EFFICIENCY MEASUREMENT OF EDUCATION SECTOR IN INDIA THROUGH DEA

Abstract

Data Envelopment Analysis (DEA) is a widely used non-parametric technique for measuring efficiency of organization that use multiple inputs to produce multiple outputs, when prices of inputs or outputs are not known. This dissertation uses DEA to explore the technical efficiency of different DMUs (States, Districts, Schools and Departments of an Institution) for education sector in India. This study gives special emphasis to testing the sensitivity of DEA scores to different input/output variable selection. The DEA scores of DMUs( State, Districts, Schools and Departments of an Institution) vary according to the inputs and outputs included in the model. Sensitivity analysis is used for different analysis through out the work. For applying DEA technique we choose the education sector, as education is the most powerful weapon which can be used to change the world. To measure the efficiency of education sector, four applications are carried out. The analysis is dived at four levels. Firstly performance of 24 states in India is assessed for three years 2005-2007, then performance of 440 districts from all over India is estimated for the cross sectional for the year 2005-06. Thirdly, a case study for the schools in Uttar Pradesh, India is carried out for the year 2000. Lastly performance assessment for academic departments of IITR has been made for the year 2003-2004. For the state level assessments 24 states out of 28 states are selected. The CCR and BCC models are used for the analysis of elementary education in India. TE, PTE and SE are calculated for each state. Two types of variables "General Input-output" and "Quality Input- output" are used for the performance of the states. Inputs are the resources provided by the school management and the Government. Outputs are related to the academic result of terminal classes and enrolments in elementary education. To measure the sensitivity of DEA scores three types of measures actual number, percentage and average of input-output variables are considered. Two methods: Karl Pearson and Spearman Rank Correlation coefficients and stability in terms of efficient and inefficient DMUs (states) are employed to test the robustness of the DEA results. It is observed that the states Bihar, Karnataka, Gujarat, Uttar Pradesh and West Bengal are the most efficient states working at MPSS. Karnataka is the "global leader" for all the states. Sikkim is the most scale inefficient state and it operates at. IRS. So to improve its efficiency it has to increase number of teachers and class rooms. For "Quality inputs-outputs based assessment", the state Haryana, Himachal Pradesh, Jharkahand, Orissa, Punjab, Tripura. and Uttarakhand are not utilizing their resources to produce quality results effectively. We check sensitivity and robustness of DEA scores by the use of actual measure, average measure and percentage measure of input-output variables. If actual measures, average measures, and total measures are used changes in all efficiency scores are measured. There is not measured any effect on inefficient states of changing the measures. For district level assessment 440 districts, out of 604 districts in India are selected for assess the technical efficiency analysis. Two types of inputs School facility index and teacher quality index are formed for the analysis. We propose the use of PCA in formation of index of variables. Four output variables related to academic results are used for the analysis. The analysis is carried out for the cross sectional data for the year 2005-06. CCRO and BCCO models are employed for this assessment. Two types of analysis (General Result based assessment" and Quality result based assessments") are ii performed with different variable specifications. Under BCC model, 334 districts relatively perform better than remaining 106 districts as they get above average efficiency scores. Under CCR model of "Quality result based assessments", only 6 are efficient and under BCC model 13 districts are efficient. In the school level study we have studied the efficiency differences among 348 schools in Uttar Pradesh state of India. We have assessed the schools in multidimensional setting i.e. with eight inputs and three outputs. Inputs include school resources (teaching, physical and ancillary facilities, teachers' qualities) and home environment of schools' students (parents' education and occupation) while output comprise school wise average marks in environment studies, mathematics, language. Mixing these input-output, four models are used to analyze the performance of schools. All four models are based on the CRS assumption of CRS. One another purpose of the study was to analyze how efficiency rankings of schools depend on inputs and outputs included in DEA. Two methods: quartile formation and calculation of correlation coefficients are used. Average efficiency in the most extensive model (model 4) is 70.58, ranging from 19.44 to 100 percent. According to this model 67 schools (19.25%) are efficient (efficiency score=100%). Both the testing methods shows rankings of schools are stable. This study evaluates the performance of academic departments of IIT Roorkee, India through DEA models using different combinations of input and output variables. The principle objective is to analyze activity-wise performance assessment of the departments. This means that we want to evaluate which department is good for which specific activity (like teaching, placements and research). For this purpose, we did four assessments, namely, overall performance assessment, research performance 111 assessment, teaching performance assessment for all departments and enrolment-placement assessment for engineering departments by using 10 models. We have used sensitivity analysis for these assessments and to check the robustness to these models by changing inputs-outputs. Overall performance assessment is good for all the science departments. Only Electronics and Computer engineering and Management studies departments are doing well for placement activity. Only ECE and Chemical Engineering departments are the best for both UG and PG programs among all engineering departments. Other engineering departments are advised to change their inputs and outputs to become efficient iv