Considerable efforts have been made over the past decade to raise the awareness among public health professionals about the importance of incorporating an equity perspective into health-related policies and interventions in low- and middle income countries. These efforts were accompanied by an increased interest in the use of an asset index to measure socio-economic position (SEP). An asset index combines information on ownership of assets, dwelling characteristics, and access to basic services. The different variables included in an asset index are often combined using principal components analysis (PCA). This procedure is typically applied at national level, combining rural and urban settings.

Drawing on data from the Indian National Family Health Survey, this dissertation addresses the following research question: Is it appropriate to use PCA-based asset indices generated at national level, or should PCA-based asset indices be generated separately for specific geographic sub-groups?

Based on the findings, the general recommendation is that PCA should be conducted separately for rural and urban settings and at the smallest possible geographical scale, sample size and representativeness allowing. This recommendation involves a dear trade-off between the validity of the measure of SEP and the sample size. Two alternative solutions are proposed for situations in which it is not feasible to adhere to this recommendation due to sample size considerations.

The findings from this research have important implications for health-related policy and programming. These are illustrated for a number of applications of the asset index, including the measurement of socio-economic inequalities in indicators of health status and health service utilization, the targeting of the poor for subsidized services, and the assessment of whether services reach the most deprived.