Evidence of Complex Contagion of Information in Social Media: An Experiment Using Twitter Bots (1703.06027v1)

Abstract: It has recently become possible to study the dynamics of information diffusion in techno-social systems at scale, due to the emergence of online platforms, such as Twitter, with millions of users. One question that systematically recurs is whether information spreads according to simple or complex dynamics: does each exposure to a piece of information have an independent probability of a user adopting it (simple contagion), or does this probability depend instead on the number of sources of exposure, increasing above some threshold (complex contagion)? Most studies to date are observational and, therefore, unable to disentangle the effects of confounding factors such as social reinforcement, homophily, limited attention, or network community structure. Here we describe a novel controlled experiment that we performed on Twitter using `social bots' deployed to carry out coordinated attempts at spreading information. We propose two Bayesian statistical models describing simple and complex contagion dynamics, and test the competing hypotheses. We provide experimental evidence that the complex contagion model describes the observed information diffusion behavior more accurately than simple contagion. Future applications of our results include more effective defenses against malicious propaganda campaigns on social media, improved marketing and advertisement strategies, and design of effective network intervention techniques.

An Analysis of Complex Contagion Via Twitter Bots

The paper presents a paper on the dynamic nature of information spread within techno-social systems, focusing specifically on Twitter. The research examines whether information diffusion on social media platforms operates under simple or complex contagion models. In simple contagion, each exposure to a piece of information carries an independent probability of adoption, while complex contagion posits that the probability of adoption increases with multiple exposures from different sources after surpassing a threshold. The authors employ a controlled experimental setup using Twitter bots to rigorously investigate these competing hypotheses.

Experimental Design and Implementation

The researchers deployed a network of Twitter bots programmed to emulate human behavior convincingly. These bots disseminated information using hashtag campaigns to a geographically localized audience. The intervention involved specific hashtags, such as those promoting vaccinations and positive social interactions, ensuring these were novel introductions to the platform to control exposure accurately.

Several unique strategies ensured the bots' successful integration within the Twitter network: they exhibited human-like tweeting patterns, circulated content at non-uniform intervals, and maintained low follow-back ratios to appear authentic. The botnet accrued a significant follower base, gathering around 25,000 followers, of which a substantial number followed multiple bots, thus facilitating the paper of diverse exposure scenarios.

Theoretical Models and Analysis

The researchers introduced two Bayesian models to characterize the contagion processes. The simple contagion model relied on independent exposure probabilities, while the complex contagion model accounted for exposure dependence on the number of distinct sources. For the complex contagion model, the probability of adoption was modeled as a sigmoid function of the number of unique sources, reflecting a threshold mechanism.

The experimental results were analyzed using likelihood-based Bayesian Information Criterion (BIC) scoring to evaluate model fit while accounting for complexity. Across various simulations and parameter configurations, the complex contagion model consistently outperformed the simple contagion model, highlighting its relevance in describing information spread dynamics.

Implications and Future Directions

The findings underscore the importance of multi-source exposure in enhancing information adoption probabilities on social media. This revelation carries significant implications for fields ranging from marketing to public health messaging, where understanding the mechanisms of information spread can aid in designing more effective communication strategies.

The paper addresses a critical gap in social contagion research by providing conclusive experimental evidence supporting complex contagion dynamics at a large scale. Future research could explore exploring individual variations in susceptibility to social influence, potential effects of network structure anomalies, and further differentiating the factors driving complex contagion, such as social reinforcement or cognitive processing limitations.

Overall, this controlled examination of information diffusion through complex contagion contributes valuable insights into the understanding of social media dynamics, opening new avenues for optimizing information dissemination strategies across technologically mediated communication platforms.