Monday, August 09, 2010

A Newbie Surveyor’s Guide to Rural India


Its been a whirlwind two months in India and this is a post which has been in the pipeline for a while (hence the length). For those unaware of my current research project, I’m working on a study of Microfinance Self-Help Groups in villages near Sanchi, Madhya Pradesh. It is the first time that I have conducted a real survey. Needless to say, it has given me a deep appreciation of the process of data collection and work of the surveyor.

First, designing a questionnaire is not simply about putting questions on paper; Murphy’s Law rules big time. Framing of questions is key – anything remotely ambiguous WILL cause confusion. Leaving any decision making to your interviewer and interviewee WILL come back to haunt you. Less words, the better. More options, even better.

Second, Lost in Translation is not just the name of a sappy movie. It is a huge pitfall whenever you are relying on an interviewer to translate questions and record responses. While I hired a local person to accelerate the rate of data collection, the main challenge was to have him understand the entire survey and be able to code responses in Bundeli, the local Hindi dialect, onto the English questionnaire. Fortunately (and it is good fortune in a country where the dialect changes every 100 kilometers), I understood Bundeli well enough to oversee that process and ensure that erroneous recording of responses was minimized.

While questionnaire design is demanding, the actual interviews of people were the most fun and rewarding part of the survey. Their responses often threw all the planning of a first-time surveyor like me out of whack. Observing features of demography and village life which veteran surveyors (and astute casual observers) of rural India know very well, I would often have to retract, change my interview questions or simply remember to account for those features when interpreting my results. In any case, here they are - A Newbie Surveyor’s Guide to Rural India.

· aG Age: Women simply don’t know it. Tellingly, they don’t care enough to understate it. Women who are 30 think they are 40 and vice-versa. For all the influence of cable TV on rural women, propensity to understate age is not one of them. Caveat # 1: this might be a local phenomenon as renowned Indian demographer, Ashis Bose, claims otherwise. Caveat # 2: Women under 30, who tend to have attended school, do know their age, often reminding me of the 60s anti-establishment dictum, “Don’t trust anyone over 30” when asking the age question.

Education: Hardly any woman over the age of 30 had attended school. Almost every woman under 25 had some schooling. Clearly, some education policy began working in the area in the early 1990s. Two frequently cited reasons for dropping out of school were 1) corporal punishment, and 2) inability to pass the class V or class VIII exams, common dropout points.

Time: The Gregorian calendar has little to no meaning for the rural woman’s calendar. Time is marked by festivals, crop cycles and the Hindu calendar (Panchang), based on lunar movements. This meant that I finally put my mind to understanding how the months of the Hindu calendar map onto the Gregorian calendar!

Time II: The Pan-Indian problem of “Indian Stretchable Time” (IST) is amplified many times over in villages. Anyone who has lived in India/South Asia or has a brown friend knows what I mean. Trying to quantify distance between the houses of SHG members through time, which I had initially considered a great idea, flopped big time. It does not matter whether you ask if point A is 5 minutes, 10 minutes or 20 minutes away from point B – the answer remains the same! A better question is to simply ask if point A is close or far from point B.

Assets: Using household owned assets (bicycle, fan, almirah etc.) as a measure of poverty can be tricky if you take responses at face value. Assets which come into the household by way of dowry are often not considered ‘owned’ by the household and respondents need to be reminded about them (Dowry is not restricted to Hindu marriages – Muslim marriages are also characterized by it). Moreover, the high maintenance cost of assets means that many are in a state of permanent disrepair, creating another classification dilemma for researchers.

Financial Inclusion: Commercial banks (state-owned and private) have not reached the poor and ultra-poor in Madhya Pradesh. Very rarely in my sample would I come across someone who had even attempted borrowing from a bank, forget actually taking out a loan. Reasons varied but a common refrain was that a bank location was too far, in the city. This suggests that the financial inclusion gains made between 1977 and 1990, when national banks were mandated to open 4 rural branches for every branch opening in an already banked location, have been reversed. Check out the famous Pande-Burgess paper on Social Banking to know more about the positive poverty reduction impact of that policy.

W While there are plenty of non-survey related anecdotes to keep this post going, I’ll stop here and save them for some other time and place. Stay tuned for the startling revelations of my earth-shattering paper which will be written sometime this summer.