Several studies have suggested that the rank-order stability of personality increases until midlife and declines later in old age. However, this inverted U-shaped pattern has not consistently emerged in previous research; in particular, a recent investigation implementing several methodological advances failed to support it. To resolve the matter, we analyzed data from two representative panel studies and investigated how certain methodological decisions affect conclusions regarding the age trajectories of stability. The data came from Australia (N = 15,465; Study 1) and Germany (N = 21,777; Study 2), and each study included four waves of personality assessment. We investigated the life span development of the rank-order stability of the Big Five for 4-, 8-, and 12-year intervals. Whereas Study 1 provided strong evidence for an inverted U-shape with rank-order stability declining past age 50, Study 2 provided more mixed results that nonetheless generally supported the inverted U-shape. This developmental trend held for single personality traits as well as for the overall pattern across traits; and it held for all three retest intervals—both descriptively and in formal tests. Additionally, we found evidence that health-related changes accounted for the decline in rank-order stability in older age. This suggests that if analyses are implicitly conditioned on health (e.g., by excluding participants with missing data on later waves), the decline in stability in old age will be underestimated or even missed. Our results provide further evidence for the inverted U-shaped age pattern in personality stability development but also extend knowledge about the underlying processes.
Keywords: personality development, rank-order stability, Big Five, panel studies, local structural equation modeling