Developing component-wise inter-censal population numbers in two Indian districts by linear projection, multi-stage sampling and utilization of auxiliary information in generalized regression technique

Arijit Chaudhuri, Indian Statistical Institute
Shankar Dihidar, Indian Statistical Institute

Determination of demographic particulars in inter-censal periods is a vital activity in the advanced countries. In this exercise our mission is two-fold. We undertake a suitably stratied multi-stage sample survey utilizing easily available auxiliary information to estimate certain parameters relating to the people in two Indian districts, at a relatively current time which is December, 2003, posterior to the last National population census of India in 2001. Secondly, we intend to examine how some symptomatic data may be effectively utilized employing the generalized regression method to derive more accurate estimates, possibly closer to the projected census data for the current time-period. In adjusting inter-censal population figures at lower levels of aggregation namely, the district level totals, utilization of projection and sample survey methods together and utilization of symptomatic data by dint of statistical modelling should be a common practice in our country too as in USA and Canada. The methods tried turn out promising on applying certain criteria for assessment.