An integration of stratified sampling designs and Geographic Information Systems

An integration of stratified sampling designs and Geographic Information Systems - An application in Educational Research

Satharasinghe A,Ranjith Premalal De Silva
E-mail: rpdesilva@pdn.ac.lk

Thattil R.O., Samitha S.
Postgraduate Institute of Agriculture, University of Peradeniya, Sri Lanka
Tel: +94-777-801712, Fax: +94-8-387216

Introduction
The standards of education in schools are highly diverse in Sri Lanka. Most of the facilities have been concentrated to the schools located in the major cities and suburbs. A government initiative for a major restructuring programme for education is underway aiming at improving national standards of education and minimizing the discrepancies on the quality of teaching offered at school level. This brings about the need for the evaluation of performance of schools.

In this exercise, various parameters that estimate the performances of schools and competence levels of students need to be defined. In general, these parameters could be enumerated using sample surveys and accordingly, the need for a proper sampling framework to derive the correct inferences is emphasized.

The most commonly used sampling strategy is based on the stratified sampling design. Stratified sampling designs are frequently used in demographic and socio-economic surveys. It may be possible to divide a heterogeneous population into sub populations each of which is internally homogenous. Geographical regions are generally used as strata in most of the stratified sampling surveys. If each stratum is homogenous, a precise estimate of any stratum parameter can be obtained from a small sample in that stratum. These estimates can then be combined into a precise estimate for the whole population. In this context, the value of characteristic being measured is the most suitable variable to be used for the stratification of elements in a population. However, in general, this information is not available. Therefore, other variables that are highly correlated to the characteristics of interest are used.

Further, sub populations identified based on geographical regions of populations may not be homogenous due to two reasons. The first reason is that no spatial information other than rigid regional boundaries is used in the stratification. Secondly, the factors that influence the performance of schools within a selected administrative boundary are very diverse and vary widely.

Further, it is obvious that inferences cannot be made about individual behaviour from sample data and it is often both necessary and relevant to estimate parameters at the aggregate level. In most of the surveys, these aggregate levels are geographic regions including administrative boundaries. Hence, it is required to identify the suitable variables, which influence the performance and then determine the spatial boundaries for the process of stratification.

Many factors such as characteristics of students, parents, teachers and learning environment within the school, household status, community background influence the performance and students and their schools. Some of these variables can be described as spatial variables and some others as non-spatial, attribute variables. Spatial variables are used to stratify the areas in this study while non-spatial, attribute variables are used to stratify schools. Classification of schools by type cannot be justified, as type does not account for the facilities available in the schools and the other characteristics, which influence the performance of schools. Strata determined by the spatial and non-spatial variables are used in the first stage and second stage in the two stage stratified sampling design.

Accordingly, the objective of this study is to develop and propose a suitable sampling strategy for the evaluation of the performance of schools in the Southern province of Sri Lanka.

Page 1 of 3
| Next |