In 1931 the Australian Institute of Industrial Psychology published lectures by A. H. Martin who, after dismissing astrology, palmistry and phrenology as pseudo-sciences, bravely put psychology forward as a legitimate science and confidently listed the desirable qualities for successful salesmen: extroversion, humour, resilience, diplomacy and self-confidence.
By the 1950s the use of personality tests by personnel managers and consultants was widespread and, for some people, worrying.
Personality tests are problematic for empirical, logical and ethical reasons. Empirically, personality tests do not predict job performance and cannot determine “fit” with anything except the prejudices of their users. Logically, their promoters are haunted by problems of the gap between theory and reality and circular arguments on why people behave a certain way.
Ethically, many individuals are required to undertake personality tests against their will and when results are shared with third parties, as they usually are in management, the opportunity for psychological manipulation of those tested is obvious and has been widely acknowledged, for example, by well-known US management author Peter Drucker.
More than a passing fad
Julie Marshall and Richard Trahair published a comprehensive record of Industrial Psychology in Australia to 1950 in which 21% of 1551 articles are devoted to incentive schemes and profit-sharing, 11% to personnel management and industrial psychology, and 6% to management training, and management traits and attitudes.
In his 1956 book, The Organization Man, William H. Whyte wrote:
“Regularly year after year many social scientists have assured me that this bowdlerization of psychology is a contemporary aberration soon to be laughed out of court […] Personality testing? Again, each year the number of people subjected to it has grown, and the criticism has served mainly to make organizations more adept in sugar-coating their purpose.”
And to level the playing field somewhat, he added an appendix called “How to Cheat on Personality Tests”. A half-century later Helen De Cieri and Robin Kramar in Human Resource Management in Australia reported 69% of resource managers surveyed believed personality tests were valuable tools that could be used to improve performance.
Research spanning more than 80 years has shown attempts to predict management performance from personality tests have been a spectacular failure. Meta-analyses of the validity of personality tests as predictors of job performance have shown low or no levels of predictive validity.
This is based on the use of the Five Factor Model, which reduces thousands of trait-type descriptions into just the “big five” personality traits.
Even the experts can’t agree
Of course, those in love with the Myers-Briggs Type Indicator disagree: such is the nature of this business. Personality testers cannot agree on a definition of personality or of a personality trait. They cannot agree on which traits should be measured because the number of traits in a test is determined by statistical techniques about which psychometricians disagree.
Critics of the big five argue they are too broad for understanding and predicting behaviour generally, let alone job performance specifically. And the much lauded, though modest, correlation between “conscientiousness” and job performance is rendered even more modest when the trait is correlated with objective, or “hard”, performance criteria.
It is unsurprising that test questions which invite testees to describe themselves as conscientious produce answers which correlate with managers’ subjective ratings of their personalities. It is a different story, however, when self-reports are correlated with the achievement of actual results.
Behaviourists and existentialists have long argued, from very different assumptions about human nature, that there are no such things as personality traits. B.F. Skinner favoured external rather than internal causes of behaviour and Jean-Paul Sartre argued that human beings are free to choose their behaviour so that what is called personality is the sum of our choices under the influence, but not control, of the external facts of our existence.
Interestingly, behaviourism and existentialism are products of respectable philosophical systems – British Empiricism and German Romanticism respectively - whereas personality trait psychology is a product of statistical games played with dubious self-report data which are subject to faking.
Too many assumptions are made
Because trait psychology is philosophically primitive, personality testers ignore or cannot recognise the problem of circularity of argument which haunts their enterprise. For example, they assume that personality traits are enduring, pervasive and affect all aspects of behaviour.
Since they cannot be observed, personality traits are theoretical constructs which are measured indirectly by self-report inventories. Importantly, personality traits are assumed to be “inner powers” because they allegedly “cause” individuals to act in certain ways. How these theoretical constructs get their power has never been explained. American psychologist David McClelland referred to them, with a smile, as “mental viruses”.
Personality traits, then, are inferences from the behaviour which gave rise to them. “Anxiety” is an inference from behaviour which is called “anxious”. How do we know that individuals are anxious? They are anxious because they behave anxiously. Why do they behave anxiously? They behave anxiously because they are anxious. There is no way of breaking out of this circle of spurious argument unless personality traits and the behaviour they are derived from are separately identifiable. This cannot happen because personality traits are abstractions.
Since personality tests have repeatedly failed to predict management performance, one might expect on rational grounds that managers would have by now rejected them. This happened to phrenology a century ago and it should have happened to personality testing a half-century ago.
Read more of our personality at work series: