Judging from the amount of private correspondence about my last blog, it seems that it has caused quite a stir. To be clear, I’m a fan of personality testing but am not a zealot. I believe it has a role in any thorough assessment process but:
- It should not be overpriced.
- Tests are far more similar than they are distinct.
- Its power of prediction is far lower than the public are led to believe.
- It is flawed and therefore must be used with a degree of healthy scepticism.
Self-report tools have clear limitations to the measurement of personality and this is confounded by the nonsense that we can tap into unconscious personality using self-report measures of behavioural preferences. This said, the idea that there are clusters of behaviours that can be used to predict other clusters of behaviours is the basis for construct validity and has found huge support at a global level (i.e., The Big Five). The next question for the science of trait theorists is what commonalities exist at a primary trait level. When test distributors are all distributing common models at a facet level, progress will be made.
From an application perspective, I do question the predictive power of personality tools. The fact that they predict at equally poor levels is not a convincing argument for a practitioner. I would argue that while researchers like Hogan provide a great service to the theory of personality they diminish that contribution somewhat when they try to quantify it with a degree of specificity that is unwarranted. This is a problem for a discipline when theory oversteps evidence to appear sounder. An analogy is the religious who try desperately to convince that God exists using reason when the basis of their belief is more often faith.
Personality is central to the study of psychology (which I am passionate about). The key for me is part predictive validity (application) and part understanding human behaviour (science). In terms of aiding application, the key must be system thinking. The battle is not between test providers all showing the strength of their magical question sets (and producing larger quantities of so-called validity research as the basis) but rather a shift in what constitutes evidence for application. We need more complex examinations of how personality assessments interplay with environments and all other manners of variables (both individual and external) to give practitioners tangible results that they can use. This requires both stronger theory building, complex modelling, and computational analysis.
Test providers should be exploring different and more comprehensive methods for determining test utility. In this regard, what is important is solid theory and solid science rather than using weak paradigms of what constitutes evidence to justify test sales.