On “New” Personality Types: An Industrial, Work, and Organizational Psychology Perspective

Deniz S. Ones, University of Minnesota-Twin Cities, and Brenton M. Wiernik, University of South Florida

To much media fanfare, a recent article presented findings proclaiming the discovery of four “personality types” (Gerlach et al. (2018, https://doi.org/10/gd8ds5).  We consider the implications for industrial-organizational psychology science and practice.  There is no compelling evidence overturning the scientific consensus that personality is composed of multiple continuous dimensions, not a limited set of discrete types.

Personality types are, again, in the limelight, having captured new attention from data scientists and the public. In September, Gerlach, Farb, Revelle, and Amaral (2018) published an article in Nature Human Behavior proclaiming they had identified four personality types that robustly replicated across 4 data sets. Its publication set off a media frenzy, with the authors’ university and major news outlets announcing this as a major breakthrough—for example, “Scientists identify four personality types” (Guarino, 2018), “Big data gives the ‘Big Five’ personality traits a makeover” (Smith, 2018), “These are the four big personality types, according to science.” (Ducharme, 2018), “Comprehensive data analysis dispels established paradigms in psychology” (Northwestern University, 2018a, 2018b), “These are the only personality types backed by science”(Stillman, 2018), among dozens of others (Daily Sabah, 2018; Dier, 2018; Duff, 2018; Holohan, 2018; e.g., Nauert, 2018).¹The publication of the Gerlach et al. article and the ensuing media frenzy had an immediate impact on us. First, as personality scholars, we were interested in the authors’ methods and conclusions. Second, we were left fielding emails, text messages, and phone calls from our collaborators, colleagues, scientists and practitioners, organizations, and test publishers. In this article, our goal is to assist other industrial-organizational psychologists make sense of this recent research and related developments.

To accomplish our goal, we accessibly describe Gerlach et al.’s methods and findings and how they fit with existing personality research. We also review the article’s relevance and importance for the I–O psychology community and address some frequently asked questions.

What were the goals of the Gerlach et al. article? Gerlach et al. sought to identify personality types in web-based self-report Big Five personality measures. They summarized their objective as gaining “insight into whether personality types truly exist” (p. 735).

What methods were used? Gerlach et al. used four publicly available datasets with responses from over 1.5 million individuals. They used an unsupervised machine learning clustering method (Gaussian mixture models) to identify clusters of people with similar Big Five scores. Traditional cluster-determination criteria suggested 13 clusters. However, clustering methods tend to produce spurious clusters, so Gerlach et al. compared the density of each of the 13 clusters (i.e., proportion of people belonging to each of the 13 clusters) with that of a null model based on randomized data. Only four clusters survived this density constraint. Two of these clusters were relatively consistent across the four datasets; the other two were observed in only three datasets.

What were the major findings? Using their newly developed clustering approach, Gerlach et al. identified 4 major clusters in five-dimensional (Big Five) personality space. These clusters represent combinations of traits that are more likely to be observed than others. The Average cluster includes people near the mean (mean z between ±0.60 standard deviations) on all Big Five personality dimensions, considered simultaneously. The Role Model cluster includes people low on Neuroticism (mean z = −0.70) but relatively high on most of the other Big Five (mean z = 0.78 Conscientiousness, 0.62 Agreeableness, 0.52 Extraversion). The Reserved cluster includes people low on Openness (mean z = −0.77) and Neuroticism (mean z = −0.54). The Self-Centered cluster includes people high on Extraversion (mean z = +0.70) but low on Openness (mean z = −0.75), Agreeableness (mean z = −0.56), and Conscientiousness (mean z = −0.44). Within each cluster, there was substantial heterogeneity in levels for each of the Big Five traits. As the relatively small differences in z values indicate, the four clusters were all quite close to each other in personality space, with the centers of all four clusters within ±1 standard deviations from the mean on all of the Big Five.

Were new personality types discovered? Personality’s multidimensional space is lumpy. The Big Five personality dimensions (and their narrower aspects and facets) are correlated to varying degrees (Davies, Connelly, Ones, & Birkland, 2015; DeYoung, 2015; Digman, 1997). The consequence of these intercorrelations is that some trait combinations are more likely than others—for example, more people will be high on both Conscientiousness and Agreeableness than high on Conscientiousness but low on Agreeableness.2 That some personality trait configurations are more common than others is already well-known (Ones, Wiernik, Wilmot, & Kostal, 2016). In the industrial-organizational psychology literature, many measures have been developed and applied that capture people’s similarity to these high-density configurations. For example, Integrity tests assess a configuration of high Conscientiousness, high Agreeableness, and low Neuroticism (Giordano, Ones, & Viswesvaran, 2017; Ones & Viswesvaran, 2001), and Narcissism measures capture a configuration of high Extraversion but low Agreeableness (O’Boyle, Forsyth, Banks, Story, & White, 2015). These measures are called compound personality scales and are among the most widely assessed personality scales in practice (Connelly, Ones, & Hülsheger, 2018).

Do personality types exist? Demonstrating that personality forms “types” requires identifying unambiguous characteristics that indicate differences in kind, in contrast to dimensional characteristics that describe differences in degree. For personality to function as “types,” members of different classes (e.g., Type A versus Type B) must have clearly distinct levels on a trait or cluster of traits, with little overlap between groups. Decades of personality research have found that personality traits do not work this way—people’s personalities are best understood as profiles of high, moderate, or low (continuous) levels on a variety of continuous dimensions (e.g., the Big Five; McCrae & Costa, 1989). Indeed, the last bastions of types in personality theory have recently fallen (e.g., see Wilmot, 2015, for Self-Monitoring; and Wilmot, Haslam, Tian, & Ones, 2018, for Type A), with data clearly supporting dimensional interpretations over discrete personality types. The basic units of personality are continuous traits, a fact undisputed by Gerlach et al. The density clusters identified by Gerlach et al. overlap substantially in their Big Five trait levels and cannot be interpreted as discrete types. Rather than “types,” they are better as described as “lumps in the batter” of personality traits (William Revelle, as quoted in Ouellette, 2018). Any scheme that places individuals into a limited number of clusters (e.g., 16 types in the Myers-Briggs Personality Type Indicator) necessarily breaks continuous personality dimensions into discrete units and therefore loses valuable individual differences information.

What are the implications for scientists who measure personality in their studies? There is no scientific support for abandoning dimensional personality measurement in favor of using types in research. The pan-hierarchical five factor model provides an integrative taxonomy of traits at all levels of the personality hierarchy (Stanek & Ones, 2018) and can help scientists better select and use appropriate dimensional personality measures.

What are the implications for practitioners who use personality assessments in organizations? Many practitioners who contacted us wondered whether they should incorporate Gerlach et al.’s four “types” into their assessments. We do not see much value in this—for example, there is little to be gained by declaring a person is “average”—most people will fall into this cluster. Any decision based on personality density-cluster membership (“type”) is bound to lose highly valuable predictive information about the constituent traits. However, direct dimensional measures tapping compound personality traits may be useful for organizational applications (e.g., self-centeredness, a compound trait similar to one of the Gerlach et al. clusters).

What are the implications for personality publishers and consultancies? Most companies offering personality assessments have much larger test-taker databases than those used by Gerlach et al., with responses from hundreds of thousands, if not millions, of people. These data are a rich resource for identifying multidimensional personality density distributions. Publishers and consultancies can apply computational approaches similar to Gerlach et al. to explore the nomological networks of their measures. If clusters similar to those described by Gerlach et al. emerge, publishers and consultancies can explore whether cluster membership predicts work-relevant variables. In doing so, it is critical to remember that differences among individuals are ones of degree and not type!

If readers of this short note have additional questions about Gerlach et al.’s research—or more generally personality types—please let us know. We hope to compile the above and other questions into an expanded FAQ resource. Deniz Ones may be reached at deniz.s.ones-1@umn.edu, and Brenton Wiernik may be reached at brenton@wiernik.org.

Notes

1 For more nuanced or critical perspectives in popular outlets, see pieces published in Psychology Today (Hutson, 2018), Live Science (Pappas, 2018), or Ars Technica (Ouellette, 2018).

2 Density patterns become even more complex if we make finer attribute distinctions by considering more specific personality aspects and facets. In this case, we would examine the distribution of people in N-dimensional space.

References

Connelly, B. S., Ones, D. S., & Hülsheger, U. R. (2018). Personality in industrial, work and organizational psychology: Theory, measurement and application. In D. S. Ones, N. Anderson, C. Viswesvaran, & H. K. Sinangil (Eds.), The SAGE handbook of industrial, work and organizational psychology (2nd ed., Vol. 1, pp. 320–365). London, UK: Sage. https://doi.org/10.4135/9781473914940.n13

Davies, S. E., Connelly, B. L., Ones, D. S., & Birkland, A. S. (2015). The general factor of personality: The “Big One,” a self-evaluative trait, or a methodological gnat that won’t go away? Personality and Individual Differences, 81, 13–22. https://doi.org/10/bc98

DeYoung, C. G. (2015). Cybernetic big five theory. Journal of Research in Personality, 56, 33–58. https://doi.org/10/33h

Dier, A. (2018, September 18). Scientists have pinpointed 4 personality types. Retrieved from http://www.newser.com/story/264776/self-centered-scientists-say-theres-hope.html

Digman, J. M. (1997). Higher-order factors of the Big Five. Journal of Personality and Social Psychology, 73(6), 1246–1256. https://doi.org/10/b53p9f

Ducharme, J. (2018, September 18). These are the four big personality types, according to science. Time. Retrieved from http://time.com/5398396/personality-types/

Duff, A. (2018, September 17). Scientists discover four new personality types... which one are you? The Sun. Retrieved from https://www.thesun.co.uk/fabulous/7281117/are-you-average-reserved-self-centered-or-a-role-model-scientists-have-discovered-four-new-personality-types/

Gerlach, M., Farb, B., Revelle, W., & Amaral, L. A. N. (2018). A robust data-driven approach identifies four personality types across four large data sets. Nature Human Behaviour, 2(10), 735–742. https://doi.org/10/gd8ds5

Giordano, C. A., Ones, D. S., & Viswesvaran, C. (2017). Integrity testing and antisocial work behavior. In B. J. Carducci (Ed.), The Wiley encyclopedia of personality and individual differences (Vol. 4). Hoboken, NJ: Wiley.

Guarino, B. (2018, September 17). Scientists identify four personality types. Washington Post. Retrieved from https://www.washingtonpost.com/science/2018/09/17/scientists-identify-four-personality-types/

Holohan, M. (2018, September 20). Study finds 4 main personality types — which one are you? Retrieved from https://www.today.com/health/personality-types-average-self-centered-role-model-or-reserved-t137902

Hutson, M. (2018, September 17). Are there really four personality types? Psychology Today. Retrieved from https://www.psychologytoday.com/blog/psyched/201809/are-there-really-four-personality-types

McCrae, R. R., & Costa, P. T. (1989). Reinterpreting the Myers-Briggs Type Indicator from the perspective of the Five-Factor Model of personality. Journal of Personality, 57(1), 17–40. https://doi.org/10/cxz4kd

Nauert, R. (2018, September 18). Big data study challenges thinking on personality types. Retrieved from https://psychcentral.com/news/2018/09/18/big-data-study-challenges-thinking-on-personality-types/138792.html

Northwestern University. (2018a, September 17). Scientists determine four personality types based on new data: Comprehensive data analysis dispels established paradigms in psychology [Press release]. Retrieved from https://news.northwestern.edu/for-journalists/press-kits/scientists-determine-four-personality-types-based-on-new-data/

Northwestern University. (2018b, September 17). Scientists determine four personality types based on new data: Comprehensive data analysis dispels established paradigms in psychology. Retrieved from https://www.sciencedaily.com/releases/2018/09/180917111612.htm

O’Boyle, E. H., Jr., Forsyth, D. R., Banks, G. C., Story, P. A., & White, C. D. (2015). A meta-analytic test of redundancy and relative importance of the Dark Triad and five-factor model of personality. Journal of Personality, 83(6), 644–664. https://doi.org/10/gckf5d

Ones, D. S., & Viswesvaran, C. (2001). Integrity tests and other criterion-focused occupational personality scales (COPS) used in personnel selection. International Journal of Selection and Assessment, 9(1/2), 31–39. https://doi.org/10/bzfvgm

Ones, D. S., Wiernik, B. M., Wilmot, M. P., & Kostal, J. W. (2016). Conceptual and methodological complexity of narrow trait measures in personality-outcome research: Better knowledge by partitioning variance from multiple latent traits and measurement artifacts. European Journal of Personality, 30(4), 319–321. https://doi.org/10/bp27

Ouellette, J. (2018, September 17). Study: People tend to cluster into four distinct personality “types.” Ars Technica. Retrieved from https://arstechnica.com/science/2018/09/study-people-tend-to-cluster-into-four-distinct-personality-types/

Pappas, S. (2018, September 18). Are these four personality types for real? Live Science. Retrieved from https://www.livescience.com/63613-are-personality-types-legit.html

Scientists say there are four personality types, not all are interesting. (2018, September 19). Daily Sabah. Retrieved from https://www.dailysabah.com/science/2018/09/20/scientists-say-there-are-four-personality-types-not-all-are-interesting

Smith, D. G. (2018, September 18). Big data gives the “Big 5” personality traits a makeover. Scientific American. Retrieved from https://www.scientificamerican.com/article/big-data-gives-the-big-5-personality-traits-a-makeover/

Stanek, K. C., & Ones, D. S. (2018). Taxonomies and compendia of cognitive ability and personality measures relevant to industrial, work, and organizational psychology. In D. S. Ones, N. Anderson, C. Viswesvaran, & H. K. Sinangil (Eds.), The SAGE handbook of industrial, work and organizational psychology (2nd ed., Vol. 1, pp. 366–407). London, UK: Sage. https://doi.org/10.4135/9781473914940.n14

Stillman, J. (2018, September 21). These are the only personality types backed by science. Inc.Com. Retrieved from https://www.inc.com/jessica-stillman/these-are-only-personalty-types-backed-by-science.html

Wilmot, M. P. (2015). A contemporary taxometric analysis of the latent structure of self-monitoring. Psychological Assessment, 27(2), 353–364. https://doi.org/10/38w

Wilmot, M. P., Haslam, N., Tian, J., & Ones, D. S. (2018). Direct and conceptual replications of the taxometric analysis of Type A behavior. Journal of Personality and Social Psychology. Advance online publication. https://doi.org/10/cp7n