The goal of this meta-analysis was to improve the understanding of life satisfaction and domain satisfaction, by synthesizing the available longitudinal data. First, we meta-analyzed the stability of individual differences in these constructs, using multilevel mixed-effects models. Second, we meta-analyzed their concurrent associations, using multilevel random-effects models. Third, we tested for their prospective effects, using two types of models with different modeling assumptions (cross-lagged panel models, dynamic panel models). We included seven domain satisfactions (e.g., romantic relationships, health). Data came from 98 samples, including 252,647 participants. The results indicated high rank-order stability of life satisfaction and domain satisfaction and moderate to strong concurrent correlations between these constructs. Hence, the results support the notion that life satisfaction and domain satisfaction can be understood as trait-like characteristics and are substantially associated with each other. As regards their prospective effects, however, the two models suggested a different pattern of findings. More precisely, the findings indicated that controlling for stable-between differences in the dynamic panel models altered the overall pattern of prospective effects. This suggests that explanations other than reciprocal effects should be considered and examined in future research (e.g., genetics, dispositions) and highlights the crucial role of modeling decisions when analyzing cross-lagged effects.
Bühler et al. (Wed,) studied this question.