Personality models: perhaps we’re all wrong!

November 15, 2010

In Japan many people believe that a person’s blood type is predictive of their personality. Being asked about your blood type is commonplace and most Japanese would be puzzled if you didn’t know what it was. Blood type is used in recruitment and matchmaking, and for example many Japanese celebrities and politicians include it in their personal profiles. The gist is that if you’re Blood Type A you’re a responsible perfectionist; B, a practical individualist; O, a confident leader (and naturally this is considered to be the ‘best’ blood type!); and AB, that you’re unpredictable because you’re a mix of other types. Now it’s easy to take a smug, Western view of this, and whilst there is no proven link between blood type and personality, there is a heritable component to personality. Thus at the population level, mean estimates of heritability for the Big Five traits, which are not the same as the ABO ‘traits’ just mentioned, are: Openness (57%), Conscientiousness (49%), Extraversion (54%), Agreeableness (42%) and Neuroticism (48%)* – Bouchard & McGue, 2003.

Why do I mention all this? We are uncertain about what really ‘causes’ personality. It is likely to be influenced by heredity and environment, but how? And the other big question, what about the structure of personality itself? This last issue brings me to a recent and fascinating piece of research. Read et al (2010), using a neural network model of human personality, suggest that the Big Five is a representation of the structure of personality across a large group, based on the covariance (a measure of how pairs of variables change together) of characteristics within that group, not the structure that you would observe for an individual. The punch line: if you created a large number of virtual humans, specified by a random set of characteristics, you would expect the pattern across individuals to end up looking like the Big Five. Oops, amongst other obvious questions, if this is true and structure operates at a group (population) level, what does the score for a scale on a personality questionnaire for an individual actually mean? It looks like we may need to do some serious thinking about the structure of personality.

Bouchard, T.J. &  McGue, M. (2003). Genetic and Environmental Influences on Human Psychological Difference. Journal of Neurobiology, 54, 4-45.

Read, S.J., Monroe, B.M., Brownstein, A.L., Yang, Y., Chopra, G. & Miller, L.C. (2010). A Neural Network Model of the Structure and Dynamics of Human Personality. Psychological Review, 117, 61-92.

* Note. This does not imply that for an individual 57% of their ‘openness’, for example, is caused by their genetic makeup. It means that the (observable) phenotypic variation is 57% due to genetic variation.


Get every new post delivered to your Inbox.

Join 88 other followers