The effects of noncognitive traits such as personality on various life outcomes are well-documented, though potentially different across cultures. In particular, personality traits have been found to have a significant effect on income.
A question that remains to be answered, however, is through what channels personality traits operate. Do they affect wages because they make people more productive, or do people with certain traits self-select into occupations with higher wages? Maybe, certain personality traits are just correlated with better bargaining skills?
Cubel et al. (2014) ran an experiment with 359 undergraduate students from the University of New South Wales to investigate this question.
(Note: the post uses Big Five personality traits. If you’re unfamiliar with these, read up on them briefly here.)
Past research suggests that conscientiousness has a positive effect on wages while agreeableness and neuroticism have negative ones. In a previous post of mine I discussed the link between personality and career choice. Indeed, there seem to be significant career choice differences across personalities.
That post suggests that conscientious people are more likely to study medicine, a high paying occupation in the U.S., but not necessarily so elsewhere (plus this result may be particular to the Netherlands where the study in question was conducted, see details in the post). Agreeable males are less likely to study business, perhaps hurting their earnings prospects on average, while neurotic people are more likely to study humanities, clearly a low wage major.
As for extroversion, there is a higher chance of studying business and law (both well-paying careers), but a lower chance of studying STEM subjects. These two effects could reasonably cancel each other out. For openness, there is a higher chance to study law for males, and a higher chance to study humanities for females. These two effects could once again cancel each other out, as the former is a high-wage, the latter is a low-wage major.
So yes, from this simplistic analysis it does seem that self-selection into particular occupations can explain at least some of the wage differences across personality traits. What Cubel et al. (2014) investigate, however, is whether personality traits also affect productivity directly.
As mentioned above, they run an experiment. It involves undergraduate students who have to solve a mentally taxing task. Specifically, they are asked to add five 2-digit numbers up, again and again and again for 10 minutes. The more such problems they can solve, the more they get paid. They get $4 per correct answer, in Australian dollars.
The task is meant to be gender-neutral. While we know that there are gender differences in math performance, the authors cite some paper that claims that this applies to more abstract problems and not to simple arithmetic/algebra. A claim I cannot confirm, but find reasonable enough to believe.
Also performance on this task is not affected by the self-selection of particular personalities into occupations, so we can control for that. Neither is it affected by bargaining power. It is meant to be a measure of pure productivity; specifically of concentration, effort, stress management (as the task is timed) and perseverance/industriousness.
So let’s look at the results of the experiment by personality trait. It must be noted that the authors controlled for age, gender, major, level of study (BA, MA, PhD), parental income and GPA (as a proxy for cognitive skills).
Conscientiousness. Seems to significantly positively influence performance. There doesn’t seem to be a difference in the effect of this trait across genders or majors.
Neuroticism. Negatively influences performance, and this one seems to have the most robust effect. There is no difference between how this trait affects females and males, but non-science majors are not hurt as severely by a given level of neuroticism as science majors are.
Openness. Doesn’t really have a robust effect, though the sign is always negative but mostly insignificant. However, there is interesting heterogeneity by gender and major. Females and non-science majors are hurt significantly by high levels of openness whereas males and science majors are not. For the latter group, the trait is neutral. The authors posit this may be because females and non-science majors are more likely to score high on openness because they’re artistic and similar, whereas males and science majors score high on other dimensions of openness such as inventiveness.
Agreeableness. Similarly to openness, there is no robust effect but the sign is always negative. No heterogeneity either.
Extroversion. There is no effect on performance, but here as well there is interesting heterogeneity. Females, who generally score high on the warmth, gregariousness and positive emotions dimensions of extroversion, are actually hurt by this trait. Males, who generally score high on the assertiveness and activity dimensions of extroversion, are helped by this trait.
No significant heterogeneity by family income or parental education was found.
Thus the conclusion of the authors is that neuroticism and conscientiousness are important for wage determination because they significantly affect productivity. On the other hand, agreeableness and openness probably operate more through other channels such as selection into particular occupations.
It may be noted here that as both the effects and the frequency of certain personality traits differ by gender, it is quite likely that personality differences can at least partially account for the gender wage gap.
Finally, it’s interesting to see how different the things a given Big Five trait covers can be. For instance, openness can be divided into two quite distinct dimensions of artistic openness/creativity and scientific curiosity/inventiveness. And as we can see these dimensions within a particular trait can have different effects. This is something to pay attention to in future empirical research, and consequently it may be important to examine whether these sub-dimensions within traits could be theoretically formalized and standardized as well. But this is probably something for psychologists to do.