Personality theory has made huge strides since the beginning of the lexical movement, pioneered by Cattell in the 1940’s. Through iterative research, the Big Five emerged from the storm of factor analyses, and has ever since been atop of the personality ladder. The most academically established, strongly validated and generally agreed upon framework for measuring and describing human personality – The Big Five Factors; Extraversion (E), Neuroticism (N), Openness (O), Agreeableness (A) and Conscientiousness (C).
Cue subtle rumblings in the literature a number of decades ago suggesting a higher order dimension of personality – a global factor of personality (GFP) analogous to the ‘g’ which represents human intelligence. This school of thought has gained ground of late, perhaps because if the huge meta-analysis conducted by van der Linden, Nijenhuis & Bakkar (2010), which shows strong evidence for a hierarchical structure of personality above the Big Five. This topic area was explored in an excellent symposium at the recent BPS Division of Occupational Psychology (DOP) conference, with two of our regular contributors, Dr Rainer Kurz (Saville consulting) and Rob Bailey (OPP), arguing for and against, respectively a GPF.
In the for-GPF corner, Dr Kurz discussed the meta-analysis by van der Linden et al (2010), which consisted of 212 studies, amassing a sample of over 144,000, exploring the inter-correlations between the Big Five Factors. The thought alone of the SPSS data file is a scary prospect. Using factor analysis, they were able to confirm the existence of two meta-factors; Alpha – consisting of C, A and –N (emotional stability) and Beta consisting of O and E. These two factors then loaded (fed upwards) onto a global factor. See the diagram below, which has been copied from the paper and shows the factor structure along with the factor loadings of each of the Big Five factors onto the GFP. So it would seem that human personality can be explained by one factor. However, is it that simple? Does personality boil down to numbers, factor analyses and cold, hard data?
What about the common sense argument that one factor of personality is too crude to predict the various criterion outcomes i.e. having predictive validity? Such predictors would include job performance, contextual performance, training proficiency and a whole host of other organisational outcomes. Having all bases covered or so it would seem, van der Linden et al (2010) did in fact find significant correlations between the GFP and these criterion variables, hence demonstrating its predictive validity. However, another issue is that essentially, someone who is low on the GFP has an undesirable personality. Characteristics include lacking altruistism, being tense, having low well-being, being dissatisfied with life, having low self-esteem and a lack of emotional intelligence (Bailey, 2013; DOP Symposium). With individuals high on the GFP having the inverse characteristics, it’s clear to see who wins. I’m not sure I’d want to deliver feedback to someone who is low on the GFP factor. As if delivering feedback for a highly neurotic individual wasn’t difficult enough. But of course, this doesn’t negate the need to get personality correct. If this is the true structure of personality, then so be it. I’d rather have the truth, than only half the story – even if that meant discovering the ‘undesirable’ qualities of my personality.
Returning to Rob Bailey’s research at the symposium, he in fact found no evidence for a single factor in US (N over 30,000) or UK (N over 1000) data. However, in both instances, the fewest possible factors were two. Along with this, the argument was made that more ‘granular’, specific personality constructs have stronger predictive power (for job performance, etc) than broad or generic personality factors. However, a review of the personality literature by Rothstein & Goffin (2006) suggests that both narrow and generic personality factors have their place in predicting criterion variables. It’s all proportional, if the maximum predictive power is to be realised. This means that if a specific criterion variable is to be predicted, a narrow personality concept i.e. facet level, will have stronger predictive power. And of course, with broader criterion variables, more generic personality factors will exhibit stronger predictive validities. Not only supported by meta-analytical research but it has a certain intuitive basis, in that a broader criterion variable i.e. leadership, will encompass a wider variety of personality elements.
So the moral of the story would seem to be that personality structure should be used in a ‘fit-for-purpose’ manner in a practical setting. This means that HR practitioners and assessment and selection specialists must understand the competencies and outcomes they want to measure. Using this information, they can thus decide the appropriate level of personality structure. In terms of personality theory, it is encouraging to see that we continue to push the boundaries of our understanding. However, in our efforts to marry theory and practice, we must be cautious. The Global Factor of Personality may not necessarily have the same use as the ‘g’ in cognitive ability. Furthermore, there is a definite need for the GPF to be more theory driven, as opposed to data driven. Whether it is an evolutionary, genetic account or not, the establishment of the theoretical basis for the GPF will help towards its credibility. This is a fascinating area of research, and one in which I’m sure more research will unfold. Until then, only time will tell us what the significance of the GFP is.
Blog posted by Raj Chopra – follow me on Twitter at @Raj_Glowatwork.
van der Lindin,D., Nijenhuis, J. & Bakkar, A.B. (2010). The General Factor of Personality: A meta-analysis of Big Five intercorrelations and a criterion-related validity study. Journal of Research in Personality, 44, 315–327.
Rothstein, M.G. & Goffin, R.D (2006). The use of personality measures in personnel selection: What does current research support? Human Resource Management Review, 16, 155–180.