B6
Heuristics & BiasesNarrative Fallacy
Compelling stories should not override evidence
4
Models Tested
2
Confirmatory
0.605
Mean Effect
1.055
Max Effect
Theoretical Context
Theoretical Anchor
Taleb (2007)
Normative Violation
Compelling stories should not override evidence
Cross-Model Comparison
Effect sizes for Narrative Fallacy across all tested models
OpenAI
GPT-5.2
The Steady Traditionalist
h = +1.055Confirmatory
Anthropic
Claude Sonnet 4.6
The Cautious Contrarian
h = +0.707Confirmatory
OpenAI
GPT-5.3 Instant
The Directive Optimist
h = +0.385
OpenAI
GPT-5.4 Thinking
The Deliberative Calibrator
h = +0.274
Statistical Details
Full results with confidence intervals and sample sizes
| Model | n (A) | n (B) | Cohen's h | 95% CI | Status |
|---|---|---|---|---|---|
| GPT-5.2 | 50 | 50 | +1.0550 | — | Confirmatory |
| Claude Sonnet 4.6 | 50 | 50 | +0.7069 | — | Confirmatory |
| GPT-5.3 Instant | 50 | 50 | +0.3849 | — | Exploratory |
| GPT-5.4 Thinking | 50 | 50 | +0.2738 | — | Exploratory |