Personality Factors in 4-D: Characteristic Constellations

Hesitate to call something as complex as extraversion or introversion "a" trait.

Posted Jun 30, 2019

Wikimedia Commons
Source: Wikimedia Commons

We have traits of many kinds. As opposed to physical traits, our specific bodily characteristics, personality traits are specific psychological characteristics or predispositions to behave in certain ways. Trait theorists are primarily interested in studying human personality by measuring habitual patterns of actions, thoughts, and feelings - in other words, by measuring personality traits. Gordon Allport, often called the father of personality, identified thousands of terms to describe personality (Allport & Oddbert, 1936). Personality psychologists trying to detect some order in that chaos observed that certain traits tend to group together in trait clusters (Cattell, 1943), constellations of characteristics that are commonly correlated with each other among many people.

Sometimes, though, we are simply seeing patterns that are not there, Illusory correlations, variables we mistakenly perceive as related even though they are not. Even when a pattern is present, its existence is not self-explanatory: The correlation does not prove its own causation. Correlation (the statistic identifying that variables are related) does not explain why they are related. Outgoing people tend to be less fearful than others, which might happen if developing an outgoing nature reduces fear, but the reverse is possible as well if fearfulness makes a person less outgoing. In many cases, some other variable (say, brain cell activity levels) influences both with no causal relationship between the things we see as correlated. 

Trait clusters are whole groups of characteristics that correlate together in both positive and negative directions, better known these days as personality factors because factor analysis identifies the groupings. Sometimes these clusters only show up in certain research samples; that is, the group of people surveyed in those cases sows a lot of correlations that other people do not. Personality researchers sought to identify which personality factors are fairly universal (meaning the traits will cluster together in any group of people measured) and orthogonal (statistically unrelated to each other). Two of the earliest to emerge as universal and orthogonal were the factors of extraversion/introversion and neuroticism/emotional stability. Eventually, popular models emerged in which five (the "Big 5") or six (HEXACO) personality factors, including those two, appeared to be fairly universal and orthogonal, although not everyone agrees. It is incorrect or, at the very least, a drastic oversimplification to refer to the "Big Five" as five personality traits because each factor is made up of many traits.

Take extraversion, for example. If you call someone an extravert, you're referring to many characteristics, subsuming them under one umbrella term. While the first adjective that springs to mind when people think of extraversion tends to be outgoing, the extravert is also likely to be assertive, bold, bored if alone, externally focused, gregarious, risk-taking, socially interactive, and talkative. A single person is unlikely to have every extraverted trait, but when the individual regularly shows many such characteristics, we label him or her an extravert (Loo, 1979).

While calling extraversion or introversion a trait as though it existed on a linear high-low scale is a drastic oversimplification, those who see it as a personality factor more often refer to it as a dimension. A two-dimensional model works better than a one-dimensional line. Two people can achieve identical scores on a numerical extraversion/introversion test, and yet the scores are not equivalent because different traits get them there. Suppose someone scores 67% in the extraverted direction on whatever scale is used. If it's treating individual responses as binomial (either/or, yes/no), two people with that same score (2/3 of the way to the 100% mark) might have given identical answers on only 1/3 of the test items. 

Person A 

first third extravert + second third extravert + last third introvert = 2/3 extravert.

Person B

first third extravert (same as A) + second third introvert (opposite of A) + last third extravert (opposite in the other direction) = also 2/3, equally extraverted despite having a majority of traits unlike those shown by the other person.

So calling extraversion/introversion a trait or even a dimension might still fall short of offering an adequate model, thereby making it necessary to thinking of the personality factor more like a two-dimensional plot or even something more complicated than that. And that's just one personality factor, one constellation of characteristics. A model of personality that assesses the individual in terms of multiple clusters might need to be three-dimensional or, better yet because we fluctuate over time, four dimensions.

See Langley (2016) for the first part of this line of thought. 


Allport, G. W., & Oddbert, H. S. (1936). Trait-names: A psycho-lexical study. Psychological Monographs, 47(1), i-171.

Cattell, R. B. (1943). The description of personality: Basic traits resolved into clusters. Journal of Abnormal & Social Psychology, 38(4), 476-506.

Langley, T. (2016). The two factors - extraversion and neuroticism. In T. Langley (Ed.), Doctor Who psychology: A madman with a box. New York, NY: Sterling.

Loo, R. (1979). A psychometric investigation of the Eysenck Personality Questionnaire. Journal of Personality Assessment, 43(1), 54-58.