Skip to content

Q+A: How (and why) policymakers are turning to large-scale, representative surveys

©Mansi Midha/Getty Images/Images of Empowerment

An interview with DataDelta India Lead Diksha Radhakrishnan.

Q. We see that policymakers often lack up-to-date, reliable data and insights to inform urgent priorities. How is DataDelta working to improve that in India?

Diksha Radhakrishnan: At DataDelta, our primary objective is to empower government and social sector leaders with data that is not only reliable and representative but also deeply rooted in the experiences of the people they aim to serve.

I can share one such instance of support that the DataDelta India team is providing. , When Nationally representative surveys such as those conducted by the National Sample Survey Office (NSSO) and the National Family Health Survey (NFHS) in India offer valuable data to policymakers but they quickly become out of date and don’t have decision-specific data provided. IDinsight is collaborating with a southern state government to help them better measure the effectiveness of their work across health and education departments. In addition to supporting them to identify key performance indicators, we are also collecting granular, representative data – for example from diverse sub-groups such as pregnant women, lactating mothers, mothers of children 7-72m, parents of children aged 6-10 years and community leaders.

These insights have informed decision-making across departments. For instance our data helped identify discrepancies between the beneficiary registration lists of pregnant women, maintained by the two different departments  of Health and  Women & Child Development. This finding  resulted in the departments unifying the registration process to ensure all eligible beneficiaries receive access to benefits from schemes run by both departments.

Discussions are already underway for a second round of the panel survey,  bolstering our confidence in the theory of change: when high-quality granular data is supplied to government partners, there is an increased utilization of the data for decision-making.

Q. You are working on a multi-state household panel survey in India. What was the motivation for building this and how will it be used? 

Diksha: The most recent population census occurred in 2011-12. More than 12 years ago. Using this data can potentially result in basing policy decisions  on outdated or inaccurate  estimates. Additionally, some of the major  Indian surveys suffer from  delayed timeliness, quality issues, and representativeness. Without access to timely, reliable, and decision-relevant evidence, policymakers may struggle to grasp the ground truth and have to rely onanecdotal evidence to make critical decisions.  To address this challenge, DataDelta is  working to establish a state-representative, multi-state household panel. This will provide policymakers and social sector leaders with a robust mechanism to generate data and insights in a rigorous, cost-effective, and timely manner. 

We currently aim to focus the panel on the energy sector. India’s journey of becoming Viksit Bharat by 2047, goes hand in hand with the goal of achieving green transition, and the commitment, made at the G20 summit, to achieve net-zero emissions by 2070. As we push on the pedal of economic development, we anticipate a surge in energy demands by 2030.

Unfortunately, no regular surveys in India capture detailed household energy consumption and demand dynamics. The panel will fill in these gaps with data on households’ energy usage patterns, awareness levels, and readiness for a shift towards clean fuels and decentralized renewable energy sources. It will also offer insights into the impact of government interventions like thermal comfort regulations and building codes on household decisions.

Through this panel survey, IDinsight aims to bridge critical data gaps, thereby fostering a just and equitable transition to a sustainable energy future for India.

Q. Last month, we featured the stories of three female enumerators in the Philippines. You and many others at IDinsight think a lot about enumerators  — from how we hire to how we support them to do their jobs well. Why is this important?

Diksha: DataDelta is driven by three core pillars – innovative sampling approaches, custom-built survey software for high-efficiency management and hyperlocal field teams that come from the communities we survey. The third pillar is critical to the existence of DataDelta. Without robust field teams, our ability to execute complex, large-scale surveys would be severely compromised. These teams not only adeptly collect data, but are also invaluable guides in contextualizing surveys and fostering equitable communication with respondents.

We are committed to providing a safe environment to the enumerators – for this, we have initiated an ethical hiring process. Enumerators receive compensation in accordance with state regulatory standards, accompanied by benefits such as ESIC and EPF contributions. 

We are also committed to ensuring gender equity within our field teams. To drive this we have taken various initiatives to identify barriers & bottlenecks that prevent female enumerators from applying and securing enumerator roles. You can read more about this work in the blog here.

Q. What work at DataDelta are you most excited about this year?

Diksha: Over the last 5 years at DataDelta, having delivered various customized large-scale surveys to a range of clients, we now feel prepared to transition to a more product-based approach to our work. This would mean developing productized services that allow us to streamline our operations to enhance efficiency and offer tools & platforms that can take a more plug-and-play approach for partner needs. Some of the products we are introducing this year include: 

A Data Quality (DQ) Diagnostic Tool, developed in collaboration with the Data Science & Engineering team (DSEM). This tool is designed to automate diverse data quality checks tailored to the specific requirements articulated by our partners. Such a tool could be especially useful for social sector partners who invest heavily in M&E systems but are unsure of the accuracy of their outputs and partners building AI/ML tools and want to ensure that any quality issues inherent in the data do not get reflected in the model. The tool would provide partners with a diagnostic report on data challenges, their implications on data use, and actionable context-based recommendations on ways to improve their data systems. Stay tuned for our upcoming blog series highlighting our work and learnings in DQ.

Another tool is our Gender Data Tool, designed to bridge gender data gaps within administrative and programmatic datasets. This tool is specifically crafted to assess data systems through a gender lens and assist partners in transitioning their databases from a gender-blind approach to a gender-intentional perspective.

Finally, we’re excited to introduce our emerging Humanizing AI service, which focuses on the collaborative efforts between DataDelta and DSEM teams to support the entire lifecycle of building and deploying machine learning models. This service delves into various aspects such as problem identification, model development, testing, refinement, and impact evaluation. By combining our expertise, we aim to provide a comprehensive approach to ML model building, maximizing its value and impact.

Building, socializing, and deploying these tools & the panel platform would be the big focus of 2024 and something I am extremely excited to take on.