We deposit here the results of a household-based survey that was used to understand the situation of Dalits and Adivasis in five rural localities in Kerala, Tamil Nadu, Maharashtra, Telangana and Himachal Pradesh, India. It was part of a wider project based on participation observation - living with the Adivasi and Dalit communities in question for a year -that asked how and why Dalits and Adivasis remained at the bottom of the Indian economic and social hierarchy despite economic growth. This survey sought to document how people made a livelihood (occupation/employment) as well as levels of education by caste/tribe (and within that by gender).In recent decades India has experienced exceptionally high economic growth rates, becoming one of the world's fastest growing major economies. Yet, the redistribution of the fruits of economic growth - the trickle down effects of growth - have been negligible for vast swathes of India's population, most of whom live in the countryside. The demographics of the poor are starkly socially marked. Economists tell us that India's dalit and adivasi communities, who account for almost 25% of the country's population and were historically seen as 'untouchable' and 'savage', suffer from disproportionate levels of poverty, remaining worse off than other groups almost everywhere across the country. But econometric analysis is unable to tell us how and why this is the case. This project consists of the following three components: 1) primary anthropological research to understand the processes by which poverty is reproduced through agrarian relations and the shift from farm-based social and economic hierarchies towards new forms of power and exploitation off the farm which lead to the persistence of dalit and adivasi marginalisation across India. 2) analytical development in the study of poverty and its persistence which, informed by recent statistical research and policy shifts at the national and state levels, crafts a more critical and powerful alternative to poverty measurements by ethnographically exploring the relationship between political and economic transformations in rural-based dalit and adivasi lives, and the transformations taking place at the macro level. Crucially this involves analytically establishing a research field which structures ethnography in the framework of political economic theory and brings this combination to the centre of understandings of poverty. 3) some of the first historically situated ethnographic studies which are comparative, not only in their regional distribution, but also in their underlying theoretical and methodological bases. The three cases - from central and eastern India - will be integrated at the level of planning, midterm goals and findings. Methodologically, the project will thus establish the value of an underdeveloped systematic ethnographic approach to poverty which will foreground the comparison of the consequences of rural political and economic transformation on dalit and adivasi lives across a number of different scales and settings in the most under-researched parts of the subcontinent.

Five field locations were selected in five Indian states: Kerala, Tamil Nadu, Telangana, Maharashtra and Himachal Pradesh in order to cover a set of different conditions of Dalit and Adivasis across India. In Tamil Nadu, Telangana and Himachal Pradesh, the 'location' was defined as a revenue village. In Kerala, the location was a Tea Plantation and in Maharashtra 3 interrelated villages were covered. In each location a postdoctoral research fellow undertook a household survey, covering employment and education of all adult household members of all Dalit castes and all Adivasi groups. Where relevant, other caste groups were also included. Sampling varied across sites. Where possible the entire village/site was surveyed - as in the case of Himachal Pradesh and the Kerala tea estate. In Tamil Nadu almost all the Adivasi and Dalit castes were surveyed but there was also data included on a smaller random sample of other castes just for illustration of difference. In Telangana, a 30% random sample across castes was conducted. In Maharashtra, where the focus was on three villages, all villages were surveyed and included 40% of all houses in the hill village (fully Bhil), 15% of all houses in the resettled village (fully Bhil) and 10% of all households in the plain villages (and which included Bhils and Gujars). The researcher carried out participant observation living for a year with the communities. As is common in many anthropological studies of this kind consent was oral and based on the longterm trust developed with informants. Information about survey was also delivered verbally, Detailed information is available in: Alpa Shah, Jens Lerche, Richard Axelby, Dalel Benbabaali, Brendan Donegan, Jayaseelan Raj and Vikramaditya Thakur, 2018: Ground Down by Growth. Tribe, caste, class and inequality in twenty-first-century India. Pluto Press and OUP India.