Feedback dynamics in Politics: The interplay between sentiment and engagement

TL;DR

This work investigates whether politicians adapt the sentiment of future social-media messages in response to audience engagement, positing a closed-loop sentiment–engagement dynamic. It analyzes $1.58$ million MP tweets from the UK, Spain, and Greece in 2021 using a simple autoregressive model with $\hat{S}_{t+1} = \alpha S_t + \beta r^+_t + \gamma r^-_t$ and extends to individual-level parameters $\alpha_i, \beta_i, \gamma_i$, revealing inertia and differential sensitivity to positive versus negative feedback across parties. The study provides empirical evidence of feedback in political communication, highlights cross-country and cross-party differences (e.g., opposition more responsive to negative engagement, government more responsive to positive signals), and develops a theoretical analysis of the linear and nonlinear closed-loop dynamics, including saturation, fixed points, and potential oscillations. These findings offer a quantitative, control-oriented lens on adaptive online political discourse and point to implications for understanding polarization and for designing interventions to moderate negativity in social media.

Abstract

We investigate feedback mechanisms in political communication by testing whether politicians adapt the sentiment of their messages in response to public engagement. Using over 1.5 million tweets from Members of Parliament in the United Kingdom, Spain, and Greece during 2021, we identify sentiment dynamics through a simple yet interpretable linear model. The analysis reveals a closed-loop behavior: engagement with positive and negative messages influences the sentiment of subsequent posts. Moreover, the learned coefficients highlight systematic differences across political roles: opposition members are more reactive to negative engagement, whereas government officials respond more to positive signals. These results provide a quantitative, control-oriented view of behavioral adaptation in online politics, showing how feedback principles can explain the self-reinforcing dynamics that emerge in social media discourse.

Figures (7)

• Figure 1: Schematic representation of the feedback mechanism in political communication.
• Figure 2: Pairwise scatter plots between future sentiment $S_{t+1}$ and the current variables: sentiment $S_t$, positive retweets $r^+_t$, and negative retweets $r^-_t$.
• Figure 3: Prediction of future sentiment $S_{t+1}$ using current sentiment and engagement metrics. The model effectively captures sentiment trends despite variability.
• Figure 4: Z-scores of the difference between positive and negative retweet coefficients, grouped by party (UK 2021 dataset). Positive scores indicate greater influence of positive feedback; negative scores, greater influence of negative feedback.
• Figure 5: Z-scores of the difference between positive and negative retweet coefficients, grouped by party (Spain 2021 dataset).