[Paper Review] #FailedRevolutions: Using Twitter to Study the Antecedents of ISIS Support

Walid Magdy, Kareem Darwish+1 more|arXiv (Cornell University)|Mar 9, 2015

Terrorism, Counterterrorism, and Political Violence4 citations

TL;DR

This study uses Arabic Twitter data to predict ISIS support or opposition with 87% accuracy by analyzing pre-ISIS social media behavior. It identifies frustration with failed Arab Spring uprisings as a key antecedent to support, while opposition is linked to support for rival rebel groups, regional regimes, and Shia sectarianism.

ABSTRACT

Within a fairly short amount of time, the Islamic State of Iraq and Syria (ISIS) has managed to put large swaths of land in Syria and Iraq under their control. To many observers, the sheer speed at which this "state" was established was dumbfounding. To better understand the roots of this organization and its supporters we present a study using data from Twitter. We start by collecting large amounts of Arabic tweets referring to ISIS and classify them into pro-ISIS and anti-ISIS. This classification turns out to be easily done simply using the name variants used to refer to the organization: the full name and the description as "state" is associated with support, whereas abbreviations usually indicate opposition. We then "go back in time" by analyzing the historic timelines of both users supporting and opposing and look at their pre-ISIS period to gain insights into the antecedents of support. To achieve this, we build a classifier using pre-ISIS data to "predict", in retrospect, who will support or oppose the group. The key story that emerges is one of frustration with failed Arab Spring revolutions. ISIS supporters largely differ from ISIS opposition in that they refer a lot more to Arab Spring uprisings that failed. We also find temporal patterns in the support and opposition which seems to be linked to major news, such as reported territorial gains, reports on gruesome acts of violence, and reports on airstrikes and foreign intervention.

Motivation & Objective

• To identify linguistic and behavioral antecedents of ISIS support among Arabic Twitter users prior to their overt expression of support.
• To investigate whether pre-ISIS social media activity can predict future support or opposition to ISIS with high accuracy.
• To uncover the underlying motivations behind ISIS support, particularly political grievances related to the Arab Spring.
• To analyze temporal and topical patterns in pro- and anti-ISIS Twitter discourse in response to major events.
• To examine the representativeness and limitations of Twitter data in studying militant group support, especially given account deletions and data access constraints.

Proposed method

• Collected 123 million Arabic tweets from ~57,000 users using Twitter Streaming and REST APIs, focusing on those mentioning ISIS.
• Classified users as pro- or anti-ISIS based on lexical cues: full names and 'state' descriptors indicate support, while abbreviations indicate opposition.
• Trained a supervised classifier on pre-ISIS tweet content (before first pro/anti-ISIS tweet) to predict future stance with 87% accuracy.
• Conducted temporal trend analysis to correlate spikes in pro- or anti-ISIS sentiment with major news events (e.g., territorial gains, beheadings, airstrikes).
• Performed hashtag and topic analysis on pre-ISIS content to identify distinguishing themes between supporters and opponents.
• Addressed data limitations by acknowledging missing data due to account deletions, privacy changes, and Twitter’s 3,200-tweet limit per user.

Experimental results

Research questions

• RQ1 Can pre-ISIS social media behavior predict whether a Twitter user will later express support or opposition to ISIS?
• RQ2 What linguistic and topical features in pre-ISIS tweets distinguish future supporters from opponents of ISIS?
• RQ3 How do major news events such as territorial gains or reports of violence correlate with spikes in pro- or anti-ISIS sentiment?
• RQ4 What are the primary political grievances that drive support for ISIS, as reflected in users’ pre-ISIS discourse?
• RQ5 How do the profiles and interests of ISIS supporters differ from those of opponents in the pre-ISIS period?

Key findings

• The study achieved a 87% accuracy in predicting future ISIS support or opposition using only pre-ISIS Twitter content.
• ISIS supporters were significantly more likely to reference failed Arab Spring uprisings in their pre-ISIS tweets, indicating political frustration as a key driver.
• Pro-ISIS sentiment spikes were temporally concentrated and often coincided with major news events such as territorial gains or reports of violence.
• Anti-ISIS sentiment was strongly linked to support for other Syrian rebel groups, existing Middle Eastern regimes, and Shia sectarian identities.
• The classifier’s high accuracy suggests that distinct behavioral and linguistic patterns in pre-ISIS discourse reliably signal future stance, even with data limitations.
• Despite data biases from account deletions and tweet limits, the core findings—especially the link between Arab Spring frustration and ISIS support—remain robust and consistent with prior research.