Quantifying representational mismatch between narratives and structured reports in near-death experiences
Cristian Pulido, Francisco Gómez, Prejaas K.b. Tewarie, Roxane Hoyer, Steven Laureys
Study at a glance
AI-extracted from the abstract| Characteristics | Observational study comparing questionnaire-derived and narrative-derived labels |
|---|---|
| Sample size | 190 |
| Population | Near-death experience reports |
| Measures | NDE-C, LCI-R |
| Topics | Near-death experience |
| Key findings | Questionnaire-derived and narrative-derived labels for near-death experiences agreed systematically but incompletely, with stronger correspondence for experiential features (NDE-C macro-F1 = 0.62) than life changes (LCI-R = 0.52) and only partial correspondence for narrative tone versus self-reported valence (macro-F1 = 0.43). The authors argue that this disagreement reflects a structured property of reporting, which they call representational mismatch, rather than measurement or extraction error alone. |
Abstract
Near-death experiences (NDEs) are often studied using both structured questionnaires and narrative reports, yet these representations cannot be assumed to provide interchangeable representations of the same experience. Narratives allow flexible expression of different aspects of the experience, whereas questionnaires organize reports into predefined dimensions and retrospective evaluations. How disagreement between these formats should be quantified and interpreted remains unclear. This study operationalized this disagreement as representational mismatch by comparing questionnaire-derived and narrative-derived labels across 190 NDE reports. Narratives were coded using 12 large language models (LLMs), complemented by human-reference annotation, a lexical baseline, external sentiment benchmarks, and preprocessing sensitivity analyses. Correspondence was evaluated across emotional tone, experiential features (NDE-C), and post-experience life changes (LCI-R).Agreement was systematic but incomplete and differed across representational domains. Experiential features showed stronger correspondence than life changes (NDE-C macro-F1 = 0.62 [0.60, 0.65] vs. LCI-R = 0.52 [0.50, 0.54]). Narrative tone contained meaningful affective information but corresponded only partially with global self-reported valence (macro-F1 = 0.43 [0.41, 0.44]), with disagreement concentrated mainly in mixed and neutral narrative-tone assignments rather than direct polarity reversals. Standard sentiment benchmarks showed substantially higher correspondence (macro-F1 = 0.78 ± 0.16), making a purely general model-limitation explanation less plausible.These findings support interpreting cross-format disagreement as a structured property of reporting rather than measurement or extraction error alone. Representational mismatch provides a framework for identifying where different reporting formats converge, where they cease to be interchangeable, and how the elicitation format shapes what becomes observable in research on subjective experience.
Comparable studies
Other cross-sectional and survey studies on near-death experience, most cited first.
| Study | Year | Design | Participants |
|---|---|---|---|
| Fantasy Proneness Correlates With the Intensity of Near-Death Experience Individuals who have experienced a near-death experience (classical or NDE-like) and... | 2018 | Cross-sectional survey | n = 228 |
| A systematic analysis of distressing near-death experience accounts Individuals who had a near-death experience and wrote down their account | 2019 | Cross-sectional study | n = 123 |
| Reduced death threat in near-death experiencers Adults who had near-death experiences, came close to death without NDE, or never came... | 1992 | Cross-sectional study | n = 290 |
| Near-Death Experiences in a Psychiatric Outpatient Clinic Population Psychiatric outpatients | 2003 | Cross-sectional survey | n = 832 |
| Near-Death States Reported in a Sample of 50 Misusers Previous ketamine misusers who reported a ketamine-related near-death experience | 2010 | Cross-sectional survey | n = 50 |
Citations in the library
Cites 12
- Near-death experience in survivors of cardiac arrest: a prospective study in the Netherlands The Lancet December 1, 2001
- The Near-Death Experience Scale The Journal of Nervous and Mental Disease June 1, 1983
- Distressing Near-Death Experiences Psychiatry February 1, 1992
- Near-Death Experience as a Probe to Explore (Disconnected) Consciousness Trends in Cognitive Sciences January 22, 2020
- The Near-Death Experience Content (NDE-C) scale: Development and psychometric validation Consciousness and Cognition November 1, 2020
and 7 more in the library