Research by Hugi Hernandez, Founder of Egreenews
Research from 2025 to 2026 reveals that online hate and digital bullying targeting teens are shaped by a critical factor often overlooked in policy debates: emotional vulnerability drives content-seeking behavior, rather than passive algorithmic exposure alone. Longitudinal research from Chinese universities demonstrates that adolescents who experience cyber-ostracism are significantly more likely to develop hostile attribution biases and subsequently engage in online deviance, with this pathway being particularly pronounced among boys .
European research spanning six countries finds that perceived discrimination is a direct predictor of both intentional and unintentional exposure to cyberhate, while specific dimensions of digital literacy—particularly communication literacy—can serve as protective factors . A separate cross-national study involving Brazil, Germany, Italy, and Spain reveals significant national differences in cyberbullying intentions, suggesting that cultural context moderates the effectiveness of interventions .
The report identifies a persistent “awareness gap” documented in Hungary: while most young users can intuitively identify overt hate speech, their formal understanding of platform rules is minimal, eroding trust in governance systems . Key unknowns include the long-term effectiveness of current moderation systems, for which university researchers lack access to proprietary algorithmic data.
Introduction
For the average teenager in 2026, social media platforms like TikTok, YouTube, and Instagram are not optional accessories but primary environments for social development. The past 18 months of university research have moved the conversation beyond simple “screen time” concerns toward a more nuanced understanding of how emotional states, cognitive biases, and platform design interact to produce harmful outcomes.
A landmark study from the University of Navarra, surveying 2,020 Spanish children and adolescents aged 8 to 18, found that 29% of minors admit to feeling “hooked” on social media, with 18% reporting anxiety when unable to check their accounts. The same study documented that cyberbullying manifests most commonly as insults (18% of adolescents), teasing (11%), and the unauthorized sharing of images (7%) .
This report synthesizes peer-reviewed university research published between 2025 and May 2026 to answer a specific question: What do methodologically rigorous studies tell us about the mechanisms, regional patterns, and governance gaps in online hate and digital bullying targeting teens on platforms like TikTok and YouTube?
The evidence presented prioritizes longitudinal designs, cross-national surveys, and systematic reviews over anecdotal case studies. Where data is incomplete or regionally absent, that limitation is clearly stated.
Section 1: The Psychological Pathway from Exclusion to Deviance
The most significant mechanistic finding from the 2025-2026 period concerns the directional relationship between online social rejection and subsequent deviant behavior. A longitudinal study by Zhang, Liu, and Zhao (2026), published in Studies of Psychology and Behavior, followed 646 adolescents in China’s Liaoning province across three consecutive time points .
The results establish a clear causal direction: cyber-ostracism at the pretest significantly and positively predicted online deviant behaviors at the posttest, while the reverse relationship—deviant behavior predicting subsequent ostracism—was not supported. This finding refutes simple “bad kids seek bad content” explanations in favor of a social-pain model: exclusion drives deviance, not the other way around.
Critically, the mediating mechanism differed by gender. Boys who experienced cyber-ostracism were more likely to develop hostile attribution biases—interpreting neutral social cues as intentionally hostile—which then led to online deviance. Among girls, this mediating pathway was not statistically significant; instead, a reciprocal relationship between ostracism and deviance was observed, suggesting different psychological mechanisms operate across genders .
“Boys are more likely to stimulate hostility attribution after experiencing cyber-ostracism, and then produce online deviant behaviors, while this mediating mechanism is not significant among girls.” — Zhang et al., Shenyang Normal University, 2026
For parents and educators, this implies that interventions targeting emotional regulation and hostile attribution biases may be more effective for boys, while girls may require different approaches addressing bidirectional social dynamics.
Section 2: Digital Literacy as Both Shield and Risk Factor
A three-wave longitudinal study from Masaryk University (Czech Republic) and partner institutions across six European countries—Estonia, Finland, Germany, Italy, Poland, and Portugal—surveyed 2,660 adolescents aged 14 to 17 to examine how digital literacy dimensions affect intentional versus unintentional cyberhate exposure .
The findings challenge simplistic assumptions about digital skills. Technical and operational skills—the ability to use digital tools effectively—were associated with increased intentional cyberhate exposure. This suggests that digitally skilled adolescents may actively seek out transgressive content rather than merely stumbling upon it. Similarly, programming skills increased both intentional and unintentional exposure.
However, other literacy dimensions proved protective. Communication and interaction literacy—the ability to communicate appropriately and interpret social cues online—decreased intentional cyberhate exposure. Information navigation literacy—the ability to find, evaluate, and process information—decreased unintentional exposure. These findings have direct implications for curriculum design: teaching technical skills without accompanying social-emotional competencies may backfire.
The study also found that perceived discrimination directly predicted both types of cyberhate exposure, confirming that marginalized youth face elevated risk regardless of their digital competence .
Section 3: Cross-National Variation in Cyberbullying Intentions and Empathy
A cross-national experimental study from Universitat Pompeu Fabra (UPF) Barcelona, involving 332 adolescents from Brazil, Germany, Italy, and Spain, evaluated an empathy training designed to reduce cyberbullying intentions. While the training itself did not produce statistically significant reductions, the cross-national differences documented are themselves instructive .
Participants’ intentions to bully were significantly higher in Spain and Brazil than in Germany. Cognitive and affective empathy levels were significantly higher in the Italian sample than in the German sample. These findings suggest that cultural context moderates both baseline risk and the potential effectiveness of universal interventions.
The study’s null result for the empathy training is methodologically important: it demonstrates that brief, digitally delivered interventions may be insufficient to shift deeply embedded behavioral patterns. The authors call for improved training designs that account for cross-national differences in social norms and empathy expression .
Complementing this, a systematic review from the University of Louisville (2026) examined existing TikTok cyberbullying research and identified that repeated exposure to harmful content may normalize cyberbullying over time, especially when negative behavior receives positive social reinforcement (likes, shares) from peers. The review also highlights how anonymity features on TikTok can produce online disinhibition, where users feel comfortable saying things online they would not say in person .
Section 4: The Governance Gap – What Teens Don’t Know About Platform Rules
Perhaps the most policy-relevant finding comes from Széchenyi István University in Hungary, where researchers surveyed 301 young people aged 14 to 34 about their understanding of hate speech governance on Meta platforms (Facebook, Instagram). The study identifies what it terms an “awareness gap”: while a majority of young users can intuitively identify overt hate speech, their formal understanding of platform rules is minimal .
Only a small minority of respondents could accurately describe Meta’s actual policies on specific types of hate speech. Furthermore, users’ sanctioning preferences—what punishment they believed a given violation deserved—often diverged from the platform’s actual enforcement practices. This gap signals what the authors call a “structural weakness that erodes user trust.”
The Hungarian research frames this gap not merely as a knowledge deficit but as a symptom of a deeper legitimacy crisis in platform governance. Users respond to opaque, commercially driven, and unaccountable moderation systems with “digital resignation”—a rational withdrawal from engagement with governance mechanisms. Even under the European Union’s Digital Services Act (DSA) framework, the study finds that transparency and remedy mechanisms remain inadequate .
At the Chinese University of Hong Kong, the Gender Research Centre hosted an international conference in May 2026 addressing technology-facilitated violence. Professor Susanne Choi’s research, funded by Hong Kong’s Research Grants Council, focuses on online sexual harassment and cyber dating abuse across four East Asian cities. The conference highlighted that technology-facilitated violence is “highly concealed yet spreads rapidly,” disproportionately affecting young people and reshaping intimate relationship dynamics in the digital age .
| Finding | Mechanism | Population | University Source |
|---|---|---|---|
| Cyber-ostracism predicts online deviance | Hostile attribution bias mediates (boys only) | Chinese adolescents | Shenyang Normal/Liaoning Normal, 2026 |
| Technical skills increase cyberhate exposure | Active seeking of transgressive content | European teens (6 countries) | Masaryk University, 2025 |
| Communication literacy decreases exposure | Improved social cue interpretation | European teens (6 countries) | Masaryk University, 2025 |
| Bullying intentions higher in Spain/Brazil vs. Germany | Cultural variation in norms | Brazil, Germany, Italy, Spain | UPF Barcelona, 2025 |
| Awareness gap in platform governance | Digital resignation to opaque systems | Hungarian youth | Széchenyi István University, 2026 |
| TikTok exposure normalizes cyberbullying | Online disinhibition, social reinforcement | Global (systematic review) | University of Louisville, 2026 |
Section 5: Emerging Risks – Misinformation, Radicalization, and Tech-Facilitated Violence
Research from Newcastle University (UK, 2026) examines how young adults become vulnerable to exploitation by malicious actors online, particularly during “emotionally vulnerable life events.” The study, based on stakeholder workshops and co-design sessions with young adults aged 18 to 25, emphasizes the need for empirical research within enclosed online communities such as gaming voice channels, “where opinions can become radicalised, emotions intensified, and young adults desensitised” .
The Newcastle researchers advocate for harm-reducing tools that increase young people’s “individual agency”—equipping them with skills to recognize, assess, and address misinformation while simultaneously enhancing algorithmic and new media literacy. They stress the importance of “reciprocal interactions and collaboration between mainstream and marginalised communities” .
At University College London, a graduate module on “Online Extremism and Hate Crime” (SECU0067) codifies the current academic understanding of these phenomena. The module, led by Dr. Sandy Schumann, covers trends in online hate speech and harassment, how extremist actors use the internet for propaganda and recruitment, and the overlap between crime, terrorism, disinformation, and hate speech. Students evaluate both human-centered and technology-centered approaches to detection and countermeasures .
The UCL module’s existence as a formal academic offering in 2026 signals that online extremism targeting youth has matured from an emerging concern to an established field of scholarly inquiry. The curriculum explicitly addresses the “reciprocal relationship between online and offline hate crime and extremism”—a bidirectional dynamic that complicates simple “online causes offline harm” narratives .
At the Chinese University of Hong Kong, the conference on “Technology, Intimacy and Violence” featured keynote speeches on tech-facilitated gender-based violence and on digital misogyny in Japan and South Korea. The event produced a bilingual booklet titled “Digital Safety and Challenges – Comparative Studies of Four East Asian Cities: Cyber Dating Abuse and Online Sexual Harassment,” suggesting that practical educational resources are emerging alongside academic research .
A separate study from Griffith University in Australia, forthcoming in Technological Forecasting and Social Change (July 2026), examines TikTok cyberbullying, anxiety, depression, life satisfaction, and sex differences among young adults. While focused on young adults rather than younger teens, this research extends the gender analysis to the TikTok platform specifically—a notable contribution given TikTok’s dominance among teen users .
Summary of Known Unknowns
- Algorithmic transparency data: University researchers cannot access platform recommendation algorithms for independent audit. The Hungarian study explicitly identifies opacity as a “structural weakness” that erodes user trust, and no university has published a validated audit of TikTok’s or YouTube’s hate speech recommendation systems from 2025-2026 .
- African teen data: No university study from an African institution specifically examining TikTok or YouTube hate targeting teens was found in the search results. The nearest available coverage comes from the Newcastle University study and UPF cross-national study , neither of which includes African countries.
- Long-term outcomes of tech-facilitated violence: While CUHK’s conference addresses the prevalence of online sexual harassment and cyber dating abuse, longitudinal data on long-term psychological outcomes for victims remains limited.
- Effectiveness of current moderation: Studies can measure exposure (e.g., University of Navarra’s survey on what teens report seeing ) but cannot easily measure what was successfully filtered before reaching users. This creates a survivorship bias in the evidence base.
- Cross-platform persistence: No longitudinal study in the search results tracks the same teen cohort’s exposure to hate across TikTok, YouTube, Instagram, and Discord simultaneously to map how abuse migrates between platforms.
Methodology Note
This report synthesizes exclusively university-affiliated research published between January 1, 2025 and May 18, 2026. Sources include peer-reviewed journal articles from MDPI, Taylor & Francis (via DOI), and university research repositories; longitudinal studies from Chinese and European universities; experimental cross-national research; and systematic reviews from US universities. Priority was given to studies using quantitative methods (longitudinal cohorts, cross-national surveys, systematic reviews) over purely theoretical work.
Regarding geographic diversity: While the search included terms for African university research, no eligible study from Africa specifically measuring teen hate on TikTok or YouTube within the date range was identified in the search results. South America is represented via the UPF cross-national study which includes Brazil. Australia is represented via Griffith University . North America is represented via the University of Louisville .
Regarding the date range: The search results did not contain any source from 2023 or 2024. All sources identified are from 2025 or 2026, which is within the requested range but narrower than the full 2023-2026 window. This is noted as a search result limitation rather than an exhaustive literature claim.
Citation List
Willett, E. (University of Louisville, Department of Criminal Justice). (2026). TikTok usage and cyberbullying: a systematic review of existing studies to identify correlations. College of Arts & Sciences Senior Theses, Paper 357. https://ir.library.louisville.edu/honors/357ents to analyze information effectively.
Zhang, Y., Liu, Z., & Zhao, L. (Shenyang Normal University, Liaoning Normal University). (2026). The Associations Among Cyber-Ostracism, Hostile Attribution Bias, and Online Deviant Behaviors in Adolescents: A Longitudinal Study. Studies of Psychology and Behavior. https://psybeh.tjnu.edu.cn/EN/abstract/abstract2711.shtml
San Miguel, P., Sánchez-Blanco, C., & Arrese, Á. (University of Navarra). (2026). Three out of ten minors consider themselves addicted to social media. https://en.unav.edu/news/-/contents/2026/05/04/20260504comredessociales
Masaryk University et al. (2025). Adolescents’ intentional and unintentional cyberhate exposure in Estonia, Finland, Germany, Italy, Poland, and Portugal. Journal of Children and Media. https://doi.org/10.1080/17482798.2025.2480091
Borsting, J., Schwarze, V., Theophilou, E., Sanchez-Reina, J.R., Odakura, V., Taibi, D., Scifo, L., Fulantelli, G., Hernandez-Leo, D., & Eimler, S.C. (Universitat Pompeu Fabra Barcelona). (2025). An Empathy Training for Sensitizing Adolescents for Cyberbullying on Social Media: A Cross-National Study. International Journal of Bullying Prevention. https://www.upf.edu/ca/web/tide/publications
Zhang, V., Vlachokyriakos, V., Johnson, I., & Durrant, A. (Newcastle University). (2026). Understanding and responding to complex online harms: misinformation, fake news, and young adults. Frontiers in Computer Science, 8. https://eprints.ncl.ac.uk/311101
Choi, S. Y. P. (Chinese University of Hong Kong Gender Research Centre). (2026). CUHK Gender Research Centre hosts international conference on technology, intimacy and violence prevention. https://www.cpr.cuhk.edu.hk/en/press/cuhk-gender-research-centre-hosts-international-conference
Alavi, M., Hillier, S., & Clough, B. (Griffith University). (2026). Exploring the Nexus: TikTok cyberbullying, anxiety, depression, life satisfaction, and sex differences among young adults. Technological Forecasting and Social Change, 228. https://doi.org/10.1016/j.techfore.2026.124665
Széchenyi István University (Department of Modern Technology and Cyber Security Law). (2026). Digital Resilience and the “Awareness Gap”: An Empirical Study of Youth Perceptions of Hate Speech Governance on Meta Platforms in Hungary. Journal of Cybersecurity and Privacy, 6(1), 3. https://www.mdpi.com/2624-800X/6/1/3
Schumann, S. (University College London, Department of Security and Crime Science). (2026). Online Extremism and Hate Crime (SECU0067) Module Catalogue. https://www.ucl.ac.uk/module-catalogue/modules/online-extremism-and-hate-crime-SECU0067
