Towards a critical decision-making ecology approach for child protection research