Elamur Research⌕
☰
home / papers / 2026 / Semantic Factor Analysis: Validating Personality Structure Recovery from empirically-mediated Word Embeddings

Semantic Factor Analysis: Validating Personality Structure Recovery from empirically-mediated Word Embeddings

Summary

The paper asks whether Big Five personality structure can be recovered from word embeddings in small samples, where classic psychometric methods break down. It proposes Semantic Factor Analysis (SFA). The Word2Vec vectors (Google-News-300) for the 50 IPIP-50 marker adjectives from Goldberg (1992) are each multiplied by the mean participant rating of the matching IPIP-50 statement. The vectors are then standardized, reduced to five components with PCA, and grouped with K-means (k=5). Factor purity measures how often adjectives land in their theoretical dimension.

Fifty-five participants completed IPIP-50 and the 18-item HSP-R sensitivity scale. They then rated three multimedia stimuli: two controversial news items and one advertisement. Weighted SFA recovered all five traits with 62.5% average purity, compared with 52.0% for an unweighted Word2Vec baseline. Standard CFA showed factor collapse and poor fit. Latent Class Analysis produced data-driven cutoffs and supported a three-group Environmental Sensitivity model. The results suggest that combining embeddings with modest survey data can aid personality research when CFA is infeasible. They also lend computational support to the lexical hypothesis.

Key points

  • SFA pipeline: Google-News-300 Word2Vec (300 dimensions, 3M-word vocabulary) → element-wise weighting by item means → z-scoring → PCA to 5 dimensions → K-means (k=5) → factor purity. The first five components explain only 30.1% of the variance.
  • Weighted SFA reached 62.5% average purity, compared with 52.0% for the unweighted baseline. The paper calls this a 10% relative gain, though the raw gap is about 10.5 percentage points. The baseline replaced four missing adjectives, for example untalkative→silent and short-tempered→cranky.
  • CFA was fit in R lavaan with maximum likelihood on n=55 responses. It suffered factor collapse and failed the conventional CFI, TLI, RMSEA and SRMR fit criteria.
  • HSP-R scores were related to the Big Five with R²=0.464, close to the large-scale value of 0.44 reported by Pluess et al. (2024). The LCA thresholds of 61.2% and 76.5% separate Dandelions, Tulips and Orchids.
  • Environmental Sensitivity predicted emotional reactions to controversial content but not purchase likelihood. Highly open participants resisted the advertisement rather than responding to it.
  • Limitations include:
    • a small, young, mostly Western English-speaking sample;
    • single-item multimedia measures;
    • no individual-level scores;
    • dependence on one embedding model.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages · History & humanities

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.latechclfl-1.17

@inproceedings{muller2026semantic,
  title = {{Semantic Factor Analysis: Validating Personality Structure Recovery from empirically-mediated Word Embeddings}},
  author = {Müller, Oliver},
  booktitle = {Proceedings of the 10th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature 2026},
  year = {2026},
  doi = {10.18653/v1/2026.latechclfl-1.17},
  url = {https://doi.org/10.18653/v1/2026.latechclfl-1.17}
}

References

22 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Gordon W Allport and Henry S Odbert. 1936. Trait-names: A psycho-lexical study. Psychological Monographs, 47(1):i–171. find
  2. Elaine N Aron and Arthur Aron. 1997. Sensory-processing sensitivity and its relation to introversion and emotionality. Journal of Personality and Social Psychology, 73(2):345–368. find
  3. Vincenzo Bove, Georgios Efthyvoulou, Armine Ghazaryan, and Harry Pickard. 2024. The emotional effect of terrorism. Scientific Reports, 14(1):26525. find
  4. Giulio Carducci, Giuseppe Rizzo, Diego Monti, Enrico Palumbo, and Maurizio Morisio. 2018. Twitperson-ality: Computing personality traits from tweets using word embeddings and supervised learning. Information, 9(5):127. find
  5. Andrew L Comrey. 1973. A first course in factor analysis. Academic Press, New York. find
  6. Colin G. DeYoung, Jordan B. Peterson, and Daniel M. Higgins. 2002. Higher-order factors of the big five predict conformity: Are there social determinants of personality? Journal of Personality and Social Psychology, 83(2):533–552. find
  7. Colin G DeYoung, Lena C Quilty, and Jordan B Peterson. 2007. Between facets and domains: 10 aspects of the big five. Journal of Personality and Social Psychology, 93(5):880–896. find
  8. Mirta Galesic and Michael Bošnjak. 2009. Effects of questionnaire length on participation and indicators of response quality in a web survey. Public Opinion Quarterly, 73(2):349–360. find
  9. Lewis R Goldberg. 1992. The development of markers for the big-five factor structure. Psychological Assessment, 4(1):26–42. find
  10. Listou H Grimen and Åse Diseth. 2016. Sensory processing sensitivity: Factors of the highly sensitive person scale and their relationships to personality and subjective health complaints. Perceptual and Motor Skills, 123(3):637–653. find
  11. Judith R Homberg, Daniel Schubert, Esther Asan, and Elaine N Aron. 2016. Understanding the molecular mechanisms underlying environmental sensitivity. Neuroscience & Biobehavioral Reviews, 71:476– 490. find
  12. Amirmohammad Kazameini, Samin Fatehi, Yash Mehta, Sauleh Eetemadi, and Erik Cambria. 2020. Personality Trait Detection Using Bagged SVM over BERT Word Embedding Ensembles. arXiv preprint arXiv:2010.01309. — [arXiv:2010.01309](https://arxiv.org/abs/2010.01309) find
  13. Ema Kusen and Mark Strembeck. 2023. Investigating the impact of personality traits on emotional communication in social media: A study on twitter users during crisis events. Computers in Human Behavior, 141:107623. find
  14. Francesca Lionetti, Elaine N Aron, Arthur Aron, Denise N Klein, and Michael Pluess. 2018. Dandelions, tulips and orchids: Evidence for the existence of low-sensitive, medium-sensitive and high-sensitive individuals. Translational Psychiatry, 8(1):1–11. find
  15. Robert R McCrae and Oliver P John. 1992. An introduction to the five-factor model and its applications. Journal of Personality, 60(2):175–215. find
  16. Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781. find
  17. Oliver Müller and Stefania Degaetano-Ortlieb. 2025. Embedded personalities: Word embeddings and the “big five” personality model. In Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2025), pages… find
  18. Gregory Park, H Andrew Schwartz, Johannes C Eichstaedt, Margaret L Kern, Michal Kosinski, David J Stillwell, Lyle H Ungar, and Martin E P Seligman. 2015. Automatic personality assessment through social media language. Journal of Personality and Social Psychology, 108(6):934–952. find
  19. Michael Pluess, Francesca Lionetti, Elaine N Aron, and Arthur Aron. 2024. The highly sensitive person scale-revised: Psychometric properties and recommendation for use. Assessment, 31(2):289–304. find
  20. Yves Rosseel. 2012. lavaan: An r package for structural equation modeling. Journal of Statistical Software, 48(2):1–36. find
  21. H Andrew Schwartz, Johannes C Eichstaedt, Margaret L Kern, Lukasz Dziurzynski, Stephanie M Ramones, Megha Agrawal, Achal Shah, Michal Kosinski, David Stillwell, Martin EP Seligman, and 1 others. 2013. Personality, gender, and age in the language of social media: The open-vocabulary approach. PloS on… find
  22. Ernest C Tupes and Raymond E Christal. 1961. Recurrent personality factors based on trait ratings. Journal of Personality, 60(2):225–251. Original work published 1961. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/muller-2026-semantic-factor-analysis-validating-personality/