Pythia
The science behind Pythia

The science of retention.

People leave jobs for many reasons, and personality is only one of them. But it's one you can see at the hiring stage, and decades of research show it matters. This page covers what the evidence says, and where its limits are.

Why people leave

Turnover research spans more than a century. Its central finding is that leaving is rarely sudden: it builds from how well the job matches what a person needs, how attached they become to the work and the people, and what alternatives they have (Griffeth, Hom & Gaertner, 2000; Hom, Lee, Shaw & Hausknecht, 2017). Many of those factors are set the moment you hire.

Personality and staying power

A meta-analysis of personality and turnover found that traits such as conscientiousness, agreeableness and emotional stability are linked both to intentions to quit and to actually leaving, while openness is linked slightly with leaving more often (Zimmerman, 2008). The effects are modest, which is why Pythia reports them as an outlook, not a prediction.

Fit matters as much as the person

The same person can thrive in one job and leave another within months. Research on person–job and person–organization fit shows that people whose traits and values match their role are more satisfied, more committed and less likely to plan to leave (Kristof-Brown, Zimmerman & Johnson, 2005). That's why Pythia compares every candidate with the role you describe, instead of scoring them in a vacuum.

The same holds for performance. Different roles activate different traits: conscientiousness matters most in structured, execution-heavy work, while extraversion and openness matter more in interpersonal roles (Barrick & Mount, 1991; Hurtz & Donovan, 2000; Judge, Bono, Ilies & Gerhardt, 2002). Across occupations, conscientiousness remains the most consistent predictor of how well people work (Wilmot & Ones, 2019; van Aarde, Meiring & Wiernik, 2025).

What else predicts who stays

Personality isn't the whole story. In a study built specifically to reduce avoidable turnover, the strongest signals came from things a good interview can bring out: how much the candidate wants this particular job, and whether they already know people at the company (Barrick & Zimmerman, 2005). Giving candidates an honest picture of the job before they accept also helps (Earnest, Allen & Landis, 2011).

The Big Five: the model we rely on

Pythia uses the Five-Factor Model of personality, commonly known as the Big Five: openness, conscientiousness, extraversion, agreeableness, and neuroticism (emotional stability). It is the dominant framework in personality psychology, with strong cross-cultural replication (McCrae & Costa, 2008; Goldberg, 1993). Unlike type-based instruments such as the MBTI, it measures traits on continuous spectra, and it has accumulated substantial evidence of reliability and predictive validity (Roberts, Kuncel, Shiner, Caspi & Goldberg, 2007). Looking within each trait, at its facets, sharpens those predictions further (Soto & John, 2017).

Why the interview

Structured interviews are among the best-validated tools for predicting job performance (Sackett, Zhang, Berry & Lievens, 2022). Recent work shows large language models can score Big Five traits from brief open-ended speech and text with accuracy comparable to established benchmarks (Wright et al., 2025). Pythia combines the two: the interview you already run, read by a model grounded in personality science.

Featured finding

Wright et al. (2025), Nature Human Behaviour — LLM-derived personality scores from short open-ended text converged with self-report measures and predicted daily behaviour and mental health, outperforming traditional closed-vocabulary text analysis.

This builds on a longer tradition of reading personality from language and behaviour: from social media posts (Park et al., 2015), digital footprints (Youyou, Kosinski & Stillwell, 2015), and smartphone use (Stachl et al., 2020).

How strong is the signal?

Personality-to-outcome correlations typically sit in the .10 to .30 range. That might sound modest, but these effect sizes are comparable to, and often larger than, those found for socioeconomic status or cognitive ability in predicting life outcomes (Roberts et al., 2007). They also replicate: Soto (2019) found that 87% of published Big Five–outcome links held up in a large preregistered study, though at about 77% of their original size — a useful reminder to be conservative about any single prediction.

How we use this responsibly

Pythia reports, and you decide. A profile never accepts or rejects anyone: it is one input to weigh alongside experience, references and your own judgement, as guidance from the American Psychological Association and the Society for Industrial and Organizational Psychology recommends. Scores are read in context, against the role, never as pass/fail. And we show a degree of confidence, not certainty.

References

  1. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44(1), 1–26.
  2. Barrick, M. R., & Zimmerman, R. D. (2005). Reducing voluntary, avoidable turnover through selection. Journal of Applied Psychology, 90(1), 159–166.
  3. Earnest, D. R., Allen, D. G., & Landis, R. S. (2011). Mechanisms linking realistic job previews with turnover: A meta-analytic path analysis. Personnel Psychology, 64(4), 865–897.
  4. Goldberg, L. R. (1993). The structure of phenotypic personality traits. American Psychologist, 48(1), 26–34.
  5. Griffeth, R. W., Hom, P. W., & Gaertner, S. (2000). A meta-analysis of antecedents and correlates of employee turnover. Journal of Management, 26(3), 463–488.
  6. Hom, P. W., Lee, T. W., Shaw, J. D., & Hausknecht, J. P. (2017). One hundred years of employee turnover theory and research. Journal of Applied Psychology, 102(3), 530–545.
  7. Hurtz, G. M., & Donovan, J. J. (2000). Personality and job performance: The Big Five revisited. Journal of Applied Psychology, 85(6), 869–879.
  8. Judge, T. A., Bono, J. E., Ilies, R., & Gerhardt, M. W. (2002). Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology, 87(4), 765–780.
  9. Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. (2005). Consequences of individuals' fit at work: A meta-analysis of person–job, person–organization, person–group, and person–supervisor fit. Personnel Psychology, 58(2), 281–342.
  10. McCrae, R. R., & Costa, P. T. (2008). The Five-Factor Theory of personality. In O. P. John, R. W. Robins, & L. A. Pervin (Eds.), Handbook of personality: Theory and research (3rd ed., pp. 159–181). Guilford Press.
  11. Park, G., Schwartz, H. A., Eichstaedt, J. C., Kern, M. L., Kosinski, M., Stillwell, D. J., Ungar, L. H., & Seligman, M. E. P. (2015). Automatic personality assessment through social media language. Journal of Personality and Social Psychology, 108(6), 934–952.
  12. Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Goldberg, L. R. (2007). The power of personality: The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science, 2(4), 313–345.
  13. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040–2068.
  14. Soto, C. J. (2019). How replicable are links between personality traits and consequential life outcomes? The Life Outcomes of Personality Replication Project. Psychological Science, 30(5), 711–727.
  15. Soto, C. J., & John, O. P. (2017). The next Big Five Inventory (BFI-2): Developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power. Journal of Personality and Social Psychology, 113(1), 117–143.
  16. Stachl, C., Au, Q., Schoedel, R., Gosling, S. D., Harari, G. M., Buschek, D., Völkel, S. T., Schuwerk, T., Oldemeier, M., Ullmann, T., Hussmann, H., Bischl, B., & Bühner, M. (2020). Predicting personality from patterns of behavior collected with smartphones. Proceedings of the National Academy of Sciences, 117(30), 17680–17687.
  17. van Aarde, N., Meiring, D., & Wiernik, B. M. (2025). Personality and job performance: A review of trait models and recent trends. Current Opinion in Psychology.
  18. Wilmot, M. P., & Ones, D. S. (2019). A century of research on conscientiousness at work. Proceedings of the National Academy of Sciences, 116(46), 23004–23010.
  19. Wright, A. G. C., et al. (2025). Assessing personality using zero-shot generative AI scoring of brief open-ended text. Nature Human Behaviour.
  20. Youyou, W., Kosinski, M., & Stillwell, D. (2015). Computer-based personality judgments are more accurate than those made by humans. Proceedings of the National Academy of Sciences, 112(4), 1036–1040.
  21. Zimmerman, R. D. (2008). Understanding the impact of personality traits on individuals' turnover decisions: A meta-analytic path model. Personnel Psychology, 61(2), 309–348.

Pythia scores the Big Five traits, and in Kaleidoscope their 30 facets, from the interview, and reports them against the role you describe. We continually review new research and update our approach accordingly. Questions about our methodology? Get in touch.