Monthly Archives: April 2012

Agency and Governance in Violence Prone Neighbourhoods: tracing households across Maharashtra

Posted by Jaideep Gupte

I’m writing this from Pusad, in Yavatmal district, where we have begun the meticulous task of tracing 1100 households in 45 violence prone neighbourhoods across nine districts in the western Indian state of Maharashtra. We have three enumeration teams in place. One has begun the tracing exercise from the coast, in Mumbai and Thane. Along with two doctoral students, Alia A and Yashodhan G, I am travelling with the second and third teams which have begun a westward journey from Nagpur district, roughly 850 kms east of Mumbai, by jeep.

As with the first wave of our longitudinal study[1] in 2010, the start of the second wave has been marked with the sombre news of a high casualty attack involving Maoist rebels. This time, 12 Central Reserve Police Force (CRPF) personnel were killed in an IED explosion in the neighbouring district of Gadchiroli. Initial reports suggest that 25kgs of explosive had been buried in the middle of the road, and detonated by a solitary person from less than 70 feet away. The explosion left behind a crater 6 feet deep and 15 feet wide. The CRPF men were travelling in a mini-van type vehicle, with no additional protection, and so stood no chance of survival. Discontent and grievances in these isolated tribal areas continue to be extreme, but recent government efforts pushing for inclusion have also met resistance. The union rural development minister, Jairam Ramesh, was in the area to promote tribal rights over selling bamboo, something which is currently controlled by the Forest Department. Tribals from Mendha-Lekha, a near-by village have recently become the first in India to officially be given the right to sell bamboo harvested from the surrounding forest areas. At the same time however, the villagers continue their closeness with the Maoist rebels – local police had noticed villagers had stopped using the roads in the days leading up to the blast.

The urban and peri-urban areas which our study focuses on are prone to low-intensity violence and crime. We are using GPS aided-GIS (open source Quantum GIS) to locate households and are collecting survey data on Androidtouch-screen tablets. We are using an open source mobile data capture platform called Open Data Kit, which had seed funding from Google and is now being developed by PhD students at the University of Washington.[2] Programming our questionnaires onto the tablets, localising the text into Marathi and Hindi, training the enumeration team, and building in protocols to safeguard respondents, researchers and the data from any harm has taken a huge amount of effort. But now that we’re able to start working on the data literally minutes after our enumerators leave the respondent household, is extremely exciting. For one, data cleaning can be done in real-time, and any mistakes corrected at source. Plus, we’re not carrying around stacks of paper questionnaires in our already over packed jeeps. Most importantly for our panel study however, we can also monitor attrition rates precisely.

An example from yesterday: we used GIS information from our last round to locate the exact house in which our survey was conducted (last time we randomly selected neighbourhoods, then randomly selected households within those neighbourhoods). When our enumerator starts a survey call at a house, she confirms the same family name at the address as last time. But as the data from her tablet is downloaded onto my laptop, I can compare ages and relationships to see that the current occupants are in fact members of the younger brother’s family, while the older brother has moved out with his wife and children. This changes how we interpret this household’s vulnerabilities and decisions. I can now include this household, trace the older brother’s household or replace it with a completely new household. Having real-time feedback such as this vastly improves our decisions in the field and can make tracing much more precise and cost effective. I can also see that the quantities of lentils, ghee, fruit and meat consumed or the numbers of goats or buffalo owned seem abnormal in certain households – I can see who the enumerators for those particular households are, and immediately ask them to confirm. In most cases they remember the exact details from the interview so can confirm if that was a indeed a typo, and any discrepancies can be smoothed out. In case there is still doubt, re-visiting the household is a ready option since we’re still at the site!

In some survey sites, we are seeing vast differences since we last visited them two years ago. In Mumbai, we seem to be racing against the Slum Rehabilitation Authority, who are clearing some of our survey sites for in-situ redevelopment. Slum dwellers are being given compensation to rent elsewhere while the original sites are being developed into formal/legal buildings.[3]  These families are almost impossible to trace. If we are lucky, we find that they have constructed a shack across the road or are staying on illegally to have continued access to their original sources of livelihood or access to local clinics where they have built a familiarity with the doctors. If they have moved elsewhere, even within the city, tracing them implies a huge cost and provides little added value to a neighbourhood based study.

While the enumerators are conducting their calls, it gives me a chance to briefly interact with some of the neighbourhood residents (some remember me from last time, so the conversation can last longer), take notes on the spatial layout, and to make any corrections to our maps. Here in Yavatmal district I noticed brand new high-tension electricity wires installed over one of our survey sites. Casual chit-chat at a local drinks stall reveals excitement that such development has come to the area. But, as one boy explains, the high-tension wires make it harder to splice on illegal connections, so they have to find other means of getting by – interestingly, from the data that’s already on my laptop, I can see on average the site receives electricity for 18 hours a day, which is about the same as two years ago. But there are a few households in our sample who now report less than 6 hours of electricity, while one is not connected to electricity at all. More detailed analysis might reveal patterns of who is cut-out when service provision is formalised, and whether such formalisation can reinforce segregation. In another district, two of our sites were severely affected by arson, in one almost all the shacks were burned, while in the other, the residents managed to control the blaze to small section. Again, we are hoping our panel study can reveal in more detail how this has impacted our sample, and how households cope with such shocks.

We still have a long way to go – just under forty more sites to survey. We are fortunate to have a very dedicated research team to work with, so I am hoping to report the successful completion of the data collection exercise by mid-April!

[1] With Patricia Justino and Jean-Pierre Tranchant. Funded by the European Commission ( and the DFID-ESRC (“Agency and Governance in Contexts of Civil Conflict”

[2] Recently, the Dispensers for Safe Water (DSW) program at Innovations for Poverty Action (IPA) (supported with a grant from the Bill & Melinda Gates Foundation) used ODK to significantly shorten the feedback loop from data collection to course-correction, allowing them to identify challenges with real-time data and address issues at a rapid pace –

[3] I have written elsewhere about the pros and cons of such redevelopment – Gupte, Jaideep. 2010. Security Provision in Slum Re-Settlement Schemes in Mumbai: A Case Study of the Lallubhai Compound Settlement, Mankhurd. Mumbai Reader 09 (1): 263-279.