Sources of Cumulative Continuity in Personality:A Longitudinal Multiple-Rater Twin StudyChristian Kandler, Wiebke Bleidorn,and Rainer RiemannBielefeld UniversityFrank M. SpinathSaarland UniversityWolfgang Thiel and Alois AngleitnerBielefeld UniversityThis study analyzed the etiology of rank-order stability and change in personality over a time period of13 years in order to explain cumulative continuity with age. NEO five-factor inventory self- and peerreport data from 696 monozygotic and 387 dizygotic twin pairs reared together were analyzed using acombination of multiple-rater twin, latent state-trait, and autoregressive simplex models. Correcting formeasurement error, this model disentangled genetic and environmental effects on long- and short-termconvergent valid stability, on occasional influences, and on self- and peer report-specific stability.Genetic factors represented the main sources that contributed to phenotypic long-term stability ofpersonality in young and middle adulthood, whereas change was predominantly attributable to environ-mental factors. Phenotypic continuity increased as a function of cumulative environmental effects, whichbecame manifest in stable trait variance and decreasing occasion-specific effects with age. This study’sfindings suggest a complex interplay between genetic and environmental factors resulting in the typicalpatterns of continuity in personality across young and middle adulthood.Keywords: personality, continuity, genetics, environment, twin studyThe meta-analysis by Roberts and DelVecchio (2000) pro-foundly established that rank-order stability in personality in-creases across the life course until it reaches its peak in lateradulthood after age 50 (Caspi & Roberts, 2001; Fraley & Roberts,2005). This increasing continuity in personality proved to be arobust finding across self- and other reports, independent of genderand the specific trait considered (e.g., Costa & McCrae, 1988;Terracciano, Costa, & McCrae, 2006). Three prominent theoriesprovide rather conflicting etiological explanations for the cumula-tive rank-order continuity with age: the genetic set-point model(Carey, 2002), the genetic maturation hypothesis (McCrae et al.,2000), and a model proposing transactions between genetic andenvironmental factors (Caspi, Roberts, & Shiner, 2005). We ana-lyzed personality assessments of twins on global personality traitsof the five-factor model (McCrae & John, 1992) across threewaves of measurement over a period of 13 years in order to test theadequacy of predictions from each of these etiological theories.Because this is the first longitudinal twin study that included self-and peer reports, we were able to generalize our findings acrossmultiple raters demonstrating convergent validity (Campbell &Fiske, 1959).The Genetic Set-Point HypothesisA number of behavioral genetic studies have led to the conclu-sion that genetic factors primarily contribute to stability in person-ality traits (e.g., McGue, Bacon, & Lykken, 1993; Viken, Rose,Kaprio, & Koskenvuo, 1994). Carey (2002) interpreted this findingin terms of a set-point model, in which environmental fluctuationsare assumed to affect short-term changes (a few days, weeks, oreven several months) in personality, whereas genetic factors de-termine individual set-points to which individuals will regress ona long-term basis. These assumptions are in accordance with theintriguing finding that parental environments seem to have nolong-term influence on personality in adulthood (e.g., Kandler,Riemann, & Ka¨mpfe, 2009; Krueger, Markon, & Bouchard, 2003).According to this theory, variance due to individual geneticset-points does not change across time. The cumulative stability ofpersonality across the life span should thus result exclusively fromdecreasing effects of environmental fluctuations with age. To thedegree that more of the important life transitions occur in youngadulthood (e.g., vocational training, finding a job, starting a fam-ily), this seems to be a plausible assumption. However, there ispowerful evidence that personality stability decreases as the timeinterval between measurement occasions increases (Fraley & Rob-erts, 2005), reflecting long-term rank-order change that is notcompatible with the genetic set-point hypothesis. If phenotypicscores get closer and closer to the genetic set point as a function ofdecreasing short-term environmental influences, then the correla-tion among scores of initial intervals of time will necessarilyincrease across a series of later intervals, even though the retestinterval is also increasing. This prediction is not in line with theresults and continuity functions presented by Fraley and Roberts.Christian Kandler, Wiebke Bleidorn, Rainer Riemann, Wolfgang Thiel,and Alois Angleitner, Department of Psychology, Bielefeld University,Bielefeld, Germany; Frank M. Spinath, Department of Psychology, Saar-land University, Saarbru¨cken, Germany.Correspondence concerning this article should be addressed to ChristianKandler, Department of Psychology, Bielefeld University, Universita¨tsstr, 25,D-33615 Bielefeld, Germany. E-mail: [email protected] of Personality and Social Psychology © 2010 American Psychological Association2010, Vol. 98, No. 6, 995–1008 0022-3514/10/$12.00 DOI: 10.1037/a0019558995Despite these conflicting findings, the genetic set-point model isstill a quite appealing and parsimonious model of personalitydevelopment. Even though the entire model seemed to be toorestrictive, the specific assumption of decreasing effects of situa-tional fluctuations might be a relevant mechanism of increasingstability and thus for personality development, which is worth-while to study in more detail.The Genetic Maturation HypothesisThe five-factor theory (FFT) provides an alternative explanationof personality development proposing that both rank-order conti-nuity and change in traits, considered as basic tendencies (abstractpsychological potentials), are exclusively mediated by geneticfactors (McCrae et al., 2000). According to this hypothesis, sig-nificant environmental effects on traits should merely result fromshort-term contextual influences and systematic as well as randommeasurement error.Researchers of a number of biometric studies have providedsome support for this genetic maturation hypothesis because theyfound stability to be primarily influenced by genetic factors andhave even obtained evidence for the appearance of new