Blog

“But Where Can We Put the Scale?”: Challenges of Data Collection in Rural India

“Data collection.” Sounds straightforward enough, right? During the Fellowship, I have led two different studies on malnutrition in the Sawai Madhopur area: one in three government schools and the other in Gramin Shiksha Kendra’s NGO schools. When people read through large reports, such as the Rapid Survey on Children (RSOC) or the National Family Health Survey (NFHS), and derive information on health, they may not know how much goes into the collection of this data. Especially in rural India, there are circumstances that make it difficult to collect accurate information. I’ll spend this blog post discussing some of the obstacles I faced, not as a project of self-gratification, but to illuminate some of these realities at the field level.

Consent

An important, internationally accepted first step in any research project is to obtain the consent of the people with whom we intend to work. With children, the standard is to obtain consent from parents as well before conducting any research. We took this very seriously. In the first survey, we handed out Hindi consent forms in school for the students to take home to their parents for a signature. However, many of the parents in these communities were not literate. We asked students to read them aloud to their parents so that they would be adequately informed. This was a long, labor-intensive project since children often either forgot the forms at home or did not come to school at all. Because neither the teachers nor the children were familiar with these procedures, it was difficult for them to understand why we were so adamant about them. For me, this brought up questions about what consent really means and how we can obtain it in a meaningful way while working with large sample sizes.

In the next survey, I was able go directly into the community to gain verbal consent from parents and confirm for myself that they really understood the point of the project. I sat with them and spoke about GSK’s plan to begin a health program and the importance of learning about their children’s nutritional habits. I felt much more confident that this was truly informed consent, since participants were able to ask questions and express confusion.

Locating Participants

As mentioned above, the progress of both surveys was slowed significantly by a high rate of school absences. Chronic absence is an enormous problem in both types of schools in India, with some students missing over half of the school days in an academic year. The reasons vary—illness, house or field work, family weddings—but the effect is the same. Many children are simply not in class. Some days we were able to complete 20 surveys, other days only 3. This significantly impacted the efficiency of data collection.

Even finding the children or parents within their homes was a challenge. Because there are no markers to any home (such as address or name) that would allow me to find it on my own, I had to rely on asking others how to locate each family within the village. The most helpful guides actually turned out to be the other school children. They were knowledgeable about the location of each house and generously spent time helping me get there.

Records

In both the government and NGO schools where I worked, paper-based data entry systems are being used. Nothing was digitized. This meant that randomly sampling was very time-intensive. To save time, we elected to assign a number to each child and then randomly draw numbers from a hat, rather than digitizing all student names and using a software to sample with a computer. Teachers volunteered to help with this process, but given the time it took, eventually became impatient and instead began suggesting names of children for the survey.

Furthermore, because these surveys were intended to measure levels of malnutrition, we had to record the age of every participant. In order to determine whether or not the child is stunted (i.e., low height for age), or underweight (i.e., low BMI for age), an exact date of birth is required. In the NGO schools, some of the children’s birth dates were simply not available. Many parents do not actually have a birth certificate for their children, and do not remember the exact birth date. According to the RSOC 2014, only 47.6% of rural children under 5 years had a registered birth, and only 9.6% have a birth certificate. [1] GSK does not push parents to provide birth certificates for entrance into the school, because then they would end up providing forged ones anyway. The records simply do not exist. Surveyors often tackle this problem by leading the parents through a recall of the year and season in which the children were born (in winter, before or after Holi, etc). I used this technique during surveys with mothers, with varying levels of success.

IMG_5744

Tools and Environment

One unforeseen but significant challenge in taking anthropometric measurements was managing the tools in the environment. Using public transportation, which was already quite unreliable and precarious, was made significantly more difficult by carrying the heavy tools required for the surveys, including a stadiometer, scale, and hemoglobin test kit.

It also proved difficult to find floor space in classrooms that was flat enough to use the digital scale. The floors are made from rough-hewn slabs of rock that are often uneven or jagged. This was especially important, since scale errors could affect the accuracy of the BMI calculations and create a false representation of the child’s nutrition status. The children in our schools thoroughly enjoyed watching me scurry around the classroom each day hunting for a perfectly flat spot to place the scale.

IMG_6912

I have gained a newfound appreciation for field workers who make sure their data is accurate no matter the obstacle. They allow us to make important decisions about health. This data offers a portrait of those who are not always seen, forcing us to address inconvenient truths. Their hard work creates the foundation for better policy decisions and more targeted interventions.


Notes: 

[1] Rapid Survey on Children (RSOC) 2013-14. New Delhi: Government of India, Ministry of Women and Child Development, 2014. Page 5. http://wcd.nic.in/acts/rapid-survey-children-rsoc-2013-14.

 

2 Comments
  1. Janan Dave
    Janan Dave

    Annika – I’ve faced a lot of similar challenges in trying to collect data in my work– happy to have you as a friend to trouble-shoot and commiserate with. Though the context has been challenging, I’ve seen you become such a skilled researcher this year.

  2. David

    doing things the right way and refusing to compromise on the quality of your research despite many of the above barriers Is commendable. Growing each day to meet this challenges has been great to see and hear about. Loved this blog post and looking forward to reading more.

Leave a Reply

Your email address will not be published. Required fields are marked *