Modeling the Impact of Group Interactions on Climate-related Opinion Change in Reddit

TL;DR

This work tackles climate-related opinion dynamics on Reddit by introducing a time-varying hypergraph framework that captures high-order group interactions within conversational threads. It combines a nonlinear multi-body consensus model with a temporal hypergraph, enabling evaluation against ground-truth stance labels generated by GPT-3.5 Turbo and measured with C1 (first drift), C2 (timing), and C3 (final opinion) metrics. Across six subreddits and various cross-community configurations, the hypergraph-based diffusion consistently improves identification of initial opinion shifts and final opinions compared with dyadic graphs, though ground-truth data remain scarce and challenging to generalize to offline behavior. The findings underscore the importance of high-order interactions for modeling online opinion dynamics and point to the need for broader, richer ground-truth datasets to advance robust, real-world predictive capabilities.

Abstract

Opinion dynamics models describe the evolution of behavioral changes within social networks and are essential for informing strategies aimed at fostering positive collective changes, such as climate action initiatives. When applied to social media interactions, these models typically represent social exchanges in a dyadic format to allow for a convenient encoding of interactions into a graph where edges represent the flow of information from one individual to another. However, this structural assumption fails to adequately reflect the nature of group discussions prevalent on many social media platforms. To address this limitation, we present a temporal hypergraph model that effectively captures the group dynamics inherent in conversational threads, and we apply it to discussions about climate change on Reddit. This model predicts temporal shifts in stance towards climate issues at the level of individual users. In contrast to traditional studies in opinion dynamics that typically rely on simulations or limited empirical validation, our approach is tested against a comprehensive ground truth estimated by a large language model at the level of individual user comments. Our findings demonstrate that using hypergraphs to model group interactions yields superior predictions of the microscopic dynamics of opinion formation, compared to state-of-the-art models based on dyadic interactions. Although our research contributes to the understanding of these complex social systems, significant challenges remain in capturing the nuances of how opinions are formed and evolve within online spaces.

Figures (10)

• Figure 1: An example of a hypergraph with four hyperedges.
• Figure 2: A toy example of a Reddit discussion (left) and its representation as a hypergraph (right).
• Figure 3: The construction process of a time-varying conversational hypergraph. In this example, $D_{s\Delta} = [d_1,..., d_i,..., d_{i+6}]$, while $\widetilde{ \mathcal{D}}_{s\Delta} = \{d_{i+4}, d_{i+5}, d_{i+6} \}$.
• Figure 4: Average accuracy scores for the criterion C1 with standard deviation values.
• Figure 5: Avg. anticipated vs. postponed simulation intervals in identifying the first opinion drift. Hypergraph results are shown in blue, clique results in orange, and graph results in green.