Let the Map Speak: Use of Mapping Technique for Policy Analysis

Abstract

Mapping is a powerful tool for policy analysis. Mapping unveils the hidden trends of the attributes that are not readily apparent in traditional statistical analysis. However, the treatment of spatial effect and the visualization of spatial dependence are rather cursory and often limited to visual inspection of the data. The present paper attempts to go beyond this state of affairs by identifying statistically significant clusters of nuances at different levels, their strength and movement over space with time. Girl child deficit in India is one such issue that attracts national as well as international attention. Existing literature has already point out the high spatial dependence of this issue. But the spatial presentation of such matter often limited to arbitrary classifications of the data and the maps are not always policy sensitive. Using scan statistics and exploiting Indian district level database on child sex ratios, the study tries to detect the geographical clusters of female deficit and its statistical inference on one hand and also tries to quantify its strength over India’s landscape, on the other. In doing so, the study proposes a newer extension of the existing toolbox of spatial autocorrelation and offers a new measure of semi-local autocorrelation. The paper ends up with multi layered maps strength of which does not depend on the visual capacity of the researchers rather rests on its own statistical strength. It is hoped that the methodological advancement proposed in this paper may be found useful for policy analysis. The main objective of this paper is, therefore, to illustrate a newer methodology for making the mapping analysis policy sensitive.