Résumé
Demographers often use maps and regional tabulations to evidence the spatial dimension of population characteristics. While they display considerable ability to test the strength of statistical links between phenomena, they rarely attempt to use similar tools to investigate the nature and extent of spatial correlation between demographic variables. This paper examines this question by combining theoretical and practical issues in the measurement of spatial autocorrelation, using recent Indian demographic variables. After a brief summary of the basic tools commonly used to examine geographical differentials, the paper includes a review of recent spatial exploratory and geostatistical tools. The second part of the paper is devoted to the application of these tools to a set of district-level variables derived from the Indian 2001 census, representing the major population characteristics as well other non-demographic variables. The conclusion sums some of the benefits and difficulties associated with spatial analysis in demography.