This article identifies elements and connections that seem to be relevant to explain persistent aggregate behavioral patterns in educational systems when using complex dynamical systems modeling and simulation approaches. Several studies have shown what factors are at play in educational fields, but confusion still remains about the underlying mechanisms driving observed outcomes and therefore more guidance is needed. The framework suggested here throws some ideas in that direction stressing the relevance of nonlinear complex interactions via feedbacks between education systems' transition rates--intake, repetition, dropout, and promotion--and schooling outcomes. Schooling outcomes reciprocally influence transition rates in the system generating aggregate patterns that continuously change (and are changed by) the inputs that endogenously determine them. Furthermore, this paper underscores practical and theoretical limitations of traditional quantitative models that can be addressed with a complex systems analysis and suggests future lines of investigation. Specifically, this article advocates a complexity approach ruled by the laws of thermodynamics to help detect corrupt practices in education systems and improve accountability and governance in such systems.