The challenge of separating effects of simultaneous education projects on student achievement

When multiple education projects operate in an overlapping or rear-ended manner, it is always a challenge to separate unique project effects on schooling outcomes. Our analysis represents a first attempt to address this challenge. A three-level hierarchical linear model (HLM) was presented as a general analytical framework to separate program effects while taking into account the hierarchy in educational data. The HLM model was then applied to data from the Commonwealth Accountability Testing System that the State of Kentucky has implemented for years, in an attempt to separate the effects of two education projects aimed at improving mathematics and science education in the Appalachian region: the Appalachian Rural Systemic Initiative (ARSI) closely followed by the Appalachian Mathematics and Science Partnership (AMSP). Even though the HLM model successfully separated ARSI and AMSP effects, relevant statistical issues were discussed to improve future efforts in separating effects of simultaneous education projects on schooling outcomes.