Volitional Personality Change

What it is

Hudson and Fraley ran two sixteen-week intensive longitudinal studies on undergraduates at the University of Illinois at Urbana-Champaign, each with a randomised intervention arm. Participants stated change goals for each of the five broad traits, then reported their traits and their trait-relevant daily behaviour repeatedly across the sixteen weeks. Change goals predicted growth in the direction people wanted.

The effects are the point. Pooled across twelve samples in a later mega-analysis, people with high change goals grew roughly a sixth of a standard deviation more than peers with average goals over sixteen weeks for extraversion and emotional stability, and roughly a twentieth of a standard deviation more for agreeableness and openness. Those are the largest and the smallest effects in that analysis, and both are small.

The follow-up work sharpened the mechanism. Hudson with Daniel Briley, William Chopik and Jaime Derringer found in 2019 that goals alone were not enough, and that completing the behavioural challenges was what predicted change. The strongest evidence in the field is a three-month randomised controlled trial of a smartphone coaching application by Mirjam Stieger and colleagues, published in 2021, where the intervention group changed more than controls, the changes aligned with intended goals, and informant reports corroborated them, although informant-reported change was smaller than self-reported change.

In effect

The concept is also the library's cleanest specimen of a failure that follows a correct citation, and the criticism is of an argument rather than of a person. Benjamin Hardy's Personality Isn't Permanent cites Hudson and Fraley accurately, with the author order right, which is more than most books making this argument manage. He then concedes on the same page that the research is conservative in expected change, at least in the short term, which is the qualification that cuts against his own thesis and he printed it anyway.

Two sentences later the concession is converted by assertion. The modest average effects, he writes, are not evidence of limits but reflect the fact that people on average do not make purposeful changes for emotional and situational reasons, both of which are controllable. Nothing is cited for that clause, and it carries the entire prescription. It fails three ways. It is unfalsifiable as stated, because any observed limit on change is reattributed to a failure of will rather than counted as evidence about how much change is available. It inverts what the study measured, because the participants were self-selected for having change goals, so the modest effect is the effect among people who were already trying, and the explanation that people do not try is not available for that number. And the closest thing in the literature to a test of the claim, the 2019 follow-through paper, is more sober than the book is.

Both sides, kept. The research supports the smaller claim, that personality is changeable at the margins, under sustained effort, in the direction people want. It does not support the larger one, that an individual can choose a personality. This publication agrees with the smaller claim, and notes that the author's own concession is the better guide to the evidence than the sentence that overrides it.

What it does not say

It does not say the traits are plastic. The effects are small, largely self-reported, and measured over months rather than years.

It does not say that wanting is sufficient. The evidence points the other way: completing the behaviour is what predicts change.

It does not licence the self-help reading. An observed limit on how much anyone changes is evidence about the size of the effect, not evidence about anyone's willingness.

---