Identifying high impact school improvements using conditional mean independent correlations and growth functions

Abstract

In educational contexts where many domains subject to improvement are interdependent and causal evidence is frequently lacking it is difficult, if not impossible, for policymakers and educational practitioners to decide which domain should be invested in. This paper proposes a new method that uses Conditional Mean Independent Correlations (CMIC) and normative growth functions to inform such decision-making processes. In this paper, CMIC and growth functions are applied to data from a research-practice partnership to identify high impact improvements among domains that are considered important to the district’s mission and vision around student learning. The results point to improvement domains that administrators did not consider to be high impact improvements initially, suggesting that this method brings leaders food for thought around strategies for improvement efforts. The CMIC and growth functions moreover accommodate opportunities for policymakers and practitioners to base their decisions on theory and data, providing them with a stronger degree of decision-making authority for use of resources for improvement. Simultaneously, CMIC and growth functions enable researchers to test and further develop theoretical models on improvement efforts. Limitations and suggestions for further research are discussed.