The personality assessment industry is large, commercially confident, and built on a premise that the research only partially supports. Paul Costa and Robert McCrae’s Big Five model, developed through the 1980s and cross-culturally replicated since, established that five stable trait dimensions — openness, conscientiousness, extraversion, agreeableness, and neuroticism — show meaningful heritability, cross-time stability, and genuine predictive validity for broad life outcomes. Conscientiousness predicts occupational performance. Neuroticism predicts relationship difficulties. Extraversion predicts social network size. These are real and replicable findings. They are also aggregate findings, describing patterns across many people and many occasions — and that distinction matters enormously for how personality assessments should and should not be used.

Mischel’s challenge and the 0.3 correlation

Walter Mischel’s 1968 Personality and Assessment synthesised hundreds of studies on the relationship between personality trait scores and actual behaviour in specific situations. The average correlation was approximately 0.3, accounting for around nine per cent of behavioural variance. The situation, the role demands, the social context, the specific incentive structure, and the relationship between the person and the others present accounted for the substantial remainder. Mischel’s conclusion was not that personality does not exist but that trait scores, used as predictors of specific behaviour in specific situations, perform modestly.

Lee Ross’s fundamental attribution error research, published in 1977, predicts why this finding is so consistently resisted: people systematically overattribute others’ behaviour to stable personal dispositions while underweighting the situational factors that are actually determining most of the variance. The personality explanation feels complete because it accounts for the behaviour in terms of the person rather than the situation — and the situation is usually less visible than the person. The entrepreneur who dismisses a team member’s risk aversion as a personality trait is making an attribution that feels explanatory while potentially missing the situational factors, the incentive structure, the history of how risk-taking has been received, the specific decision context, that are doing most of the causal work.

The aggregation principle: what traits actually predict

Seymour Epstein’s 1979 aggregation principle provided the most important reconciliation of the trait and situationist accounts. Personality traits do not predict single behaviours reliably, but they do predict aggregated patterns of behaviour across multiple occasions reliably. The high-conscientiousness person will not inevitably behave conscientiously in any particular situation — when they are depleted, when the situational demands run strongly against it, when the social context rewards other behaviour. Across many situations and many occasions, though, the pattern is real and distinguishable from the pattern produced by lower-conscientiousness individuals.

The practical implication is specific. Personality assessment is a reasonable tool for predicting patterns and tendencies across time and contexts. It is a poor tool for predicting what a particular person will do in a particular situation, which is usually the question that matters most in hiring, team composition, and performance assessment. Treating a personality score as an explanation of a specific instance of behaviour is using the tool outside its validated range.

The situation strength moderator

The interactionist perspective, developed by David Magnusson and Norman Endler through the 1970s, and Meyer, Dalal, and Hermida’s subsequent situation strength research, adds the most important practical moderator. The same personality trait produces different behaviour depending on the strength of the situational demands. In strong situations — those with clear role expectations, high evaluation pressure, and powerful incentives — situational forces dominate behaviour and personality trait differences between individuals are suppressed. In weak situations — ambiguous, unconstrained, and low-stakes — personality traits predict behaviour much more reliably, because the situation is not determining what happens and individual differences are therefore free to express themselves.

This means that personality assessments are most predictively useful for behaviour in weak situations and least useful for precisely the high-stakes, high-pressure situations where organisations most want to predict behaviour. The assessment centre, the structured interview, the performance review — all are strong situations where the situational demands are powerful enough to suppress the individual differences that personality scores describe.

The MBTI problem

The Myers-Briggs Type Indicator is the most commercially deployed personality assessment in professional contexts and the least supported by the predictive validity research. Test-retest reliability studies consistently document that approximately fifty per cent of people receive a different type classification on retesting within five weeks. The instrument’s binary categorical structure, sorting people into types rather than placing them on continuous dimensions, misrepresents the underlying trait distributions, which are continuous and normally distributed. Most people are somewhere in the middle of the introversion-extraversion dimension, not clearly one or the other.

The MBTI’s commercial persistence alongside its poor research support is itself instructive. People find personality type frameworks useful for generating self-awareness conversations, for reducing the discomfort of interpersonal difference, and for providing a shared vocabulary for team dynamics. These are genuine benefits that do not require predictive validity. The problem arises when MBTI type classifications are used as predictive tools for hiring, team composition, or developmental planning — applications for which the instrument’s reliability and validity are insufficient.

What personality assessment is good for

The Big Five research establishes a smaller but more defensible set of uses. Conscientiousness is the most reliable single personality predictor of occupational performance across job types, with a correlation of approximately 0.25 in meta-analyses. Combined with specific ability assessments, personality measures improve hiring prediction over either alone. For understanding long-term behavioural tendencies, broad life outcomes, and the general texture of how a person approaches varied situations, well-validated trait measures provide genuine information.

The character strengths framework that Martin Peterson and Christopher Seligman developed through positive psychology adds a complementary perspective: character strengths, described as morally valued traits that the individual has agency to develop, offer a values-based vocabulary for self-description that trait dimensions do not. The distinction matters for self-knowledge. Personality traits describe statistical tendencies. Character strengths describe what a person aspires to express and can choose to develop. Both are useful; neither is the complete account.

Books worth reading on this

Personality and Assessment by Walter Mischel remains the most rigorous available challenge to the overuse of trait-based personality explanation. Mischel’s 1968 book is not a popular account; it is a systematic review of the empirical evidence for trait-based prediction, and its central finding, that trait scores account for a modest fraction of behavioural variance in specific situations, is presented with the care and detail that makes the argument hard to dismiss on methodological grounds. For anyone who uses personality assessments in professional contexts, whether for hiring, team development, or self-knowledge, Mischel’s account of what the instruments actually predict and what they do not is essential calibration. The book has been partially rehabilitated by subsequent interactionist research that confirmed traits do predict aggregate patterns reliably — but Mischel’s core point about the limits of single-instance prediction remains intact and remains widely ignored in commercial personality assessment practice. Reading it does not make personality assessment useless; it makes it possible to use personality assessment for the questions it can actually answer rather than for the broader explanatory role it is routinely but incorrectly assigned.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Costa, P.T. & McCrae, R.R. (1992), Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual, Psychological Assessment Resources. Mischel, W. (1968), Personality and Assessment, Wiley. Ross, L. (1977), The Intuitive Psychologist and His Shortcomings, Advances in Experimental Social Psychology, 10, 173-220. Epstein, S. (1979), The Stability of Behavior: On Predicting Most of the People Much of the Time, Journal of Personality and Social Psychology, 37(7), 1097-1126. Magnusson, D. & Endler, N.S. (1977), Personality at the Crossroads, Erlbaum. Meyer, R.D., Dalal, R.S. & Hermida, R. (2010), A Review and Synthesis of Situational Strength in the Organizational Sciences, Journal of Management, 36(1), 121-140. Boyle, G.J. (1995), Myers-Briggs Type Indicator: Some Psychometric Limitations, Australian Psychologist, 30(1), 71-74. Peterson, C. & Seligman, M.E.P. (2004), Character Strengths and Virtues, Oxford University Press. Mischel, W. (1968), Personality and Assessment, Wiley. Ross, L. & Nisbett, R.E. (1991), The Person and the Situation, McGraw-Hill.