Sibling Death Clustering Among the Tribes of Central and Eastern India: An Application of Random Effects Dynamic Probit Model

Abstract

The tribe population in India, according to the 2011 census, was 87 million and it constitutes around 8.2% of the total population. The infant mortality rate was 84 per 1000 live births among tribe population which was much higher than the national figure. The present research was undertaken to investigate the extent of clustering of infant deaths among families of tribes by rural-urban in the central and eastern states of India. The paper also explores whether infant deaths are uniformly distributed among tribe mothers across different states of this region after adjusting the confounding variables. Lastly, the reduction in infant deaths will be worked out by changing the level of scarring factor. Study utilised retrospective birth history information of NFHS-3 data and examined the clustering through dynamic probit model which addresses the endogeneity arises in the model due to presence of previous infant deaths in the model. It has been found that nearly 7% families in urban areas and 9% families in rural areas were experiencing frequent child loss and the extent of clustering in both urban and rural areas to such families was 52%. After adjusting for mother and child specific covariates in dynamic probit model, the infant death was more likely to occur in those families who had experienced prior child loss. The mother specific unobserved heterogeneity was not found to be significant suggests that tribal women constitute a homogenous group across central and eastern region of India and follow the similar socio-cultural practices. Probit based simulation shows that after eliminating the clustering of deaths in tribal families, the probability of infant death would be declined by 11%. The maximum reduction in infant deaths could be seen in Chhattisgarh followed by Madhya Pradesh.