Personality differences uncovered between students at different US universities

Students at more expensive institutions tended to score higher in trait Neuroticism 

By Christian Jarrett

Psychology is overly dependent on student samples, but on the plus side, you might assume that one advantage of comparing across student samples is that you can rule out the influence of complicating background factors, such as differences in average personality profile. In fact, writing in the Journal of Personality, a team of US researchers led by Katherine Corker at Kenyon College has challenged this assumption: their findings suggest that if you test a group of students at one university, it’s not safe to assume that their average personality profile will match that of a sample of students from a university elsewhere in the same country.

Corker and her colleagues measured the personality of over 8,500 students studying a range of majors (including psychology, business and nutrition) at 30 colleges and universities across 20 different US states. They found some significant differences in average student personality between different sites. The amount of difference that was explained by the site of testing was modest – about 1 to 3 per cent – but Corker and her colleagues said that this “should not be dismissed as necessarily trivial or unimportant”.

Among the site-specific effects, larger universities tended to have more extraverted students; more urban and diverse universities had more open-minded students;  universities requiring letters of recommendation had more agreeable students; public colleges had less agreeable students than private colleges; and more expensive colleges had higher trait Neuroticism. Differences like these could reflect students with particular personality profiles being drawn to particular institutions; selection could be at play, in the sense of university selectors showing a preference for particular personality types; and also students’ personalities could be shaped by the culture of their university.

“All told, these results suggest there is more variability between students at different colleges and universities than some researchers might have expected,” Corker and her team said, though they warned their colleagues not to leap to this finding as an explanation for the replication crisis in psychology (the difficulty labs often have in trying reproduce earlier findings reported by researchers based at another institution), at least not until there is more data and a thorough theory in place to account for the site differences in personality.

It’s a shame the current research didn’t include students studying a more diverse set of subjects, but if anything, this makes it likely the observed between-institution differences are an underestimate. Perhaps for now the lesson to take is that if you’re comparing data from two sets of students at different institutions, it’s unwise to be overconfident about how similar they are likely to be.

College Student Samples Are Not Always Equivalent: The Magnitude of Personality Differences Across Colleges and Universities

Christian Jarrett (@Psych_Writer) is Editor of BPS Research Digest

3 thoughts on “Personality differences uncovered between students at different US universities”

  1. Quoting from above “and also students’ personalities could be shaped by the culture of their university.”

    I think, the culture of the university is external and not so pivotal as the individual mindset and mental makeup of a student that shapes their personality. The culture at home fosters the cultivation of those habits that are valued and cherished by their family. The habits in turn shape the student who impact the environment by their contribution to the culture outside.

  2. It makes sense. For instance in research I’ve done on the Absorption scale and hypnotizability in different universities, I’ve found different results based on the location. That said, one factor the authors do not consider is that it may just be a function of sampling error.

    This can be shown using a Monte Carlo simulation. Simply generate a population, with a population relationship of r = .40. Take 50 random samples of varying sizes without replacement. This would be analogous to the different university samples. You will find that the sample estimates will vary a fair degree, in my simulation this ranged from r= .1 through r=.70. Again these are similar to what Corker et al found. Now if you average the individual results, weighted by sample size, you’ll find that the mean relationship will be very close to to the original population relationship r = .4.

    It is very probable that a similar mechanism is operating in this case.

Comments are closed.