Measuring the Determinants of School Completion in Pakistan: Analysis of Censoring and Selection Bias

Abstract

Ascertaining the sources of biasness in educational studies and highlighting factors that impact attainment of education within Pakistan.

The paper explores the determinants that impact attainment of education within Pakistan while also highlighting the censoring and sample selection bias that prevails in previous studies regarding education. Ordinary Least Square, Ordered Probit and Censored Ordered Probit methodology based on data generated from Pakistan Integrated Household Survey (1991) reveals that the sample selected can introduce bias within the results which might prevent the study from accurately providing an insight into factors that impact education. The non-trivial censoring bias produces distorting results that create negative implications for educational policies.

This paper provides an analysis based on three frameworks and holds that censor ordered probit model based on a sample that includes all home and non-home resident children is able to produce accurate results. The study concludes that parents education impacts girls and boys schooling whereas mothers schooling significantly impacts daughter’s education and fathers schooling influences son’s education. Income level is also a primary determinant of attainment of education specifically for girls. The authors believe that resources should be diverted towards increasing access to secondary schools as access to primary schools is not the major factor influencing attainment.

Publishing Year: 1999

Attached File:

cdp794.pdf - Download