Publication - Working Paper
One of the important debates in education policy has been how to improve the educational achievement within schools. It has been increasingly recognized that one of the most important factors determining student achievement is teacher quality. The importance of highly qualified teachers is increasingly recognized in public policy. For example, in the US, there exist government regulations at many levels including standards for highly qualified teachers as mandated by the Federal No Child Left Behind Act and state level licensing requirements. In the Indian context, Minister of State for Human Resource Development, Dr. Shashi Tharoor, stated (on 20th August, 2013) that to improve the quality of school teachers, the Government of India has adopted a three-pronged strategy, which includes (i) the strengthening of Teacher Education Institutions, (ii) the revision of curriculum for teacher education in accordance with the National Curriculum Framework for Teacher Education 2009 and (iii) the laying down of minimum qualifications for Teacher Educators and their continuous professional development.
In this context, it is natural to ask how one defines a good teacher. In recent times, increasing attention has been focused on outcome-based perspective, now commonly called value-added analysis, that takes the perspective that a good teacher is simply one who consistently gets higher achievement from students (after controlling for other determinants of student achievement such as family influences or prior teachers). There exist significant literature on teacher effectiveness in the US, but outside the US, very little research has been carried out on the measurement of teacher effectiveness. In this paper, we plan to delve into the teacher quality question in the Indian context using an administrative panel data that matches students to their teachers. The data come from a group of linked private schools located in one of the districts of India’s most populous state Uttar Pradesh. In our dataset, we observe students taking exams in multiple subjects for their twelfth grade at a point of time, and we also know their prior achievement (in tenth-grade) in those subjects. Students are matched to subject-specific teachers who taught them for two years. Our dataset consist scores on multiple subjects (4 or 5) for all the students who took twelfth-grade exam between 2006 and 2010 from ten schools (more than 8000 students). Hence, we have more than 39,000 observations (each observation is a pupil-subject-teacher match). The data also contain human capital and demographic characteristics of the teachers. Exploiting the fact that we observe student taking exams on multiple subjects and each subject is taught by different teachers, we use pupil fixed effects that enables us to control for the fact that teachers are not randomly assigned to students.
The paper intends to find out the extent of variation in teacher effectiveness in India. In addition, the paper will also explore what are the observed characteristics that explain the variation in teacher effectiveness. Although the results are based on a sample from one district in India, and may not be applicable to the entire India given the heterogeneity of India on geographical lines, the paper aims to shed some light on teacher effectiveness and intends to provide some policy perspective for hiring and retaining of teachers. In addition, the paper also hopes to promote the construction and use of administrative databases in policy decisions.