Sentiment Analysis of Clinical Interviews Identifies Language Markers of Identity Diffusion in Adolescents with Personality Disorder
Abstract
Identity diffusion – a fragmented sense of self and others maintained by splitting-based defenses – is a core feature of impaired personality functioning, theorized to involve heightened negative affectivity and affective instability. Evidence from other psychopathologies suggests that markers of affective disturbances can be automatically extracted from spontaneous speech, yet the linguistic correlates of identity diffusion have rarely been examined – particularly in adolescents, the age group in which personality pathology assessment is most challenging but also most consequential for early intervention. Transcripts of the Structured Interview of Personality Organization (STIPO) for adolescents were analyzed in a sample of 72 German-speaking participants (50 patients, 22 healthy controls) spanning the full spectrum from consolidated identity to severe identity diffusion. An interpretable lexicon-based sentiment analysis framework combined word frequency measures (LIWC) with continuous word-level valence indices (VADER), capturing both the prevalence and the dynamics of emotional tone. Greater identity diffusion was associated with higher affective variability (r = .48–.55), a more negative overall affective orientation (strongest effect: r = −.62), more frequent negative affective language, and more extreme negative expression, with weaker evidence for reduced positive language. Measures of positive extremes, affective expressiveness, and verbosity showed no significant associations. Identity diffusion in adolescents is thus associated with a distinctive affective language profile, detectable from clinical interview transcripts using sentiment analysis. The results are consistent with psychodynamic accounts of splitting and affective dysregulation, and support the potential of automated language analysis as a reproducible, scalable complement to clinical assessment.
Related
- Personality Inference From Semistructured Interviews Using Large Language Models: Comparative Evaluation Study
- Personality structure mediates the impact of childhood adversity on psychopathology in adolescents with personality disorder.
- Exploring the Feasibility of a Generic Early Intervention Programme for Emerging Personality Disorder in Adolescents: Guideline-Informed Treatment for Personality Disorders-Adolescent Version (GITPD-A).
- Improved Sleep and Clinical Outcomes Following a Single-Session Intervention among Adolescents with Borderline Personality Disorder Features
- Interoception and borderline personality disorder in female adolescents with restrictive eating disorders
BibTeX
@misc{lodrant2026sentiment,
title = {{Sentiment Analysis of Clinical Interviews Identifies Language Markers of Identity Diffusion in Adolescents with Personality Disorder}},
author = {Katarina Lodrant and Filip Melinscak and Ayse Nur Beris and Valentin Schneider and Klara Czernin and Anselm Bründlmayer and Paul Plener and Frank Scharnowski and David Steyrl and Clarissa Laczkovics},
howpublished = {PsyArXiv},
year = {2026},
month = oct,
doi = {10.31234/osf.io/dqbz2_v1},
url = {https://osf.io/preprints/psyarxiv/dqbz2_v1/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Sentiment Analysis of Clinical Interviews Identifies Language Markers of Identity Diffusion in Adolescents with Personality Disorder
AU - Katarina Lodrant
AU - Filip Melinscak
AU - Ayse Nur Beris
AU - Valentin Schneider
AU - Klara Czernin
AU - Anselm Bründlmayer
AU - Paul Plener
AU - Frank Scharnowski
AU - David Steyrl
AU - Clarissa Laczkovics
PY - 2026
DA - 2026/10/07
PB - PsyArXiv
DO - 10.31234/osf.io/dqbz2_v1
UR - https://osf.io/preprints/psyarxiv/dqbz2_v1/
AB - Identity diffusion – a fragmented sense of self and others maintained by splitting-based defenses – is a core feature of impaired personality functioning, theorized to involve heightened negative affectivity and affective instability. Evidence from other psychopathologies suggests that markers of affective disturbances can be automatically extracted from spontaneous speech, yet the linguistic correlates of identity diffusion have rarely been examined – particularly in adolescents, the age group in which personality pathology assessment is most challenging but also most consequential for early intervention. Transcripts of the Structured Interview of Personality Organization (STIPO) for adolescents were analyzed in a sample of 72 German-speaking participants (50 patients, 22 healthy controls) spanning the full spectrum from consolidated identity to severe identity diffusion. An interpretable lexicon-based sentiment analysis framework combined word frequency measures (LIWC) with continuous word-level valence indices (VADER), capturing both the prevalence and the dynamics of emotional tone. Greater identity diffusion was associated with higher affective variability (r = .48–.55), a more negative overall affective orientation (strongest effect: r = −.62), more frequent negative affective language, and more extreme negative expression, with weaker evidence for reduced positive language. Measures of positive extremes, affective expressiveness, and verbosity showed no significant associations. Identity diffusion in adolescents is thus associated with a distinctive affective language profile, detectable from clinical interview transcripts using sentiment analysis. The results are consistent with psychodynamic accounts of splitting and affective dysregulation, and support the potential of automated language analysis as a reproducible, scalable complement to clinical assessment.
ER -