The purpose of this research to investigate the heterogeneity of per-pupil expenditure (PPE) slope estimates in predicting student achievement. The research question guiding this project is: how does the measured relationship between per-pupil expenditure vary across studies that use different models? In concert with SREE's 2012 conference mission of "Understanding Variation in Treatment Effects", this project will inform policy-makers and practitioners on how the distribution of PPE slope estimates varies across different model specifications. A secondary purpose driving this research is to explore methods for synthesizing regression results when the models differ across studies. To date, few methods exist to synthesize regression models when those models use different sets of predictors (Becker & Wu, 2007). The data come from an NSF grant to explore methods for synthesizing regression coefficients, where the authors gathered and coded the education production function studies from Greenwald et al. (1996).