Niche diversity can explain cross-cultural differences in personality structure

Abstract

The structure of personality refers to the covariation among specific behavioral patterns in a population. Statistically derived models of personality-such as the Big Five or HEXACO models-usually assume that the covariance structure of personality characteristics is a human universal. Cross-cultural studies, however, have challenged this view, finding that less complex societies exhibit stronger covariation among behavioral characteristics, resulting in fewer derived personality factors. To explain these results, we propose the niche diversity hypothesis, which predicts that a greater diversity of social and ecological niches elicits a more diverse set of multivariate behavioral profiles, and hence lower trait covariance, at the population level. We formalize this hypothesis as a computational model in which individuals assort into niches, which influence their behavioral traits. The model provides strong support for the niche diversity hypothesis and reproduces empirical results from recent cross-cultural studies. The model also provides a novel prediction for which we find support empirically: individual trait variation increases with niche diversity. This work provides a general explanation for differences in personality structure between populations in both humans and other animals. It also suggests a radical reimagining of personality trait research: instead of reifying statistical descriptions of manifest personality structures, research should focus more attention on modeling their underlying causes.

Cross-cultural, developmental psychology: integrating approaches and key insights

“…The model provided a novel prediction: that individual trait variation should increase with niche diversity. The team then found support for this novel prediction in a large cross-cultural dataset, finding that greater socioecological complexity was indeed associated with more individual trait variation (Smaldino et al, 2019) . This integrative process of bringing together existing data with formal models to produce novel predictions that are then empirically tested is a strong example of the "virtuous cycle" of replicable science (Smaldino, 2019) and represents a tractable approach for increasing the explanatory power of our research.…”

“…The first of these is a series of related projects exploring the structure of personality across diverse contexts. The study of personality -or persistent patterns of behavior that are stable across time and contexts (Fleeson, 2001;Smaldino et al, 2019;Wood & Denissen, 2014) -and the way in which it may be structured has long been of interest to psychologists. One common approach to studying personality structure is to statistically extract patterns of covariance into factors.…”

Practical Methodological Reform Needs Good Theory

“…Specifically in this case, theory could help 1) select effects which might be sample-moderated according to prespecified theoretical expectations, rather than vaguely specified intuitions, 2) select the samples which would make strong tests of underlying constructs, xi 3) aid in designing a moderating measure that goes beyond a binary split of an amalgamation of archival data sources that correspond to letters in an acronym, and 4) evaluate the strength of underlying evidence. Alas, this test has not been conducted yet, but theoretically-driven approaches to culture tend to be able to specify when cultural heterogeneity is expected (e.g., Gervais et al, 2017;Henrich et al, 2006;Kitayama & Cohen, 2010;Kline et al, 2018;Legare et al, 2012;McNamara et al, 2019;Purzycki et al, 2016;Smaldino, Lukaszewski, et al, 2019;Willard & Cingl, 2017), while also being able to specify instances where homogeneity might instead be more likely (e.g., Apicella et al, 2012;Sznycer et al, 2017).…”

Equilibria in Personality States: A Conceptual Primer for Dynamical Analyses

“…Some researchers have developed formal models of personality and underlying processes, which translate "top down" theoretical intuitions into quantitative simulations (e.g. Read & Miller, 2002;Smaldino, Lukaszewski, Von Rueden, & Gurven, 2019). To complement these top-down approaches, we estimate here parameters from observed data, hoping that this "bottom up" approach will spur further theorizing about dynamic processes underlying personality.…”