Perceptions of how occupants adopt water conservation behaviors under psychosocial processes: A complementary dual-stage SEM-ANN perspective

This study delved into socio-psychological determinants driving residential water curtailment actions, an area relatively overlooked in the existing literature. The theory of planned behavior was innovatively expanded with moral norms, perceived expectations, self-identity, and risk-related components. The proposed framework was empirically examined by conducting an online self-administered survey of 343 citizens residing in Isfahan, Iran. What sets this research apart is introducing a new multi-analytical hybrid approach integrating linear and non-linear techniques to leverage their respective strengths. Artificial Neural Network (ANN), coupled with Sobol sensitivity analysis, was employed to robustly validate Structural Equation Modeling (SEM) results regarding the framework's explanatory power and predictors' ranking. Through providing reliable performance, the ANN analysis confirmed SEM findings, revealing that the significant influences of moral norms and self-identity on intention are mediated by attitude and perceived behavioral control. Moreover, self-identity and attitude were the strongest direct predictors of behavior and intention. Several policy recommendations were proposed, with the following highlighted as crucial: (1) fostering favorable water conservation attitudes, which entails targeting individuals' moral considerations and inducing self-reward feelings within the community; (2) cultivating a sense of water conservation identity among citizens; and (3) boosting individuals’ self-confidence and facilitating their engagement in water conservation behaviors.