The determinants of school efficiency in Oklahoma: Results from stochastic production frontier and data envelopment analysis

Scope and method of study. The primary purpose of this study is to determine factors that affect the efficiency of Oklahoma school districts. District level data are obtained from the Oklahoma Office of Accountability for the academic years 1996 through 1999. Using test scores as the output measure, the production of school districts is investigated using a frontier production function. In addition, inefficiency effects are examined. Items considered in the inefficiency model include variables over which school districts have control such as teacher years of experience, teacher salary, teachers holding an advanced degree, size of the school district measured by enrollment, and student/teacher ratio. Student characteristics that are not under the control of school districts such as percentage of minority students, poverty rate, etc. (socioeconomic variables) are also considered in the inefficiency model.;Two different methods of estimation for the production function are employed, Stochastic Frontier Regression (SFR) and Data Envelopment Analysis (DEA). The school districts are ranked according to their efficiency scores generated by SFR and DEA. SFR estimates the inefficiency model simultaneously with the production function. In the DEA model, a second stage estimation is conducted using the Tobit regression model to determine causes of inefficiency.;Findings and conclusions. The method of estimation affects the efficiency scores. Thus, the SFR and DEA efficiency scores for the majority of school districts are not identical. In general, SFR generated a more favorable score than that of DEA. The results from the two estimation methods in the inefficiency model also are different. In this study, DEA results are preferred because DEA is better suited for multiple output district level data like those available for Oklahoma.