[Paper Review] False News On Social Media: A Data-Driven Survey

Francesco Pierri, Stefano Ceri|arXiv (Cornell University)|Feb 20, 2019

Misinformation and Its ImpactsSocial Sciences64 references22 citations

TL;DR

This data-driven survey synthesizes recent advances in detecting, characterizing, and mitigating false news on social media by analyzing textual, behavioral, and network features across diverse datasets. It identifies network dynamics, psychological biases, and scalable detection techniques as central to combating misinformation, with key contributions in modeling rumor diffusion and leveraging crowd-sourced fact-checking for real-world impact.

ABSTRACT

In the past few years, the research community has dedicated growing interest to the issue of false news circulating on social networks. The widespread attention on detecting and characterizing false news has been motivated by considerable backlashes of this threat against the real world. As a matter of fact, social media platforms exhibit peculiar characteristics, with respect to traditional news outlets, which have been particularly favorable to the proliferation of deceptive information. They also present unique challenges for all kind of potential interventions on the subject. As this issue becomes of global concern, it is also gaining more attention in academia. The aim of this survey is to offer a comprehensive study on the recent advances in terms of detection, characterization and mitigation of false news that propagate on social media, as well as the challenges and the open questions that await future research on the field. We use a data-driven approach, focusing on a classification of the features that are used in each study to characterize false information and on the datasets used for instructing classification methods. At the end of the survey, we highlight emerging approaches that look most promising for addressing false news.

Motivation & Objective

• To provide a comprehensive, data-driven review of recent research on false news detection, characterization, and mitigation on social media platforms.
• To identify and classify the key features—textual, behavioral, and network-based—used in false news detection studies.
• To evaluate the datasets and evaluation methodologies employed across studies to assess reproducibility and comparability.
• To highlight promising mitigation strategies, particularly those leveraging crowd-sourced fact-checking and active user engagement.
• To outline open challenges and future research directions, including interdisciplinary coordination and real-world deployment of detection systems.

Proposed method

• Systematically surveys 190+ studies on false news in social media, focusing on detection, characterization, and mitigation techniques.
• Classifies features used in detection into textual (e.g., linguistic cues), behavioral (e.g., sharing patterns), and network-based (e.g., propagation structure) categories.
• Analyzes datasets such as Rumors, Hoaxy, and Twitter-based collections to evaluate detection model performance and generalizability.
• Evaluates detection methods using metrics like precision, misinformation reduction, and exposure prevention, particularly in temporal and stochastic modeling frameworks.
• Examines active mitigation strategies, including algorithmic selection of stories for fact-checking (e.g., CURB and adaptive flagging models).
• Integrates insights from psychology (e.g., confirmation bias, naive realism) and social network science to inform detection and intervention design.

Experimental results

Research questions

• RQ1 What are the dominant textual, behavioral, and network-based features used to distinguish false news from true news in social media?
• RQ2 How do different datasets and evaluation protocols affect the performance and comparability of false news detection models?
• RQ3 To what extent can crowd-sourced fact-checking and user flagging effectively reduce the spread of misinformation?
• RQ4 What role do social bots, echo chambers, and network structures play in amplifying false news diffusion?
• RQ5 What are the key open challenges and future research directions for building scalable, real-world mitigation systems?

Key findings

• False news spreads faster and more widely than true news on social media, with network structure and social bots playing a critical role in amplification.
• Textual features such as emotional language and linguistic complexity are useful but insufficient alone; social and network features significantly improve detection accuracy.
• Crowd-sourced fact-checking mechanisms, such as those modeled in CURB and adaptive flagging algorithms, can reduce misinformation exposure with theoretical guarantees.
• Models that account for user reliability and adversarial behavior (e.g., spammer users) show improved robustness and performance in simulated environments.
• The lack of standardized gold-standard datasets and evaluation protocols hinders reproducibility and cross-study comparison in the field.
• Future detection systems must integrate insights from psychology, journalism, and computer science to build more effective, scalable, and real-world deployable solutions.