Response to comment on estimating school efficiency

After correcting errors in our paper (Econ. Educ. Rev. 20 (2001) 417), Ruggiero (Econ. Educ. Rev. 22 (2003) forthcoming) finds that efficiency estimates provided by DEA and COLS have higher correlations with true efficiency values than indicated in our analysis. However, because Ruggiero only considers cases without measurement error, his analysis leaves the primary question of our study unanswered. Using the corrected data generation process proposed by Ruggiero, we find that the presence of measurement error substantially reduces the correlations between estimates of efficiency provided by DEA and COLS and true efficiency. Thus, the primary conclusions of our original study remain. If the administrative data sets used in school accountability programs have significant amounts of measurement error, and if the methods used to estimate efficiency require exogenous inputs, than measures of school efficiency can be quite misleading.