So, this was a longitudinal study of Big Five changes. That there are changes is already known, but this bit may be helpful to me in the future:

Later studies have since suggested that personality is instead set like “soft” plaster, in that personality does change, albeit only marginally, beyond 30 and across the entire life cycle (Srivastava, John, Gosling, & Potter, 2003). Some of this research is problematic to interpret, as it has been based on cross-sectional differences in the mean level of personality traits across age groups, which could represent either real change or simply cohort effects (where, for example, people at a certain age only appear to have different personality profiles due to events that historically happened to their cohort in youth). More recent longitudinal research, however, suggests that personality change does take place, with the same people giving different responses to personality questionnaires on different occasions (e.g. Helson, Jones, & Kwan, 2002; Lucas & Donnellan, 2011; Roberts, Walton, & Viechtbauer, 2006a; Roberts, Wood, & Caspi, 2008) and as such a relatively broad consensus that personality does change has developed (Costa & McCrae, 2006; Roberts, Walton, & Viechtbauer, 2006b)

What does this do usefully?

This approach could potentially contribute to both the economic and psychological literatures in several ways. First, if personality change can be linked to change in another variable, then it would suggest that such changes are substantive rather than simply inconsistent responding. Second, it would allow a more direct comparison of the relative degree to which personality explains life satisfaction compared to economic indicators, such as income. Such a direct comparison is not currently possible, and claims of the relative predictive ability of personality on the various components of subjective well-being are complicated by relying on comparisons across studies with non-compatible methodologies (e.g., through comparing income effect estimates based on within-person changes and personality effect estimates based on between-person estimates). Third, this direct comparison will allow an estimate of the relative magnitude of the change in personality variables compared to the changes in variables commonly considered to be variable (e.g., income).

I don't actually regard #1 as in serious question, but #2 and #3 are certainly interesting.

Neuroticism and agreeableness are the personality traits that have the strongest predicting effect and a comparison can be made with other predictive factors. For example, all things being equal, being unemployed is associated with the same amount of life satisfaction as being around 0.75 standard deviations lower in levels of neuroticism than the average, and a one percent higher income is associated with the same amount of life satisfaction as being approximately 0.01 standard deviations higher in agreeableness.

For example, our results suggest that a one standard deviation change in openness to experience is associated with approximately the same change in life satisfaction as would a AUD $61,000 (~USD $62,000) increase in annual household income. The dollar values for one standard deviation changes in the other personality traits are as follows: Conscientiousness – AUD $91,000 (~USD $92,000), Extroversion – AUD $222,000 (~USD $225,000), Agreeableness – AUD $147,000 (~USD $149,000), Neuroticism – AUD $309,000 (~USD $314,000). The average annual household income is around AUD $87,000 (~USD $88,000) each year so these dollar values could be considered as high and therefore highlight the importance of changes to personality in the relationship with higher life satisfaction.

To me, this suggests a way to test causality: we discussed earlier the recent research which seemed to demonstrate that psilocybin increased Openness. Record income for experimental and control groups, see if the relationship is maintained.

(I was going to suggest meditation as another possible intervention, but a quick googling seems to indicate meditation has no clear causation or correlation with Big Five factors.)