Can you actually change your personality?

There is a version of you that you keep almost being. Slightly calmer about the things that aren’t worth it. Slightly better at the thing you said you’d do on Monday. You have wanted this for a while, which is exactly what makes the question uncomfortable: if wanting it were enough, it would presumably have happened by now.

The research on this has improved a lot in the last decade, and it has a specific answer — including a fairly blunt finding about what wanting achieves on its own.

The short answer

Yes, and the effect is modest rather than transformative. A systematic review of 207 intervention studies found personality traits shifted by d = .37 over an average of 24 weeks, with emotional stability moving most (d = .57). The changes did not fade: at a year or more after treatment the effect was still d = .37.

But wanting it does nothing by itself. In a 15-week study of 377 people, participants could accept weekly behavioural challenges. Merely accepting them did not predict any change. Only completing them did. People who completed none were expected to decline slightly on the trait they wanted to grow.

And your personality is already changing without your involvement. Across 92 longitudinal samples, people reliably become more conscientious, more emotionally stable and more socially dominant — most sharply between 20 and 40, and continuing well beyond.

How much can actually change?

Enough to notice, not enough to become someone else — and the honest number is smaller than the headline.

The 207-study review is the best evidence base available: d = .37 overall, largest for emotional stability at .57, smaller for extraversion at .23. Type of therapy didn’t matter much. People with anxiety disorders changed most.

Now the caveat that matters, because it is the sort of thing that separates a real answer from a motivational one. When the authors restricted to experimental studies with control groups, the raw effect was d = .43 — but the publication-bias-corrected estimate fell to d = .13, with a confidence interval whose lower bound is close to trivial. Roughly a third of the headline.

So: interventions do move trait measures, durably, and the best-controlled estimate of how much is modest. Both halves of that sentence are load-bearing.

The strongest non-clinical evidence comes from a randomised trial of 1,523 people using a three-month smartphone programme. Self-reported change matched intentions (d = 0.52 for people wanting to increase on a trait), and — the part that makes it convincing — friends, family and partners also noticed (d = 0.35). The changes were still present three months after the programme ended.

One asymmetry worth knowing: for people trying to decrease on a trait, observers did not reliably detect change (d = −0.22, not significant). Others noticed the growth; they did not notice the reduction.

Why hasn’t wanting it worked so far?

Because intention and behaviour come apart, and the research now separates them cleanly enough to show exactly where.

Hudson and colleagues (2019) gave participants weekly challenges — small prewritten behavioural tasks matched to the trait they wanted to change. Participants could accept a challenge and then not do it, which made it possible to test intention against action.

Accepting challenges predicted nothing. Completing them predicted change. Someone who completed no extraversion challenges was expected to decline by 0.04 standard deviations per month. Someone completing roughly two a week gained 0.05 SD per month — about 0.17 SD over a 15-week semester.

That is a small number, and the confidence interval’s lower bound (0.03) sits near zero. But the direction of the comparison is the useful part: the difference between wanting and doing was the difference between drifting slightly backwards and moving slightly forwards.

The effects held for conscientiousness and emotional stability. They did not hold for agreeableness or openness — so not every trait is equally available to deliberate effort.

What actually works?

The most-supported technique is unglamorous and specific: decide the situation in advance.

Across two 16-week experiments, Hudson and Fraley (2015) tested two ways of supporting change. One did nothing. The one that worked trained people to form implementation intentions — “when situation X arises, I will do Y” — rather than general resolutions to be different.

This is the same mechanism that shows up in the procrastination literature, and it works for the same reason: it removes the moment where motivation gets a vote. “Be more outgoing” has no trigger and no action. “When someone asks how my weekend was, I’ll ask them a question back” has both.

The practical shape that emerges from all of this:

• Pick the behaviour, not the trait. You cannot do conscientiousness. You can put your keys in the same place.
• Attach it to a cue. The if-then structure is the specific thing that was found to work.
• Expect months, not weeks. The studies here run 15 to 24 weeks, and the smartphone trial ran three months.
• Pick a trait that responds. Emotional stability and conscientiousness moved in these studies. Agreeableness and openness did not.

Will I stay changed, or does it snap back?

The evidence available says it holds — over the timescales anyone has checked.

In the 207-study review the effect did not decay: d = .34 immediately after treatment, .48 at six months, .46 at six-to-twelve months, and .37 at a year or more. No marked decrease with time. The smartphone trial found changes still present three months after it ended, observed by other people as well as self-reported.

What nobody has is a ten-year follow-up on deliberate change. The durability evidence is good for a year and unknown beyond it.

It is also worth knowing what the background rate looks like, because it is not zero. Rank-order stability — how well your position relative to other people holds — rises from .31 in childhood to .54 in the college years, .64 by 30, and plateaus around .74 between 50 and 70. High, and never approaching 1.0, at any age. Your relative position is durable, not fixed, and the ceiling on that is set by nature rather than effort.

Meanwhile the maturity principle says the average person is already moving in the direction most people say they want: more conscientious, more emotionally stable, more socially confident, with the steepest change between 20 and 40 and meaningful change continuing into middle and old age. The old claim that personality is “set like plaster” by 30 does not survive the data.

So is it worth trying?

On the evidence, yes — with the expectations adjusted.

You will not become a different person. The best-controlled estimate of deliberate change is modest, and the traits most responsive to effort are not the whole of you. But the change is real, other people can see it, it persists for at least a year, and it is reliably driven by completing specific behaviours rather than by resolving to be different.

The most useful thing in this entire literature might be the least inspiring sentence in it: people who wanted to change and did nothing got slightly worse at the thing they wanted. Wanting is not a first step. It is not a step at all.

If you are going to try, get a baseline first. A change of 0.17 standard deviations over a semester is not something you will detect by memory, and the difference between “this is working” and “I feel like it’s working” is the whole reason to measure.