Purpose: District effectiveness research (DER) is an emerging field concerned with identifying the organizational structures, administration, and leadership practices at the school district level that help districts find success with all of their students across the schools within the system. This work has mirrored much of the early school effectiveness research (SER). However, to date across the DER literature, site selection for in-depth studies of districts deemed "effective" has been haphazard and nonsystematic. This is problematic given the long history of critiques centered on site selection in SER. The purpose of this study is to address and adapt the critiques from SER to a method of site selection for DER and test the method using a large multiyear data set to identify districts that are significantly unusual and effective. Research Design: A two-level hierarchical linear growth model that nests multiple time points per district (Level 1) within districts (Level 2) was used to predict gains in district achievement for all school districts in the state of Ohio during a 7-year period, 2001-2002 through 2007-2008. Findings: Districts that statistically significantly outperformed their predicted gains in achievement, controlling for background and demographic variables over the period, are identified as possible sites for in-depth qualitative studies for DER in comparison to districts performing at the norm. Conclusions: This study proposes and tests a method for district identification in DER that addresses the critiques from SER through controlling for achievement covariates, modeling district gains over time, and examining the population of districts within an entire state. (Contains 1 note, 2 tables, and 1 figure.)