Fig.1

Illustration of the analysed autoregressive models including measures of SWB (measured by the WHO-5 well-being index, and the satisfaction with life scale) and personality traits (measured by the big five inventory) at baseline and at 15–18-month follow-up. The arrows indicate: a = SWB Prospective Effects, i.e. effects of SWB at baseline on Trait at follow-up; b = Trait Stability, i.e. effects of Trait at baseline on Trait at follow-up; c = Trait Prospective Effects, i.e. effects of Trait at baseline on SWB at follow-up; d = SWB Stability, i.e. effects of SWB at baseline on SWB at follow-up; e = Associations between SWB and Trait at baseline; f = Associations between SWB and Trait at follow-up. The error terms for the same item (e.g. T3) were allowed to correlate across the two time points. For measurement invariance, factor loadings (e.g. t3), intercepts, and residuals of the same item were constrained to be equal at the two time points. Parameters b and d indicate rank-order stability and are not affected by possible mean level changes in SWB/personality traits