Longest ever personality study finds no correlation between measures taken at age 14 and age 77

Imagine you've reached the fine age of 77 and you hear news of a school reunion. You're going to have the chance to meet up with several of your former classmates who you haven't seen since you were fourteen-years-old. They'll look a lot different, of course, but what about their personality? Will they be broadly the same as they were back then?

Past research that's looked at trait changes from adolescence to mid-life has shown there tends to be a moderate amount of stability, so too research that's looked at changes from mid-life into old age. Put these two sets of data together and you might expect to see at least some personality stability across an entire lifespan. Your classmates probably won't have changed completely.

Yet that's not what a recent open-access study in Psychology and Aging has found: the first – to the authors' knowledge – to measure personality in the same people in their adolescence and then again in old age. By covering a period of 63 years, this in a sense is the longest ever personality study.

But contrary to what we might expect based on previous findings, Matthew Harris and his colleagues at the University of Edinburgh failed to find a significant correlation between their participants' personality scores at age 14 and their scores on the same items at the age of 77. "Personality in older age may be quite different from personality in childhood," they said.

The story begins in 1950 when a group of researchers asked teachers to rate the personality of 1,208 fourteen-year-olds in Scotland (actually a sub-sample from a much larger study examining almost every child born in the country in 1936).

The teachers rated the teens on six questionnaire items each addressing their Self-confidence, Perseverance, Stability of Moods, Conscientiousness, Originality and Desire to Learn, respectively. Ratings for these six items correlated strongly and indeed, although it's not a perfect match, the teachers were rating the teens on something very close to what today is referred to as trait Conscientiousness (Harris and his team say the items are collectively measuring another similar trait that they call "dependability"). The children also took intelligence tests.

Fast forward to 2012 and Harris and his team managed to find 635 of the teenagers who were assessed back in 1950. One hundred and seventy-four of them, now aged 77, agreed to take part in a new round of testing. The group, including 92 women, rated themselves on the same six items that their teachers had rated them on all those years ago, and they nominated a close friend or relative to rate them on the items too. The elderly participants also completed some intelligence tests and measures of their general wellbeing.

To the researchers' surprise they found there wasn't a statistically significant correlation between the ratings the participants received when they were aged 14 and the ratings they gave themselves at age 77, or the ratings they received at age 77 from a friend or relative. This was true whether looking at the individual personality items, such as Stability of Moods, or at a single "dependability" trait based on amalgamating across the six items.

Moreover, although dependability at age 77 correlated with current wellbeing, the participants' dependability at age 14 was not linked with their wellbeing in late life, seemingly contradicting past research that's found higher scores on the related trait of conscientiousness are associated with superior wellbeing decades later.

Does this mean that our personalities in old age will bear no relation to our personalities in mid-adolescence? Such an idea is not completely far-fetched. Late adolescence and early adulthood are times of significant personality development and change; old age too can be a time that our characters evolve.

Because of the impressive time span covered by this research, the participants will have lived through both these periods of major change, not to mention the accumulation of subtler personality adjustments through their lives.

"The longer the interval between two assessments of personality, the weaker the relationship between the two tends to be," Harris and his team concluded. "Our results suggest that, when the interval is increased to as much as 63 years, there is hardly any relationship at all."

That's the headline result, but a psychology study covering such a long period of time faces many hurdles, among them the fact that personality theory has changed a lot over the decades.

Today, there's a near-consensus that personality is best conceptualised as being made up of five main traits (including Extraversion, Neuroticism etc) but that wasn't the case back in 1950, which is part of the reason for the rather superficial and incomplete assessment of personality. It's possible that if the participants had been assessed on a comprehensive, modern personality scale at age 14 and again at age 77, that we would see at least some correlation in scores.

Other methodological problems include the fact the teachers' ratings of the teens were likely biased by their knowledge of the teenagers' academic prowess (indeed, the personality scores they gave the teens correlated with the teens' IQ; though even after attempting to "correct" for this in some complex analysis, there was still little evidence of personality stability across the decades).

Yet another problem is that the sample that the researchers were able to reach in 2012 was a highly selected subgroup of the original sample, scoring much higher on average on dependability and intelligence: this may also have affected the ability to detect signs of personality stability. All of which makes it very difficult to interpret these new findings, interesting though they are.

Further reading

—Personality Stability From Age 14 to Age 77 Years

About the author

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