What Is the Five Factor Model of Personality?

The ‘Five Factor Model’ is a model of personality that uses five separate factors to describe an individual’s character. Refined by Goldberg (1990) and developed into the widely used 300 item NEO-PI R personality inventory by Costa & McCrae (1985), according to this theory when an individual is scored on these factors they will produce a complete picture of that person’s personality. This essay will look at whether five fundamental traits can in fact comprehensively explain human personality on their own.

Trait theories of personality rely on factor analysis of commonly used adjectives relating to personality traits which are then grouped together under smaller headings. The first trait theory was the work of Allport (1936) who found 4,504 descriptive personality terms in the then most comprehensive English dictionary. Cattell (1946) then reduced these terms to 171 by eliminating synonyms, before using factor analysis to discover 16 key traits forming the 16PF. Subsequent research has whittled these factors down to fewer more fundamental characteristics and generated a variety of models, with Eysenck’s three (previously two) factor personality questionnaire (1975) being a particularly successful candidate. The general consensus however eventually settled on human personality being best described by five dimensions (Goldberg, 1990; Costa & McCrae, 1985). The five in question, also known as ‘The Big Five’, are: Openness (AKA Intellect), Conscientiousness, Extraversion, Agreeableness, and Neuroticism (AKA Emotional Stability), or ‘OCEAN’ for memorability. While these five traits (if effective) should be sufficient on their own to describe all facets of a personality, there also should be no correlation between the main factors as this would mean they too could possibly be combined under a larger heading. The Five Factor Model is now perhaps the most widely use trait theory of personality and has achieved the closest thing to a consensus in personality research.

One area of support comes from twin studies, which have demonstrated that scores on the five factors are at least partially heritable with genes accounting for around 50% of the variance (with environment accounting for the other 50% (Jang et al., 1996)). This is roughly what would be expected and is similar to findings regarding psychological disorders such as anorexia, giving the Five Factor Model further credibility. Sex differences have also been demonstrated across the traits with females scoring higher on neuroticism and agreeableness and males scoring higher on extraversion and conscientiousness, which seems to concur with casual observation. Additionally the Five Factor Model has been shown to be reliable across cultures, Trull and Geary (1997) for example found that the five traits could be replicated in China, while Ostendorf (1990) found the same in Germany. All this suggests that the Five Factor Model can be used reliably in a variety of contexts and has real-world validity and at least seems to be as capable of explaining personality in other cultures as it is in our own.

There is also a lot of empirical evidence supporting the Five Factor Model and it has been shown to be predictive of behaviour in a variety of contexts. In one study by Saulsman and Page (2004), the relationship between the five dimensions and the 10 personality disorder categories in the DSM-IV was examined revealing that each disorder had its own unique five-factor profile across 15 independent samples. The Five Factor Model has also been shown to be useful in the context of job performance and in a review of 117 studies (cumulatively using 162 samples or 23,994 participants in total) it was found that the Five Factor Model was significantly predictive with the most relevant traits being conscientiousness, extraversion and openness to experience (Mount & Barick, 1998). Similarly, conscientiousness, agreeableness and openness have been shown to correlate significantly with academic performance (Poropat, 2009). In particular conscientiousness was shown to be as accurate a predictor as IQ. Many other studies have been conducted showing stress to be involved in everything from health behaviours (Korotkov, 2008), to technological acceptance and use (Devaraj et al., 2008), to substance abuse (Ruiz et al., 2008; Victor et al., 1973).

Such statistical evidence demonstrates that the Five Factor Model is high in validity and useful as a predictive tool, but whether this means it can fully explain personality is another matter. It should be considered that perhaps behaviour itself might not reflect personality exactly either. For example, two very different personalities can engage in the same behaviours but with different motives. Additionally some people might behave in a way that is not ‘true’ to their personality, if perhaps they are trying to project a different façade. So to take one of the above examples; one worker might appear to be conscientious as a way to climb the ladder and get a promotion while another might genuinely be conscientious. This view is supported by McAdams (1995) who described the big five as a ‘psychology of the stranger’, in that it describes only the observable behaviours we could discern from watching a stranger rather than the more private psychological functions. Therefore to say that the Five Factor Model offers a complete explanation of human personality because it offers an explanation of behaviour may be an illogical leap. Studies looking into correlations between these traits and individuals’ beliefs however could be one way to test for a more direct relation to personality. Taylor & Macdonald (1999) among others found that the Five Factor Model significantly correlated with religious beliefs, which suggests at least some correlation at a deeper level than just observable behaviour.

This obviously is only relevant if five factors are enough to truly offer a comprehensive predictor of personality or behaviour. Many critics of the Five Factor Model accuse the five factors to be insufficient in capturing a comprehensive character profile. There are many alternate theories that use larger numbers of factors such as Cattell’s 16 Personality Factors (1946) (although Cattell also found there to be five ‘global’ traits that could offer headings for the sixteen), the Seven-Factor Psychobiological Model of Temperament and Character (Brown et al., 1993); and the 18-factor model of personality pathology (Livesley, 1986). At the opposite end of the spectrum is Eysenck’s P-E-N theory. The ‘Five Factor Model’ itself grew out of Eysenck’s early work using factor analysis that resulted in a three factor model featuring Extraversion, Neuroticism and Psychoticism. Eysenck argued that the additional two traits were in fact excessive and demonstrated overlap (Eysench, 1991), and that only three were really required. All these trait models have been shown to have their own strengths and weaknesses and the large number of different models demonstrates the subjective nature of using statistical analysis; essentially the resultant factors will be the product of the data used in the factor analysis and here is where the scope for subjectivity and error (Block, 1995) exists. Further, the selection of the factors from the data set again is left to the analyst’s discretion with no universal guidelines. In one recent analysis of English adjectives Saucier and Goldberg (1998) attempted to discover if there was indeed anything not explained by the Five Factor Model and found that the big five were adequate as is. However, in another study that analysed the same data, Paunonen and Jackson (2000; 2001) came to the opposite conclusion stating there was ‘plenty’ beyond the big five, at least 10 traits that are not described by the big five. This sheds doubt on the Five Factor Models comprehensiveness and also illustrates the subjectivity inherent in factor analysis. One study by Bagby et al. (2005) found that the Five Factor Model was in no way superior to either the seven or eighteen factor models when looking at relationships between personality traits and DSM-III disorders, though each had unique advantages. This suggests that there is nothing ‘special’ about the Five Factor Model above other similar suggestions.

Offering more serious opposition to the Big Five is the recent HEXACO model (Lee & Ashton, 2004) that is supported by lexical studies across several languages (Ashton & Lee, 2001; Ashton et al., 2004; Ashton, Lee, & Son, 2000) and finds personality to be better described using six traits: Honesty-humility, Emotionality, Extraversion, Agreeableness, Conscientiousness and Openness to Experience. The HEXACO Personality Inventory has thus been developed as a potential successor to the NEO PI R and at least one study has shown it to have significant predictive validity advantages (Ashton & Lee, 2008). When Ashton and Lee then added the Honest-humility dimension to the NEO PI R it was shown to improve though still not to the level of the HEXACO PI.

This then would seem to suggest that no, five factors are not sufficient to fully explain human personality and that at least one more is required. However, as demonstrated by the variety of trait models that exist, more or fewer may be necessary and it is likely only a matter of time before another improved selection of factors becomes the accepted model. Theoretically, while different factor models use more or fewer ‘global’ dimensions, these should be made up of similarly descriptive sub-traits. In a sense the difference then is more in the presentation, meaning that arguably a Five Factor Model could be no more and no less adequate in explaining human personality as long as you examine the underlying facets. It could be argued that a Five Factor Model is adequate in explaining personality, but that it could be explained just as well with a three or eighteen factor model.

However support for any of the factor models of personality all rely on the assumption that personality and/or behaviour can be determined using a trait theory at all. This underlying assumption is challenged by situationism which states that personality is in fact an ‘illusion’ and that individuals do not act consistently across different situations. Here it is the social context, rather than any intrinsic personality, that determines our behaviour (Krahe, 1993). Studies such as Milgram’s famous obedience experiment (1963) demonstrate that seemingly anyone can be coerced into acting in unusual ways under the right circumstances. However an extreme situationist stance can not explain variance in the voltage that participants were willing to administer, and in neither this nor similar studies was an attempt to measure personality traits made (perhaps an interesting area for future research).

While it is unlikely that personality is completely a construct of the situation (otherwise why do we observe the individual differences we do?), situational factors are still likely to play a role. This is something that the Five Factor Model does not address and so again we see that it is not able on its own to explain human personality. Another more fundamental aspect that is lacking in the Five Factor Model is a description of how individuals develop their personalities. As it is it is a descriptive model that offers no real explanation for human personality. For that reason alone it is insufficient in explaining the phenomenon. There are however explanatory accounts of the Five Factor Model such as the ‘Social Investment Theory’ (Brent et al., 2004), which uses an interactionist approach explaining the factors as arising through a combination of environmental and biological influences (as supported by the aforementioned twin studies).

In conclusion then, the Five Factor Model is a useful tool for predicting behaviour and explaining why certain individuals act the way they do. It seems that on their own these five factors may not be sufficient for giving a complete character profile, but are more useful in providing a ‘snapshot’ of sorts. It certainly does not on its own ‘completely’ explain human personality, particularly when you consider environmental factors, or the fact that it offers no explanation for how personality develops among other issues. Furthermore, it is not the only useful tool for predicting behaviour and the decision to use five factors, rather than say three or sixteen, seems to be largely subjective.



Leave a Reply

Your email address will not be published. Required fields are marked *

Recommended Articles