Secondary Data Collection Methods: Sources and Precautions

In population and family health research, not every meaningful study begins with a fresh questionnaire or a field visit. A significant portion of research today relies on data that already exists – collected by governments, international agencies, hospitals, and other researchers for entirely different purposes. This is called secondary data, and learning to use it well is one of the most practical skills a researcher can develop. It saves time, opens up large-scale comparisons, and lets students explore questions that would otherwise be impossible on a limited budget. But secondary data also comes with traps: outdated figures, mismatched definitions, hidden biases, and gaps that quietly distort conclusions. Understanding the sources, weighing the benefits against the limitations, and taking sensible precautions are what separate a credible study from a misleading one.

What counts as secondary data? 🔗

Secondary data is information that was collected by someone else, at some other time, for some other purpose. The researcher using it did not design the questionnaire, train the field staff, or supervise the data entry. Instead, they access the dataset, the report, or the published figures and reuse them to answer a new research question. In population and family health studies, this could mean using Census tables to study urbanisation, NFHS reports to examine contraceptive prevalence, or WHO databases to compare maternal mortality across countries.

The distinction matters because the assumptions, definitions, and quality controls of the original study become inherited features of any secondary analysis. A researcher using secondary data is essentially borrowing someone else’s lens – and must therefore know exactly what that lens does and does not show.

Major sources of secondary data 🔗

Secondary data in population and family health studies flows from several streams. Each has its own strengths, periodicity, and limitations, and a good researcher learns to navigate them like a map.

Government reports and official statistics 🔗

The Indian government is one of the largest producers of population and health data in the world. The Census of India, conducted once every ten years, provides comprehensive demographic information on age, sex, literacy, occupation, migration, religion, and housing. The Sample Registration System (SRS), operated by the Office of the Registrar General, provides annual estimates of birth rates, death rates, infant mortality, and fertility through a dual-recording mechanism in sample units spread across the country. The Civil Registration System (CRS) captures births and deaths under the Registration of Births and Deaths Act, 1969.

Beyond these, ministries publish their own statistical handbooks. The Ministry of Health and Family Welfare releases health management information system (HMIS) data, the National Sample Survey Office produces consumption and employment surveys, and state directorates of economics and statistics issue district-level abstracts. For a student in population studies, these are usually the first place to look.

Large-scale demographic and health surveys 🔗

The National Family Health Survey (NFHS), conducted by the International Institute for Population Sciences (IIPS) for the Ministry of Health and Family Welfare, is probably the single most-used secondary dataset in Indian population research. NFHS-5 (2019-21) covered approximately 610,000 households across 707 districts and provides estimates of fertility, family planning, maternal and child health, nutrition, anaemia, domestic violence, and several emerging issues. Earlier rounds – NFHS-1 (1992-93) through NFHS-4 (2015-16) – make trend analysis possible over nearly three decades.

Other large surveys include the District Level Household and Facility Survey (DLHS), the Annual Health Survey, the Longitudinal Ageing Study in India (LASI), and the Comprehensive National Nutrition Survey (CNNS). Internationally comparable data is available through the Demographic and Health Surveys (DHS) Program, which has conducted over 300 surveys in more than 90 countries.

International organisations 🔗

For cross-country comparisons or global benchmarks, international agencies are indispensable. The World Health Organization (WHO) publishes the Global Health Observatory with indicators on mortality, disease burden, immunisation, and health systems. The UNICEF data portal houses statistics on child welfare, maternal health, and nutrition. The United Nations Population Fund (UNFPA) provides data on reproductive health and population dynamics, while the World Bank’s open data covers development indicators including health expenditure and demographic outcomes. The Institute for Health Metrics and Evaluation (IHME) publishes the Global Burden of Disease estimates, which are widely cited in epidemiological research.

Policy documents and programme reports 🔗

Policy documents are often overlooked but extremely useful. NITI Aayog publishes the SDG India Index, the National Health Policy and the National Population Policy documents lay out targets and review past performance, and programme evaluations from agencies like the National Health Mission contain implementation data that rarely appears in survey reports. For students studying programmatic outcomes – say, Janani Suraksha Yojana or Mission Indradhanush – these are essential reading.

Research publications and journals 🔗

Peer-reviewed journals are a rich source of secondary data, both as published numerical findings and as background for designing new studies. Journals like The Lancet, BMJ Global Health, Demography India, Economic and Political Weekly, and Indian Journal of Public Health regularly publish analyses of NFHS, SRS, and Census data. Databases like PubMed, JSTOR, and Google Scholar make these accessible. Theses and dissertations from institutions like IIPS, JNU, and the Tata Institute of Social Sciences also contain detailed secondary analyses worth citing.

Administrative records and other sources 🔗

Hospital records, school registers, electoral rolls, insurance claim data, and even social media activity now form part of the secondary data landscape. The Lancet Digital Health notes that real-world data sources such as epidemic surveillance, disease registries, insurance, and electronic health records are increasingly used for global health research, often combined with newer streams like wearable devices and mobile phone apps.

Advantages of secondary data 🔗

The appeal of secondary data is easy to understand once you see what it offers.

Cost-effectiveness 🔗

The most obvious benefit is cost. A survey of 600,000 households like NFHS-5 costs hundreds of crores of rupees and takes years to complete. For an individual researcher or student, conducting anything of comparable scale is impossible. Using existing data sidesteps these enormous costs entirely. The same logic applies to time – datasets that took years to build are often available within minutes of a download request.

Broader scope and scale 🔗

Secondary sources frequently provide data at scales no individual researcher can match. The Census covers every village and household in the country. NFHS-5 provides district-level estimates for 707 districts. The WHO Global Health Observatory covers nearly 200 countries. This scale enables comparisons across regions, time periods, and population subgroups that would be impossible with primary data.

Trend analysis over long periods 🔗

Because surveys like the SRS and NFHS have been conducted for decades, they make longitudinal analysis possible. A researcher can track how the total fertility rate has declined across Indian states from the 1990s to the 2020s, or how institutional deliveries changed before and after the launch of Janani Suraksha Yojana. As the Open University notes, large organisations with significant resources can sustain longitudinal data collection that individual researchers simply cannot replicate.

Methodological rigour 🔗

Major secondary datasets are usually collected with sampling designs, training, and quality controls that exceed what most individual researchers can manage. NFHS, for example, uses stratified two-stage sampling with rigorous interviewer training and biomarker testing. Using this data effectively means inheriting that rigour.

Limitations of secondary data 🔗

For all its advantages, secondary data is rarely a perfect fit. Several limitations need honest acknowledgement.

Mismatch with research objectives 🔗

Secondary data was collected for someone else’s questions. It may not align with the specific variables, definitions, or population subgroups a new study needs. A researcher interested in unmarried adolescent contraceptive use, for example, will find limited data because most fertility surveys focus on ever-married women.

Concerns about adequacy 🔗

The data may not cover the geographical area, time period, or subgroup the researcher cares about. District-level estimates exist for many NFHS indicators, but block-level or village-level analysis usually requires primary collection. Similarly, rare events or small populations may be under-represented in national samples.

Reliability and quality variations 🔗

Not all secondary data is equally trustworthy. Quality depends on the original collector’s methodology, sampling, response rates, and quality controls. Debates around the data quality of certain NFHS-5 indicators, for example, have surfaced in academic and policy discussions, and researchers should be aware of these critiques before using the data uncritically.

Outdated information 🔗

Census data, in particular, can be significantly out of date by the time it is used. India’s 2021 Census was delayed, leaving researchers relying on 2011 figures for population denominators well into the mid-2020s. Health behaviours, fertility patterns, and migration flows can change considerably over a decade.

Definitional inconsistencies 🔗

What one survey calls “stunting” or “anaemia” or “currently married” may differ subtly from another. The age cut-off for “child,” the threshold for “skilled birth attendant,” or the categorisation of “rural” can vary across sources, making comparisons treacherous.

Precautions in using secondary data 🔗

Because the researcher does not control the data collection process, evaluating secondary data critically before using it is non-negotiable. Three classical tests – adequacy, reliability, and suitability – guide this evaluation.

Suitability for the research question 🔗

The first question is whether the data actually fits the research purpose. As research methodology texts emphasise, suitability is judged by comparing the nature, objectives, and scope of the original inquiry with those of the present study. Definitions, units of measurement, reference periods, and the population covered must all align reasonably with the new research question. If a study collected data on “households” using a definition that differs from yours, conclusions drawn from it may be misleading.

Adequacy of coverage 🔗

Adequacy refers to whether the data sufficiently covers the area, time period, and population of interest. A study on tribal maternal health in Jharkhand cannot rely solely on national averages – the sample must contain enough tribal respondents from the relevant geography to produce reliable estimates. Researchers should check sample sizes for the subgroup of interest, not just the headline total.

Reliability of the source 🔗

Reliability rests on the credibility of the original collecting agency, the methodology used, the response rate achieved, and the quality control procedures in place. Data from established agencies – IIPS, the Office of the Registrar General, WHO, UNICEF – generally carries more weight than figures from advocacy organisations or commercial reports that may have inherent biases. Cross-checking with at least one independent source is good practice whenever the stakes are high.

Currency of the data 🔗

Time-sensitive topics demand recent data. Using 2005-06 NFHS-3 figures to discuss current contraceptive prevalence in 2026 would be misleading. Researchers must report the year of data collection clearly and discuss any likely changes since then.

Transparent documentation of limitations 🔗

Finally, every research report using secondary data should explicitly state its sources, the year and method of collection, the definitions used, and any limitations carried over from the original study. This transparency lets readers judge the validity of conclusions for themselves and is a hallmark of good scientific practice.

Putting it together 🔗

Secondary data is one of the most powerful tools available to population and family health researchers – particularly students and early-career scholars who cannot mount large primary surveys. The richness of Indian sources like the Census, SRS, NFHS, and DLHS, combined with international databases from WHO, UNICEF, and the World Bank, makes it possible to ask sophisticated questions about fertility, mortality, morbidity, and health behaviours without ever leaving the library. But the power of these tools is matched by the responsibility to use them critically. Knowing what data exists, what it can and cannot tell you, and how to test it for adequacy, reliability, and suitability is what turns secondary data from a shortcut into genuine scholarship.

What do you think? If you were studying a sensitive topic like adolescent reproductive health in your home district, which combination of secondary sources would you trust the most, and which gaps would you fill with primary data? How would you decide when the secondary data on hand is “good enough” to draw policy-relevant conclusions?

References

1. https://censusindia.gov.in/census.website/
2. https://censusindia.gov.in/census.website/en/node/180
3. https://www.nfhsiips.in/nfhsuser/nfhs5.php
4. https://dhsprogram.com/
5. https://www.who.int/data
6. https://data.unicef.org/
7. https://www.unfpa.org/data
8. https://data.worldbank.org/
9. https://www.thelancet.com/journals/landig/article/PIIS2589-7500(22)00195-9/fulltext
10. https://www.open.edu/openlearn/money-business/using-data-aid-organisational-change/content-section-6
11. https://www.dataforindia.com/nfhs-explainer/
12. https://www.vedantu.com/commerce/precautions-to-be-taken-before-using-secondary-data