issues section where topics of migration have been discussed. We focus on the year 2015 when the migration crisis culminated in Europe.
The advantage of our data compared to data from frequently used social media, such as Twitter or Facebook, is that users can rate each other's comments with both positive and negative votes (likes and dislikes). We trace how different network communities develop through several time points based on both positive and negative ties and with subsequent content analysis of opinions within the groups, we identify polarized subgroups of pro-western and pro-Russia debaters in the issue of Ukrainian conflict and anti-immigrant versus pro-immigrant debaters in the issue of migrant crisis.
Furthermore, we analyse internal characteristics of each group such as centralization, density or size. We would like to follow up on this research with a longitudinal exponential random graph model, where we model both positive and negative ties simultaneously to discover which network mechanisms brought about the polarized structure of the network.