Abstract

In this article we defend that the adoption of a non-linear approach, theoretically framed on complexity theories can make some contribution to the bottom-up approach, which explains the levels of satisfaction with life as a whole through the combination of the levels of satisfaction in different life domains. Two approaches have been tested: (Rojas in J Happiness Stud 7:467-497, 2006) constant elasticity of substitution model and the model with quadratic terms and interaction effects (Gonzalez et al. in Soc Indic Res 80:267-295, 2006; Gonzalez et al. in Qual Quant 42:1-21, 2008). In order to prevent obtaining false non-linear relationships they have been analysed twice taking into account or not limited measurement of satisfaction with life as a whole. Results show that: (a) any of the two non-linear models fits better than the linear one; (b) any of the models failing to take into account limited measurement fits worse; (c) the non-linear model with quadratic terms and interaction effects fits better than Rojas'. The implications for the study of psychological well-being are discussed.