The notion of schools as "loosely coupled" organizations has been widely discussed in the research literature. Many argue it is either a protective mechanism for schools to buffer external pressure or a barrier for implementing new reforms. Against the backdrop of systemic change and accountability, we applied a two-level hierarchical linear model to nationally representative data in the US, testing the "loosely coupled" theory through examining the association between data-informed improvement efforts at the school level and data-informed instruction at the classroom level. Statistically significant associations were identified but with a small proportion of variance explained, indicating that the top-down systemic change strategy failed to tighten the system as intended. Alternatively, bottom-up strategies, such as professional learning communities, which operate under the assumption of working with loose coupling, should be considered.