Is there a rural bias in national surveys?

Subject :Economy

Section: National Income

Concept :

The Government of India has appointed a panel under the chairmanship of Pronab Sen, former Chief Statistician of India, to review the methodology of the National Statistical Organisation (NSO).

About

Need for review: The usage of outdated survey methodology by national surveys such as the National Sample Survey (NSS), National Family Health Survey (NFHS) and Periodic Labour Force Survey (PLFS), have systematically underestimated India’s development.
Significance of survey/data collection: National level data is a key resource for research, policymaking and development planning, so it is of utmost importance to understand and analyse both claims in the light of existing evidence.
Agencies involved: For this purpose, we will be taking a closer look at NFHS data, which is being conducted by the Ministry of Health and Family Welfare for the last 30 years with the International Institute of Population Sciences (IIPS) as the nodal agency.

Data collection survey method

Data collection surveys collect information from a targeted group of people about their opinions, behaviour, or knowledge.
Common types of example surveys are written questionnaires, face-to-face or telephone interviews, focus groups, and electronic (e-mail or website) surveys.

What are claims against the present methodology?

Rural bias: There is evidence of rural population underestimation by NFHS-3. Overestimation of rural population seems to have taken place by NFHS-2 and NFHS-5.
Only NFHS-1 and NFHS-4 estimates seem to be really close to World Bank estimates and projections based on Census data. However, these errors seem random rather than systematic.
Less scope to overcome errors: Generally, there are higher percentages of no-response in urban areas compared to rural areas. However, this also does not seem to have any systematic relation with either rural or urban bias in estimation.
No response or not, there seems to be room for improvement in minimising the errors and the way sample weights are assigned.