Big Five Personality Traits

From the lexical hypothesis (Galton, Allport, Cattell) — important individual differences are encoded in language. Researchers extracted thousands of trait adjectives from dictionaries (Allport & Odbert 1936 — 17,953 words), reduced redundancy, and factor-analyzed self- and peer-ratings. Five robust factors emerged repeatedly in multiple labs (Tupes & Christal 1961, Norman 1963, Goldberg 1981). Costa & McCrae's NEO-PI (1992) operationalized the model with 240 items and six facets per trait.

Substantial empirical literature. Conscientiousness — academic GPA, job performance, income, longevity (Friedman et al. 1995). Extraversion — subjective well-being, social network size, leadership emergence. Neuroticism — depression, anxiety, divorce, lower life satisfaction. Agreeableness — relationship quality, prosocial behavior, lower workplace conflict. Openness — creative output, political liberalism, curiosity-driven careers. Effects are real but modest — typically r = 0.2-0.4 with single outcomes; trait combinations predict more.

Twin and adoption studies converge on 40-60% heritability for each trait. The remainder is non-shared environment (life events, peer groups, idiosyncratic experience). Shared family environment matters surprisingly little once adolescents leave home. Specific genes have small individual effects; recent genome-wide association studies (GWAS) implicate hundreds of variants per trait, each contributing fractions of a percent. The genetic architecture is highly polygenic, like height or BMI.

Yes, increasingly so with age. Test-retest correlations rise from r ≈ 0.4 in childhood to r ≈ 0.7-0.8 in adulthood (Roberts & DelVecchio 2000 meta-analysis). Mean-level changes follow predictable patterns — conscientiousness and agreeableness rise into middle age (the "maturity principle"); neuroticism declines. Major life events (marriage, parenthood, occupational change) can shift traits modestly. The basic profile of a 25-year-old strongly predicts the 60-year-old profile.

Myers-Briggs (MBTI) is widely used in business but psychometrically weak. Type-based — sorts people into 16 categories — but the underlying dimensions are bimodally enforced when actual data is unimodal. Test-retest reliability is poor (around 50% land in different types upon retake within weeks). Predictive validity for life outcomes is weak. The Big Five uses continuous dimensions, has high reliability and replicated factor structure, and predicts substantially better. Most personality researchers regard MBTI as pseudoscience for serious use.

Moderate to strong but not universal. Five-factor structure replicates well in WEIRD samples and many other industrialized cultures (translations of NEO-PI confirmed in 50+ countries). Small-scale societies show weaker results — Gurven et al. (2013) with Tsimané forager-horticulturalists found only two robust factors. Whether the traits are genuinely universal or partly artifacts of the urbanized lexical pools used to derive them is an open question. The model is the best validated, not necessarily the only valid one.

Several. (1) Five factors may be too few — HEXACO model (Ashton & Lee 2001) adds Honesty-Humility as a sixth. (2) Or too many — some argue traits collapse into two metatraits (alpha/stability and beta/plasticity). (3) Self-report bias — situational variability and self-presentation distort scores. (4) Behavior is more situational than traits suggest (Mischel 1968). (5) Trait labels carry cultural baggage. Despite critiques, the model remains the gold standard for personality research, with HEXACO as the leading challenger.