Primary sources of data

Detail

What primary data are

• Primary data are data collected first-hand by a person, a group or an institution for a set purpose. Nobody has collected or published them before.
• Secondary data are the opposite. They are data already collected and published by someone else, such as government reports, gazetteers and yearbooks.
• NCERT Fig. 1.1 splits the sources of data into two branches:
• Primary sources: personal observation, interview, questionnaire or schedule, and other methods.

• Secondary sources: published and unpublished documents.

• Why primary data matter in geography: they record conditions on the ground as they are now, in the place being studied. A village-level land-use map or a ward-level literacy study often needs fresh field data that no published table holds.

Source 1 — Personal observation

• Personal observation means collecting information by watching and recording things directly in the field. It is usually done through a field survey (a visit to the study area to see, measure and record).
• What a field survey records:
• Physical features: relief (the height and slope of the land), drainage (the pattern of rivers and streams), soils and natural vegetation.

• Human features: population structure, sex ratio (number of females per 1,000 males), literacy, means of transport and communication, and urban and rural settlements.

• What the observer needs:

• theoretical knowledge of the subject, to know what to look for;

• a scientific, unbiased attitude, so that personal opinion does not change what gets recorded.

• Weakness: the result depends on the skill and honesty of one observer. Two observers with different training may record the same field differently.

Source 2 — Interview

• An interview collects information through direct dialogue (face-to-face conversation) between the researcher and the respondent (the person who answers).
• The eight NCERT precautions:

Stage	Precaution
Before	1. Prepare a precise list of items to ask about.
Before	2. Be clear about the purpose of the survey.
During	3. Assure confidentiality (promise that the answers will be kept secret), so the respondent trusts you.
During	4. Create a congenial (friendly) atmosphere.
During	5. Use simple, polite language.
During	6. Never ask questions that hurt self-respect or religious feelings.
End	7. Ask for any additional information the respondent wants to give.
End	8. Thank the respondent.

• Why these precautions matter:
• a respondent who is afraid or offended gives false or partial answers;
• false answers lower the reliability (trustworthiness) of the whole data set.

Source 3 — Questionnaire and schedule (classic trap)

• Both are printed lists of questions used to collect primary data. The difference is who fills them in.
• Questionnaire: a list of questions that the respondent reads and answers alone.
• Schedule: a list of questions that a trained enumerator (a field worker who asks each question and writes down the answer) fills in on the respondent's behalf.

Feature	Questionnaire	Schedule
Filled in by	the respondent	a trained enumerator
Who can answer	only literate people	literate and illiterate people
Cost	cheap	costly (enumerators must be trained and paid)
Reach	can be posted (mailed) to far-flung places; suits large areas	wide coverage of a whole population, if enough enumerators are sent
Questions	must be simple and self-explanatory	the enumerator can explain hard questions and probe for detail
Typical example	mailed surveys, online feedback forms	Census of India, NSS household surveys

• Trap: the Census does not use a questionnaire in the strict sense. It uses schedules, because a large share of India's people could not read or write when the Census began, and many still cannot today.
• Census 2027 changes how the schedule is filled. Enumerators will enter it on a mobile app, not on paper. This makes it India's first digital census [2].
• Phase I: Houselisting and Housing Census (HLO), held for 30 days in each state or UT between April and September 2026 [2].
• Phase II: Population Enumeration (PE), in February 2027. It covers demographic, socio-economic, cultural, migration and fertility details of each person [2].
• The HLO mobile app works offline on Android and iOS in 16 regional languages. It sends data straight from the field to the server, so no paper is needed [2].
• A Self-Enumeration (SE) portal lets a household fill in its own details online before the enumerator comes. The household gets an SE ID and gives it to the enumerator, who then verifies the data [2]. This is a questionnaire-style step built into a schedule-based census.
• The CMMS (Census Management & Monitoring System) portal lets officers at the sub-district, district, state and national levels track progress in near real time [2].
• About 30 lakh field functionaries (enumerators, supervisors, master trainers and census officers) will be deployed [3].
• Reference date (the fixed moment the count describes): 00:00 hours, 1 March 2027. For Ladakh, the snow-bound areas of Jammu & Kashmir, and Himachal Pradesh and Uttarakhand, it is 00:00 hours, 1 October 2026 [3].

• Caste enumeration will be included in Census 2027. The Cabinet Committee on Political Affairs decided this on 30 April 2025 [4].

• History: the first modern census of India was carried out from 1865 to 1872, but not in all regions at the same time. The first synchronous census (held in all regions at the same time) came in 1881 [3].

Source 4 — Other methods (instruments in the field)

• NCERT Fig. 1.2 shows instruments that measure things directly, without asking anyone.
• Soil-testing kit: measures soil properties such as pH (how acidic or alkaline the soil is) and nutrients in the field.
• Water-quality kit: tests the quality of water samples on the spot.
• Transducers: sensors that turn a physical reading into a signal. NCERT gives the example of transducers that measure crop health.
• Advantage: instrument readings are objective (they do not depend on the respondent's memory or honesty), and they can be repeated and checked.

Sampling — making large primary surveys possible

• The problem: asking every person in a population of over a billion people is very costly and slow. Only the Census, held about once every ten years, does this.
• The solution: sampling. A sample is a small part of the population, chosen so that it represents the whole. Results from the sample are then estimated for the whole population.
• Main types:

Type	How units are chosen	Use
Random sampling	every unit has an equal chance of being picked	when the population is fairly uniform
Stratified sampling	the population is first split into groups called strata (e.g. rural and urban, states), then units are picked from each group	when the population is varied; makes sure every group is covered
Multistage sampling	units are picked in steps (first villages or urban blocks, then households within them)	very large surveys spread over a large area

• NSS (National Sample Survey):
• It was set up in 1950 by the Government of India to carry out nation-wide sample surveys on socio-economic subjects [5].
• The first round ran from October 1950 to March 1951 and covered rural areas only. It was a multi-subject enquiry covering household particulars, agriculture and livestock, household industry, trade, services and consumer expenditure [5].
• It used independent interpenetrating sub-samples (IPNS): each stratum's sample is drawn as two or more independent parts, and comparing the parts checks the quality of the data. P. C. Mahalanobis developed this method at the Indian Statistical Institute (ISI) [5].

• Mahalanobis is called the "Father of the Indian Statistical System". Statistics Day marks his birth anniversary. The theme of the 19th Statistics Day (2025) was "75 Years of National Sample Survey" [6].

• NSS today:

• The Periodic Labour Force Survey (PLFS) gives labour-force (job and unemployment) estimates every quarter for urban areas and every year for rural and urban areas, at the state/UT and all-India levels [7].
• The PLFS collects data with CAPI (Computer Assisted Personal Interviewing). Here the enumerator's schedule is a tablet app. NSSO developed it with technical help from the World Bank [7].
• The revamped PLFS (from 2025) uses a multistage stratified sample design [8].
• In its two-stage stratified surveys, the First Stage Units (FSUs) are villages in rural areas and Urban Frame Survey (UFS) blocks in urban areas. Households are then picked inside each FSU [8].
• Unit-level data (data for each household or establishment) can now be obtained from the NSS by outside users, who can re-tabulate and analyse them themselves [5]. When published and reused by others, such data become secondary data for that later user.

Primary vs secondary — the same data can be both

• Census schedules filled by enumerators = primary data for the Census organisation.
• Census tables used by a student = secondary data for the student.
• So "primary" or "secondary" depends on who collected the data and who is using them, not on the data themselves.

Prelims Hooks

• Schedule is filled in by a trained enumerator. Questionnaire is filled in by the respondent. The Census uses schedules.
• A questionnaire works only for literate respondents. A schedule covers both literate and illiterate people.
• The questionnaire is the cheaper method and can be mailed to far-flung places.
• Census of India: first modern census 1865–1872 (not all regions at once). First synchronous census: 1881 [3].
• Census 2027 is India's first digital census: Phase I HLO from April to September 2026, Phase II Population Enumeration in February 2027 [2]. Reference date 1 March 2027, but 1 October 2026 for Ladakh, the snow-bound areas of J&K, Himachal Pradesh and Uttarakhand [3].
• Caste enumeration was added to Census 2027 by the CCPA on 30 April 2025 [4].
• NSS set up in 1950. Its first round (1950–51) covered rural areas only [5].
• P. C. Mahalanobis is the "Father of the Indian Statistical System". His birth anniversary is observed as Statistics Day [6].
• PLFS: quarterly estimates for urban areas, annual estimates for rural and urban areas. Data are collected by CAPI [7].
• Transducers measure crop health. Soil and water kits are "other methods" of primary data collection (NCERT Fig. 1.2).

Mains Points

• Choosing a method depends on literacy, cost and area:
• in a country with uneven literacy, schedule-based enumeration (the Census, NSS) gives the most complete coverage;

• questionnaires are cheap but leave out the non-literate. This bias hits poorer and more remote areas hardest.

• Field observation and instruments make geographic analysis stronger: a field survey of relief, drainage, soils, settlements and sex ratio, along with soil and water kits and crop-health sensors, gives location-specific data for village planning, watershed projects and soil health programmes. Published district-level data often cannot do this.

• Sampling is what lets India produce data often:
• the Census comes only once in about ten years;
• stratified multistage NSS surveys such as the PLFS give quarterly and annual estimates of jobs, consumption and health in between [7][8];

• this supports policy-making based on evidence (GS-III: economy and employment).

• Going digital improves speed and checking, but brings new issues:

• Census 2027's mobile app, self-enumeration portal and CMMS dashboard should shorten the time from field to data and allow real-time monitoring [2];
• the issues are a digital divide in self-enumeration, data privacy, and the need to keep enumerator verification so that the data stay reliable.