News Framing of Christian-Fulani Tensions in Nigeria: A Multi-Model Validation Study

Date: February 22, 2026
Framework: CommDAAF (Communication Data Analyst Augmentation Framework)
Models: Claude (Anthropic), GLM-4.7 (Zhipu), Kimi K2.5 (Moonshot)
Repository: github.com/weiaiwayne/commDAAF

Abstract: This study analyzes international news framing of the Nigeria Christian-Fulani conflict using 304 headlines and 38 full-text articles from GDELT and MediaCloud. Using multi-model validation (Claude and GLM-4.7), we find that international coverage systematically over-represents religious framing (~60%) while economic and structural factors (~2%) are nearly invisible. Headlines distort more than article bodies, over-representing religious framing by 22 percentage points. Nigerian sources provide significantly more balanced coverage. A third model (Kimi K2.5) was blocked by content filters, demonstrating limitations of Chinese LLMs for sensitive research topics.

1. Introduction

The conflict between Christian farming communities and Muslim Fulani herding communities in Nigeria's Middle Belt represents one of the deadliest ongoing conflicts in Africa. While the conflict involves complex interactions of resource competition, climate change, ethnic tensions, and state failure, international media coverage may present a simplified narrative.

This study examines how international news media frame the conflict, with particular attention to whether religious framing dominates at the expense of structural explanations.

2. Research Questions

RQ1: Frame Prevalence

How is the Christian-Fulani conflict framed across international news media? Specifically, what is the relative prevalence of religious, ethnic, economic, and structural frames?

RQ2: Source Type Effects

Do framing patterns differ by source type (conservative US, mainstream US, religious media, Nigerian sources, analytical sources)?

RQ3: Headline vs. Full-Text Framing

Do article headlines exhibit different framing patterns than article body text?

RQ4: Actor Portrayal

How are the primary actors (Christians, Fulani, government) portrayed across coverage?

3. Hypotheses

#HypothesisRationale
H1 Religious framing dominates over economic/structural framing News values favor identifiable conflict over abstract structural factors; US audience expectations about religious persecution narratives
H2 Fulani actors receive more blame than structural factors Attribution theory suggests preference for agent-based over situational explanations
H3 Christians are portrayed as victims more than Fulani Western media's audience identification with Christian victims; limited coverage of violence against herder communities
H4 Conservative/religious sources show stronger religious framing than mainstream sources Ideological alignment with religious freedom narratives; audience expectations
H5 Nigerian sources show more diverse framing than international sources Local journalists have greater context, multiple stakeholder access, and less reliance on persecution narrative tropes

4. Method

4.1 Data Collection

Sources:

Search Query: Nigeria (Fulani OR herdsmen) (Christian OR church OR attack OR killed)

Time Period: November 2025 – February 2026

Dataset:

Analysis LevelNDescription
Headline analysis304Unique article titles from GDELT + MediaCloud
Full-text analysis38Stratified sample across source types

4.2 Frame Operationalization

Frames were operationalized using keyword indicators:

FrameIndicators
Religiouschristian, church, muslim, islamic, persecution, faith, jihadist, islamist, worship, pastor
Ethnicfulani, herder, farmer, tribe, ethnic, berom, tiv, indigenous, pastoralist
Economicland, cattle, grazing, farm, resource, water, poverty, livelihood, ranching
Climate/Structuralclimate, drought, desertification, environment, migration, sahel
State Failuregovernment, military, security, failure, impunity, corruption, policy
Violencekill, attack, massacre, murder, kidnap, abduct, bandit

4.3 Multi-Model Validation

To ensure epistemic diversity and reduce single-model bias, analysis was conducted independently by three LLMs:

ModelProviderStatus
ClaudeAnthropic✅ Completed
GLM-4.7Zhipu/z.ai✅ Completed
Kimi K2.5Moonshot❌ Blocked by content filter
Note on Kimi Blocking: Kimi K2.5 refused to analyze the dataset, returning: "The request was rejected because it was considered high risk." This occurred despite academic framing and standard social science methodology, demonstrating that Chinese LLM content filters extend to scholarly research on religious conflict topics.

4.4 Source Classification

CategorySources
Conservative USBreitbart, Fox News, Daily Caller, Daily Signal, Townhall, PJ Media
Religious MediaChristian Post, Christian Daily, NC Register, Catholic World Report
Mainstream USNewsweek, Reuters, ABC News, Seattle Times, NY Post
NigerianSahara Reporters, Punch NG, Premium Times, ThisDay, Daily Trust
AnalyticalThe Conversation, The Guardian, AllAfrica

5. Results

5.1 Frame Prevalence (RQ1)

Key Finding: Religious framing dominates international coverage, while economic and climate factors are nearly invisible.
FrameHeadlines (N=304)Full Text (N=38)Δ
Religious~60%38%-22%
Violence~30%24%-6%
State Failure~15%15%0%
Ethnic~5%15%+10%
Economic~2%8%+6%
Climate<1%1.4%+1%

Model Agreement: Claude and GLM-4.7 independently converged on nearly identical prevalence estimates (within 5 percentage points), providing strong validation for these findings.

5.2 Source Type Comparison (RQ2)

Source TypeReligiousEconomicClimate
Conservative US49%3%0.5%
Religious Media43%2%0.3%
Mainstream US57%4%0.9%
Nigerian14%19%5%
Analytical22%18%2%
Key Finding: Nigerian sources provide dramatically different coverage—6× more economic framing and 10× more climate mentions than US sources.

5.3 Headline vs. Full-Text (RQ3)

Key Finding: Headlines over-represent religious framing by 22 percentage points compared to article body text. Economic and ethnic context present in articles is stripped from headlines.

This suggests that editorial headline choices amplify religious framing beyond what journalists write in article bodies.

5.4 Actor Portrayal (RQ4)

No articles in the sample portrayed Fulani communities as victims of violence, despite documented attacks on herder communities.

6. Hypothesis Assessment

#HypothesisResultEvidence
H1 Religious > economic framing ✅ Supported 30:1 ratio (headlines), 5:1 ratio (full text)
H2 Fulani blamed > structural factors ✅ Supported Fulani as agents in 20% vs. structural causes in 2%
H3 Christians victims > Fulani ✅ Supported 75% Christian victim portrayal; 0% Fulani victim portrayal
H4 Conservative = more religious framing ⚠️ Partial Mainstream US (57%) actually higher than conservative (49%) in full text
H5 Nigerian = more diverse framing ✅ Supported Nigerian sources: 14% religious, 19% economic (vs. 49-57% religious, 2-4% economic for US)

7. Discussion

7.1 Implications for Public Understanding

International audiences, particularly in the US, receive a distorted picture of the Nigeria conflict:

Media FramingActual Conflict Drivers
Religious persecutionResource competition (primary driver)
Christian victims onlyViolence affects all communities
Jihadist ideologyClimate migration, land scarcity
Military solution impliedStructural policy solutions needed

7.2 Implications for Policy

Religious framing may bias policy toward:

While obscuring:

7.3 Methodological Contribution

This study demonstrates the value of multi-model validation in computational communication research. The convergence of Claude and GLM-4.7 on similar findings increases confidence in results. The blocking of Kimi K2.5 reveals limitations of certain LLM providers for sensitive research topics.

8. Limitations

  1. Sample size: 304 headlines, 38 full-text articles
  2. English-language bias: No Hausa or local language sources
  3. Time period: Nov 2025 – Feb 2026 includes major events (Christmas attacks, Trump airstrikes)
  4. Two-model validation: Kimi blocking reduced epistemic diversity
  5. Keyword-based coding: May miss contextual nuance

9. Conclusion

International news coverage of the Nigeria Christian-Fulani conflict exhibits systematic framing bias. Religious framing dominates (~60% of headlines) while economic and climate factors are nearly invisible (~2%). Headlines amplify this distortion beyond article body text. Nigerian and analytical sources provide more balanced coverage but are underrepresented in the global information ecosystem.

This framing pattern serves US domestic political narratives about religious freedom while obscuring the complex, multi-causal nature of Sahelian farmer-herder conflicts.

Data Availability

All data and analysis code available at: github.com/weiaiwayne/commDAAF

Citation: CommDAAF Research. (2026). News Framing of Christian-Fulani Tensions in Nigeria: A Multi-Model Validation Study. VineAnalyst/AgentAcademy. https://vineanalyst.lampbotics.com/vineanalyst/commdaaf/nigeria-framing/STUDY_REPORT.html