OK Boomer: Probing the socio-demographic Divide in Echo Chambers

Abstract

Social media platforms such as Twitter or Reddit have become an integral part in political opinion formation and discussions, accompanied by potential echo chamber forming. In this paper, we examine the relationships between the interaction patterns, the opinion polarity, and the socio-demographic characteristics in discussion communities on Reddit. On a dataset of over 2 million posts coming from over 20k users, we combine network community detection algorithms, reliable stance polarity annotations, and NLPbased socio-demographic estimations, to identify echo chambers and understand their properties at scale. We show that the separability of the interaction communities is more strongly correlated to the relative socio-demographic divide, rather than the stance polarity gap size. We further demonstrate that the sociodemographic classifiers have a strong topical bias and should be used with caution, merely for the relative community difference comparisons within a topic, rather than for any absolute labeling.

Online Fake News Opinion Spread and Belief Change: A Systematic Review

“…HOMY is an important one. The results in [104] showed the effects of HOMY in the formation of echo chambers. It also demonstrated a moderate to high resemblance of the echo-chamber phenomenon for network topologies of abortion, capitalism, and feminism.…”

Stance Prediction from Multimodal Social Media Data

“…Dataset sizes are often within a few thousand instances, which is close to the text-only SemEval-2016 stance corpus in (Mohammad et al, 2016) with 4.2K labelled instances. We notice, however, that the two largest stance corpora in the present review -used in (Magdy et al, 2016) and (Geiss et al, 2022) -are not manually annotated at the text level, resorting to label propagation or similar methods instead.…”

“…Computational models of stance prediction, which often take social media text as an input, have been applied to a wide range of topics, including moral or social issues (Pavan et al, 2023;Geiss et al, 2022), politics (Darwish et al, 2017;Lehmann and Derczynski, 2019;Cignarella et al, 2020), and others. The task has become particularly popular in the field since the SemEval-2016 shared task (Mohammad et al, 2016) and accompanying corpus, focusing on stance prediction from Twitter text in the English language.…”

“…These include, for instance, the use of demographics information (e.g., users' gender, age, etc.) (Lehmann and Derczynski, 2019;Geiss et al, 2022), network structure (e.g., social media friends, followers, etc.) (Lai et al, 2020b), interactions (e.g., mentions, replies, retweets, etc.…”