💻 computer science

Schadenfreude in the Digital Public Sphere: A cross-national and decade-long analysis of Facebook news engagement

This study analyzes a decade of Facebook engagement across the US, UK, and India to reveal that schadenfreude is a significant, patterned phenomenon in online discourse that is most prevalent in political contexts, varies by cultural and ideological affiliation, and exhibits distinct "power-licensed" versus "power-compensatory" dynamics depending on whether an outlet's aligned party is in government or opposition.

Original authors: Nouar Aldahoul, Hazem Ibrahim, Majd Mahmutoglu, Hajra Tarar, Muhammad Fareed Zaffar, Talal Rahwan, Yasir Zaki

Published 2026-03-20

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Nouar Aldahoul, Hazem Ibrahim, Majd Mahmutoglu, Hajra Tarar, Muhammad Fareed Zaffar, Talal Rahwan, Yasir Zaki

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the internet as a massive, global town square. In this square, people gather to watch the news, but instead of just listening, they are constantly shouting, laughing, or booing. This paper is a ten-year detective story about a very specific, slightly dark human habit that has become a huge part of how we talk online: Schadenfreude.

Schadenfreude is a fancy German word for "taking pleasure in someone else's misfortune." It's that little smirk you get when your rival trips, or the satisfaction you feel when a politician you dislike makes a huge mistake.

Here is what the researchers found, broken down into simple stories and analogies.

1. The "Smirk" vs. The "Sigh"

The researchers looked at millions of posts on Facebook from the US, UK, and India. They wanted to know: When bad news happens, do people feel sad, angry, or secretly happy?

The Surface Level (The Emoji Button): When people see bad news, they mostly click "Sad" or "Angry." It's like seeing a car crash and gasping in horror.
The Deeper Level (The Comments): But when people actually type comments, the picture changes. While sadness is still there, a huge chunk of people are actually laughing, mocking, or gloating.
- The Analogy: Think of the "Sad" emoji as a polite nod at a funeral. The comments section is the gossip afterward, where some people are whispering, "Well, they deserved it," and laughing about it. The study found that laughing at others' pain is much more common in the comments than the "Haha" button suggests.

2. The "Who" and "Where" Matters

The study found that who you are and where you live changes how much you enjoy seeing others fail.

The "India Effect": People in India showed the highest levels of this "gloating." If a rival politician or celebrity messed up, Indian audiences were the most likely to turn it into a comedy show.
The "Right vs. Left" Game:
- Right-leaning audiences (conservatives) generally laughed more at misfortunes than left-leaning audiences.
- The "Power" Twist: This is the most interesting part. The researchers found two different "modes" of gloating:
  - The "Power-Licensed" Gloat (Left-leaning): When the Left is in charge, they feel emboldened to laugh at their opponents. It's like a team that just won the championship feeling free to tease the losing team.
  - The "Power-Compensatory" Gloat (Right-leaning): When the Right is out of power, they laugh even harder at their opponents' failures. It's like a team that lost the game but is cheering wildly when the other team trips over their own shoelaces. They use laughter to cope with feeling powerless.

3. What Kind of Bad News Gets the Laughs?

Not all bad news gets the same reaction. The researchers found that the type of misfortune matters:

The "Moral" Jackpot: When someone breaks a moral rule, gets caught in a scandal, or a political rival fails, the laughter is loudest. It feels like "justice" to the audience.
The "Nature" Bummer: When a hurricane hits or a sports team loses a game, people are much less likely to laugh. They feel sad or neutral.
The Analogy: If a corrupt CEO gets fired, people might high-five each other online. But if a random person slips on a banana peel, people just feel bad. The internet loves to laugh when it feels like the "bad guys" are getting punished.

4. The Digital Amplifier

Why is this happening now more than ever? The paper suggests that social media is like a giant microphone and a spotlight.

In the past, if you felt a little bit of joy when your enemy failed, you might keep that to yourself because it feels rude.
On Facebook, that feeling is turned into a public performance. You can type a joke, share a meme, or comment with a smirk. The more you do it, the more the algorithm shows it to others, creating a feedback loop where laughing at pain becomes a normal, even celebrated, way to participate in politics.

The Big Takeaway

This study tells us that online political fighting isn't just about arguing facts; it's about emotions.

We are seeing a world where "taking pleasure in others' pain" has become a routine part of how we consume news. It's not just a few trolls; it's a structured, predictable pattern. Depending on your political team and whether you are winning or losing, you might be more likely to turn a tragedy into a punchline.

In short: The internet has turned the town square into a place where, when the other team stumbles, a significant portion of the crowd doesn't just help them up—they pull out their phones to record it and laugh. And the study shows that this behavior is deeply tied to our politics and our sense of power.

1. Problem Statement

The paper addresses the lack of empirical understanding regarding schadenfreude (pleasure derived from others' misfortunes) in the digital public sphere. While existing literature treats schadenfreude as a socially corrosive force or a response to social comparison, its prevalence, dynamics, and social patterning in online news engagement remain unquantified.

Context: Social media platforms lower barriers to publicizing private emotions, transforming schadenfreude from a private feeling into a performative act of identity signaling.
Gap: There is a need for a systematic, longitudinal analysis of how schadenfreude manifests across different political ideologies, national contexts, and over time, specifically distinguishing between superficial reactions (e.g., emojis) and deeper verbal engagement (comments).

2. Methodology

The study employs a mixed-methods approach combining large-scale data collection, human annotation, and machine learning classification.

A. Data Collection

Scope: A ten-year dataset (2015–2024) from verified Facebook pages of major news outlets in three countries: USA, UK, and India.
Sampling Strategy: Three outlets per country representing distinct political leanings:
- Left-leaning: The Guardian (UK), Indian Express (India), NBC News (USA).
- Centrist: Reuters (UK), Hindustan Times (India), CNBC (USA).
- Right-leaning: Metro (UK), OpIndia.com (India), Fox News (USA).
Volume: Up to 1,000 posts per month per outlet, including post text, timestamps, public comments, and reaction metadata (e.g., "Haha," "Sad," "Angry").
Privacy: All personally identifying information (PII) was removed; only publicly visible content was analyzed.

B. Annotation and Classification Pipeline

The authors utilized a hybrid human-AI pipeline to identify misfortune and classify emotional responses:

Misfortune Detection:
- Human Annotation: 500 posts were labeled by three annotators (Inter-annotator agreement $\kappa = 0.81$ ) to define "misfortune" (harm, scandal, failure, loss).
- LLM Classification: GPT-4o Mini was fine-tuned to classify posts as "misfortune" or "normal," achieving 94.8% accuracy, 88.9% precision, and 87.3% recall.
Comment Classification (Schadenfreude Operationalization):
- Definition: Schadenfreude was operationalized as the proportion of comments classified as "toxic" (insults, mockery, gloating, derisive language) in response to misfortune posts.
- Training Data: 1,000 post-comment pairs were annotated into four categories: Toxic, Sympathetic, Neutral, Unknown.
- Model Performance: A fine-tuned GPT-4o Mini model outperformed simple prompting strategies, achieving 87.3% accuracy and an F1-score of 87.9% on the test set.
Topic Analysis:
- Misfortune posts were categorized into eight domains (e.g., Moral/Ideological, Political, Economic, Natural/Environmental) using GPT-5.2 to analyze contextual triggers.

C. Statistical Analysis

Comparative Analysis: T-tests were used to compare schadenfreude rates across ideologies and countries.
Regression Modeling: Ordinary Least Squares (OLS) regression was applied to US data to model schadenfreude rates as a function of outlet ideology, topic, and alignment with the governing party.

3. Key Contributions

Systematic Quantification: Provides the first large-scale, longitudinal quantification of schadenfreude in news discourse, moving beyond anecdotal evidence.
Methodological Innovation: Demonstrates a robust pipeline for detecting complex emotional states (schadenfreude) in unstructured text using fine-tuned LLMs, distinguishing it from general toxicity.
Cross-National & Ideological Mapping: Maps the variance of schadenfreude across three distinct political cultures (US, UK, India) and the political spectrum (Left, Center, Right).
Theoretical Framework: Introduces the concepts of "power-licensed" vs. "power-compensatory" schadenfreude to explain how political power dynamics influence emotional expression.

4. Key Results

A. Prevalence and Emotional Landscape

Dominant Emotions: Sadness and Anger dominate reactions to misfortune posts.
Schadenfreude Frequency: While "Haha" reactions account for ~15% of engagement, toxic comments (schadenfreude) are significantly higher, accounting for 28–42% of all comments on misfortune posts.
Underestimation by Emojis: Emoji reactions substantially underestimate the frequency of schadenfreude compared to verbal comments.

B. Cross-National and Ideological Variations

National Differences: India exhibits the highest rates of schadenfreude (approx. 42% of comments), followed by the USA (32%) and UK (28%). Indian audiences respond with humor/derision more frequently than with anger compared to Western counterparts.
Ideological Differences: Right-leaning audiences consistently show the highest mean schadenfreude rates (~38%), followed by Centrist (34%) and Left-leaning (31%).

C. Contextual Triggers (Topics)

High-Trigger Topics: Moral/Ideological and Political/Religious/Institutional misfortunes elicit the highest schadenfreude (often >40%, reaching 54% in right-leaning Indian outlets). This suggests a link to justice-based and intergroup pleasure.
Low-Trigger Topics: Natural/Environmental and Sports misfortunes elicit the lowest rates (<20%), lacking the moral agency or group antagonism required to fuel schadenfreude.

D. Temporal Dynamics and Political Power (US Case Study)

Asymmetric Power Dynamics: Regression analysis reveals distinct patterns based on political power:
- Left-leaning outlets: Exhibit "power-licensed" schadenfreude. Rates increase when their aligned party is in power, suggesting political dominance emboldens gloating.
- Right-leaning outlets: Exhibit "power-compensatory" schadenfreude. Rates decrease when their party governs and intensify when in opposition, suggesting schadenfreude serves as a coping mechanism for political exclusion.

5. Significance and Implications

Affective Polarization: The study confirms that polarization is not just cognitive (selective exposure) but deeply emotional. Users derive pleasure from out-group failures, reinforcing group boundaries and inhibiting empathy.
Platform Design: Features like reaction buttons and engagement-based ranking algorithms convert fleeting feelings into public, quantifiable signals, legitimizing derision as a form of political participation.
Cultural Mediation: While the capacity for schadenfreude is universal, its public expression is culturally mediated. The study highlights how local media ecosystems and political cultures shape the "emotional tenor" of public discourse.
Future Research: The findings suggest a need to investigate how regulatory environments and linguistic norms further shape these emotional boundaries in digital spaces.

Limitations

The authors acknowledge limitations including potential data loss due to moderation/deletion, the difficulty of LLMs in detecting sarcasm/irony, the observational nature of the study (precluding causal inference), and the bias toward highly engaged/polarized users who comment on news.