Measuring efficiency in primary school: the comparison of SFA and DEA

博士 === 國立高雄師範大學 === 教育學系 === 103 === This study compared the performance of school efficiency in Penghu County measured by two approaches of measuring: Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA), thus explored the applicability of adopting these two methodologies for scho...

Full description

Bibliographic Details
Main Authors: HO-KAI CHEN, 陳河開
Other Authors: LI-JU CHEN
Format: Others
Language:zh-TW
Published: 2015
Online Access:http://ndltd.ncl.edu.tw/handle/2pf2vu
Description
Summary:博士 === 國立高雄師範大學 === 教育學系 === 103 === This study compared the performance of school efficiency in Penghu County measured by two approaches of measuring: Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA), thus explored the applicability of adopting these two methodologies for school efficiency measurement. The purposes of this study include: (1) Discussing the content of education costs and the costs of public and private sectors in Penghu; (2) Interpreting the student assessment performance in 2011SY; (3) Adopting DEA to measure school efficiency; (4) Exploiting the feasibility of adopting SFA on school efficiency measurement, and comparing the results with DEA; (5) Enquiring into the characteristics of SFA and DEA, and providing improvement strategies for schools of various types grouped by these two methodologies. In order to achieve the above purposes, the 40 elementary schools in Penghu County were adopted as the subject. The input variables were defined as the costs from both public and private sectors, the former included personnel expenditure, current operating expenditure, and grants, the latter included after-school tutoring expense and learning material expense by the families collected from questionnaires for parents. On the other hand, the output variable was defined as student test scores in 2011 assessment. Statistical measures were conducted to explore the relation between the two efficiency scores, rankings, and the background variables. The conclusions of this study were as follows: (1) Per student cost of public sector in Penghu was NT$250,000, while the costs from private sector was only NT$29,000; The results also reflected the disparities among townships; (2) The student test scores were distributed as bimodal. Also, school students enrolled in remote islands were much lower than those enrolled in other townships; (3) SFA was more suitable for measuring schools of different sizes than DEA; (4) The efficiency scores measured by SFA were highly correlated with output variables, while the DEA scores inclined to schools of bigger size and lower costs; (5) SFA could provide each school with their own target scores, while DEA only provide basic guideline; (6) Based on the ordinal values of SFA and DEA efficiency scores of each school, the four groups of schools were provided with their own strategies of efficiency improvement. Several recommendations were proposed based on the results of this study: (1) School size could cause statistic errors, especially measured by DEA. SFA, on the other hand, can take both input and output into consideration at the same time; (2) Efficiency scores should be adopted as the benchmark for schools in enhancing efficiency; (3) When ranking schools by their efficiency performance, SFA outperformed DEA; (4) When analyzing education cost, per-student and per-class cost should be taken into account, private cost from student families should also considered when allocating expenditures; (5) School authorities should provide administrative assistance to the four efficiency groups based on their specific needs. Keywords: school efficiency, education cost, stochastic frontier analysis (SFA), data envelopment analysis (DEA)