The Real and Fake Faces of Personality’s "Big One"
The so-called general factor of personality is artifactual.
Posted Mar 24, 2020
As I noted in a previous post, there is debate about the structure of personality traits. Current theories suggest that traits are organised into a hierarchical structure in which a large number of narrow specific traits can be subsumed into a smaller number of broader factors. The most famous example is the Big Five model that describes people's personality traits in terms of the five major factors of neuroticism, extraversion, agreeableness, conscientiousness, and openness to experience.
A rival theory – the HEXACO model – proposes six factors as more comprehensive, the most important difference being the addition of a factor called honesty-humility. In contrast to these multi-factor models, some scholars propose an all-encompassing general factor of personality (GFP) that combines all of the five or six factors into a single super-factor that is supposed to represent social effectiveness.
Proponents of the GFP have argued that it is a meaningful and substantive feature of human personality that emerged during human evolution because of its adaptive value. Critics argue that it is likely a statistical artefact instead. I have argued for the latter position in a number of articles.
A couple of recent studies cast further doubt on the validity of the GFP, as they show that there is not one, but several different versions of the GFP that can be extracted from personality inventories (Schermer & Goffin, 2018) and that a supposed GFP becomes stronger when people are instructed to alter their responses to a personality test to create a more positive impression of themselves (MacCann et al., 2017).
Proponents of five- or six-factor models argue that these factors fundamentally operate independently of each other. However, in practice, when people complete measures of personality, their scores on these factors tend to be inter-correlated, which has led to the suggestion that they all have a shared basis.
Proponents of the GFP have used the statistical technique of factor analysis to support this claim and argue that the GFP represents a shared factor that combines emotionally stability (i.e., the opposite pole of neuroticism), conscientiousness, agreeableness, extraversion, and openness from the Big Five (Musek, 2007), so that individuals with high GFP scores could be described as ‘‘open-minded, hard-working, sociable, friendly, and emotionally stable’’ (van der Linden et al., 2010). Similarly, researchers have attempted to extract a GFP from a measure of the HEXACO (Veselka et al., 2009), although with questionable results as the GFP in this study was most strongly related to extraversion and only weakly related to honesty-humility, with modest relations with the other factors.
Critics of these approaches have pointed out problems with the analyses, which I have discussed elsewhere. One criticism I want to focus on here is that substantially different versions of the GFP have been found in different studies depending on the methods and measures used.
In fact, one study gave the same group of people two different personality measures and used two different statistical methods to generate a purported GFP. As a result, they found four different versions that had striking divergences from each other.
Specifically, for each personality measure, they used two different methods of factor analysis to extract the GFP. Importantly, only one of these GFPs came close to resembling a classic GFP that combines all the best characteristics of the standard personality factors, in this case, it was a GFP that combined high openness to experience and extraversion, and to a lesser extent conscientiousness. The other GFP derived from the same personality measure using a different method indicated caring extraversion (caring was not a salient factor in the other GFP derived from this measure). The other two GFPs derived from the other personality measure indicated mainly either emotionally stable extraversion on the one hand or mainly submissiveness on the other (this second GFP was negatively correlated with the other GFPs). If the GFP was a reliable and universal feature underlying personality, it would be expected that different methods and measures would converge on reasonably similar substantive content. However, each of the four GFPs in this study, which were derived from the same people’s data, seemed to have little overlap. This provides evidence against the GFP being a truly general feature of personality and instead suggests that it is more likely to be a statistical artefact.
Another objection to the GFP is that the correlations between the Big Five or HEXACO represent socially desirable responding rather than substantive overlap between them, that is, people tend to respond to personality measures in a way that is considered socially acceptable, which creates illusory correlations between unrelated personality traits.
Although there is evidence that the GFP is related to socially desirable responding, proponents have argued that social desirability is a substantial feature of personality, that is, people high in the GFP actually have “better” qualities than others that are socially valued.
To get around this problem, a recent study (MacCann et al., 2017) experimentally manipulated socially desirable responding by instructing people to complete a measure of the HEXACO under two different conditions: honest responding or pretend that they were completing the measure as part of a job application and they wanted to give the impression of being an ideal employee, that is, to “fake good.” Additionally, participants completed an intelligence test. This was because proponents of the GFP have argued that it is positively associated with intelligence, that is, “good” (adaptive) personality traits go along with other adaptive qualities, including intelligence.
Comparing the results showed that participants scored higher on each of the six HEXACO factors, i.e., higher on honesty-humility, conscientiousness, agreeableness, extraversion, and openness when they were instructed to fake good than when they answered honestly, and the largest effects were for extraversion, agreeableness, and conscientiousness in particular. A GFP was extracted from the six HEXACO factors in each of the two conditions.
The correlations between the various personality traits tended to be stronger in the faking condition, and the GFP was also statistically more robust under faking as well. Additionally, in the honest condition, the GFP was unrelated to intelligence, but in the faking condition, the GFP had a significant positive correlation with intelligence.
This indicated that under instructed faking, people with higher intelligence endorsed more socially desirable traits than those with lower intelligence. However, when answering honestly, this effect did not occur. This indicated that the GFP emerges more strongly when people are trying to create a socially desirable impression, which is consistent with the GFP being an artefact of response distortion rather than a basic feature of personality. Additionally, it showed that the ability to create a socially desirable impression is related to intelligence, perhaps because more intelligent people find it easier to figure out which characteristics are more socially desirable, such as when seeking employment.
Regarding intelligence, openness to experience is the trait that is most strongly related to actual intelligence. However, it seemed that smarter people understood that they needed to be seen as higher in extraversion, agreeableness, and conscientiousness, than in the other traits including openness to experience. Hence, it seems that the GFP might represent the way smart people know how to present themselves, but not the way they necessarily are.
 There are also some minor differences between Big Five neuroticism and HEXACO emotionality, although these are not of great importance for the purposes of this article.
 For the statistically minded, the first method involved considering the first unrotated factor in factor analysis as the GFP or by extracting lower-order factors then extracting a higher-order GFP. Both of these methods have been previously used by GFP proponents.
MacCann, C., Pearce, N., & Jiang, Y. (2017). The general factor of personality is stronger and more strongly correlated with cognitive ability under instructed faking. Journal of Individual Differences, 38(1), 46–54. https://doi.org/10.1027/1614-0001/a000221
Musek, J. (2007). A general factor of personality: Evidence for the Big One in the five-factor model. Journal of Research in Personality, 41(6), 1213–1233. https://doi.org/10.1016/j.jrp.2007.02.003
Schermer, J. A., & Goffin, R. D. (2018). A tale of two general factors of personality in relation to intelligence and validity measures. Personality and Individual Differences, 124, 111–116. https://doi.org/10.1016/j.paid.2017.12.010
van der Linden, D., te Nijenhuis, J., & Bakker, 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(3), 315–327. https://doi.org/10.1016/j.jrp.2010.03.003
Veselka, L., Schermer, J. A., Petrides, K. V., Cherkas, L. F., Spector, T. D., & Vernon, P. A. (2009). A General Factor of Personality: Evidence from the HEXACO Model and a Measure of Trait Emotional Intelligence. Twin Research and Human Genetics, 12(5), 420–424. https://doi.org/10.1375/twin.12.5.420