The models above consider traits to be perfectly stable. As noted earlier, there is now extensive evidence indicating that personality shows plasticity in childhood, with long-term test-retest correlations of r ≈ 0.35, and continues to change across the lifespan, although personality consistency gradually increases up to r ≈ 0.75 after the age of 50 (Roberts & DelVecchio 2000). Models of personality-psychopathology relations can be expanded to recognize the malleability of traits (e.g., Ormel et al. 2001). For example, one can posit a dynamic precursor model4 in which early temperament defines the baseline level of risk but subsequent experiences modify personality liability to depression. This model explains variability in disorder onset as a function of the initial level of risk and steepness of the trait trajectory over time. Given the evidence on patterns of personality continuity and change (Roberts & DelVecchio 2000), it appears likely that trait vulnerability is more malleable early in life, but significant life events can alter its trajectory even in old age. A depressive disorder is thought to emerge when personality liability crosses the threshold. Thus, individuals
