Doing Sociology: Research Methods

The last chapter of Introducing Sociology asks what makes sociology a social science. Its answer is method. It discusses objectivity and reflexivity, the choice among methods, and then three methods in detail: participant observation (field work), the survey and the interview. It has five activities, a sex-ratio table and ten exercise questions. The Class 12 chapter Suggestions for Project Work applies these methods to small student projects; this page covers the principles behind them and does not repeat that one.

• Builds on Chapter 1 (sociology vs common sense; sociology and social anthropology), Chapter 2 (reflexivity in concepts) and Chapter 3 (the sex ratio). Links to Understanding Society, Chapter 5 (Indian sociologists — Srinivas's village studies).
• Pairs with live chapters: Class 12 Indian Society, Chapter 1 Introducing Indian Society (self-reflexivity), Chapter 2 The Demographic Structure (Census data) and Chapter 7 Suggestions for Project Work (practical survey, interview, observation); Social Change and Development in India, Chapter 4 Change and Development in Rural Society.
• UPSC: Prelims asked what an exit poll is — a post-election survey of voters (1994). The Essay paper set ‘What is research, but a blind date with knowledge’ (2021). Beyond these, the chapter is a toolkit for reading Census, NSS, NFHS and PLFS data critically in GS answers.

• Sociologists differ from lay people by how they know, not how much; sociology needs the outsider's observation and the insider's meanings.
• Objectivity is harder in sociology (researcher bias; many ‘truths’; multi-paradigmatic) → self-reflexivity, documentation, declaring one's background; objectivity is a continuous process.
• Methods: quantitative/qualitative, primary/secondary, micro/macro — conventions only; choose by the question; combine (triangulation).
• Participant observation: about a year living ‘as one of them’; census → map → genealogy → language → informants and daily field notes. Malinowski (Trobriands) made it the anthropologist's method; India's village studies of the 1950s.
• Survey: a representative sample — all relevant strata + random choice within them; report size, design and margin of error; gives the aggregate picture (child sex ratio) but little depth; non-sampling errors.
• Interview: a guided conversation between questionnaire and participant observation — extremely flexible, but unstable; depends on rapport.

More than any other discipline, sociology deals with things already familiar to most people — groups, institutions, norms, relationships. So what makes the sociologist different? As in all sciences, the crucial element is method: sociologists differ from lay persons not by how much or what they know, but by how they acquire knowledge. Sociology is also deeply interested in lived experience: in friendship, religion or bargaining, it wants not only what a bystander can observe but the opinions and feelings of those involved — what friendship means in different cultures, what a believer thinks a ritual does, how shopkeeper and customer read each other's words and gestures. The need to understand both the outsider's and the insider's points of view is another reason method matters.

‘Methodology’ is often used as a synonym for method but strictly means the study of method — general problems of scientific knowledge-gathering beyond any one technique. ‘Objective’ means unbiased, neutral, based on facts alone; ‘subjective’, based on individual values and preferences. Every science aims to be objective, but this is much harder in the social sciences. A geologist or botanist is not part of the world of rocks or plants; social scientists study the social world they themselves live in.

• Bias: sociologists have ordinary likes and dislikes; one studying families is a member of a family; one studying another caste or religious community may carry the attitudes of her own social environment.
• Remedies: rigorously examine one's own ideas and feelings; take an outsider's perspective on one's own work — self-reflexivity, consciously adopting the point of view of others, especially the people studied; document all procedures and cite all sources so that others can retrace the steps and one can re-check one's own argument; and, because unconscious bias remains, state the features of one's own background that might bias the work, so readers can mentally ‘compensate’.
• Many versions of ‘truth’: shopkeeper and customer differ on a ‘good’ price, young and old on ‘good food’; sociology tries not to judge which is true, because it is interested in what people think and why.
• Many paradigms: like its sister disciplines, sociology is ‘multi-paradigmatic’ — competing, incompatible schools coexist (Chapter 2).

So the old notion of an ‘objective, disinterested’ social science is widely seen as outdated, even misleading. This does not make sociological knowledge useless or objectivity a useless concept: objectivity is a goal of a continuous, ongoing process, not an end result already achieved.

With multiple truths and perspectives come multiple methods; there is no single road to sociological truth. Each method has strengths and weaknesses, so arguing over superiority is futile; the right question is whether the method suits the research question. To find out whether most Indian families are still joint, a census or survey is best; to compare women's status in joint and nuclear families, interviews, case studies or participant observation may suit. Conventional classifications: quantitative (countable variables — proportions, averages) vs qualitative (attitudes, emotions); methods studying observable behaviour vs non-observable meanings and values; secondary data (documents, records, artefacts — e.g. historical methods using archives) vs primary data (e.g. interviews); and micro methods (small settings, a single researcher — interview, participant observation) vs macro methods (large numbers of respondents and investigators — chiefly the survey).

These classifications are conventions; the lines are not sharp, and one method can be converted into or supplemented by another. The choice depends on the research question, the researcher's preferences and constraints of time and resources. The recent trend is to bring several methods to bear on the same problem from different vantage points — triangulation. NCERT stresses methods producing primary data, as most distinctive of sociology, and introduces the three most prominent: the survey, the interview and participant observation.

Popular in sociology and especially social anthropology, participant observation differs from surveys or interviews in its long period of interaction: typically the researcher spends many months — about a year or more — living among the people studied as one of them. As an outsider, the anthropologist immerses herself in the ‘native’ culture, learning the language and taking part in everyday life, to acquire the insider's explicit and implicit knowledge. The goal is to learn the ‘whole way of life’; the model is the child, who learns about the world holistically. It is also called ‘field work’, a term borrowed from natural sciences such as botany, zoology and geology, whose scientists had to leave the laboratory for ‘the field’.

Field work made anthropology a science. Early anthropologists were ‘armchair scholars’ who compiled accounts by travellers, missionaries, colonial administrators and soldiers about communities they never visited — James Frazer's The Golden Bough and Durkheim's work on primitive religion rested on such second-hand accounts. From the late 19th century, some scholars began systematic first-hand study; reliance on second-hand accounts came to be seen as unscholarly, and since the 1920s field work has been integral to anthropological training. Many places became famous through classic field studies.

What did the anthropologist do? Usually begin with a census of the community — everyone's sex, age group and family — and a map of the settlement, with houses and socially relevant sites. A key early technique is the genealogy: a family tree for individuals, extended back as far as people remember, built by asking a household head about relatives in his or her own generation, then the parents' and grandparents' generations, and cross-checking with other relatives. This reveals the kinship system and, practically, introduces the researcher to people. Then come constant language learning, observation and detailed notes on festivals, collective events, livelihoods, family relations and child-rearing — asking ‘endless questions’ about things members take for granted, like a child. The anthropologist depends on one or two informants (‘principal’ or, in early usage, ‘native’ informants), who act as teachers; and keeps field notes written up every day without fail, perhaps as a daily diary.

Sociological field work uses much the same techniques; it differs in context (where) and emphasis rather than content. Sociologists also try to become insiders, but among all sorts of communities, not only remote tribes, and do not necessarily ‘live in’, though they spend most of their time with the community. William Foote Whyte spent three and a half years ‘hanging out’ with a street ‘gang’ of mostly poor, unemployed young men — the first American-born generation of Italian immigrants — in a city slum, and wrote Street Corner Society. Michael Burawoy worked for several months as a machinist in a Chicago factory and wrote about work from the workers' perspective.

Village studies. In India, field work was used above all in village studies from the 1950s, by Indian and foreign scholars. The village served as the equivalent of the tribe: a ‘bounded community’ small enough for one researcher to know almost everyone. Anthropology was unpopular with nationalists for its preoccupation with the ‘primitive’ and its perceived colonial bias — stressing the non-modern side of colonised societies; and classic field work (Malinowski, Evans-Pritchard) was made possible by colonial empires. Studying villages seemed more acceptable. Above all, villages mattered to newly independent India: the government wanted rural development; the national movement, especially Gandhi, had run ‘village uplift’ programmes; urban educated Indians retained links to villages; and most Indians lived (and still live) in villages.

Strengths: a very rich, detailed picture of life from the insider's perspective — the main return on the heavy investment of time; knowledge of the field over a long period rather than a quick visit; correction of initial impressions, which may be mistaken or biased; the ability to track change and see how contexts matter (a good harvest year vs a bad one; employed vs unemployed). A full-time participant observer can avoid many errors of surveys, questionnaires and short-term observation.

Weaknesses: long, intensive work by a single scholar can cover only a very small part of the world — a village or small community — so we can never be sure whether what was observed is common or exceptional: probably the biggest disadvantage. We are never sure whether we hear the voice of the researcher or of the people: the researcher, consciously or not, selects what goes into notes and how it is presented, and no other version is available (though this risk exists in most methods). More generally, field work rests on a one-sided relationship: the researcher asks, answers and speaks for ‘the people’. Some scholars propose dialogic formats — translating the work into the community's language, seeking their opinion and recording it. As the gap between researcher and researched narrows, the scholar's version will more often be questioned and corrected — making research more controversial and difficult, but in the long run more democratic.

The survey is probably sociology's best-known method, now part of everyday public life — election predictions, marketing, opinion polls. It provides an overview based on information from a carefully chosen, representative set of people (respondents). It is usually done by large teams: researchers who design it and investigators or research assistants who collect data — orally in personal visits or by telephone, through written questionnaires delivered or posted, or electronically by email or web forms. Its main advantage: results can be generalised to a large population by studying only a small portion, with manageable time, effort and money.

The ‘shortcut’ rests on sampling theory and the representativeness of the sample, which depends on two principles. First, all relevant sub-groups must be represented — stratification in the statistical sense (rural and urban; villages of different sizes; class, caste, gender, age, religion), with relevance set by the study's objectives (all religions for a study of religious attitudes; workers, managers and industrialists for one on trade unions). Second, the actual unit — person, household or village — must be chosen purely by chance: randomisation, based on probability, like a coin landing heads or tails with equal 50 per cent chance. Choosing villages near the highway, or middle-class or familiar households, makes the sample purposive, not random. Lots, dice, random-number tables or computer-generated numbers ensure chance.

NCERT's example: to test whether smaller, more intimate communities produce greater inter-community harmony than larger ones in one state's rural sector, list all villages with their populations (from the Census), define ‘small’ and ‘large’, drop the ‘medium’ ones, and — giving equal weight to each stratum — draw 10 small and 10 large villages by lot. Real designs have many stages and strata, but the principle is the same: a small sample carefully chosen to stand for the population, whose results are generalised. The sample's statistical properties make it closely resemble the population; small deviations remain, and their likely size can be specified — the margin of error or sampling error, arising not from mistakes but from using a sample. Researchers must report the size and design of their sample and the margin of error. Sample sizes range from 30–40 to many thousands; bigger is more likely to be representative (the extreme being a census), but the mode of selection matters even more.

The aggregate picture. In situations where a census is not feasible, the survey is the only way to study the population as a whole, and its unique advantage is the aggregated picture: many problems become visible only at the collective level. NCERT's example is the child sex ratio, which cannot be seen by looking at individual families — one family's proportion of boys and girls could be offset by others — but only through a census or large survey.

Weaknesses. Wide coverage comes at the cost of depth: time per respondent is limited. With many investigators, complicated questions or those needing prompting may not be asked identically, and differences in asking or recording introduce errors — so the questionnaire (‘survey instrument’) must be designed very carefully, with little chance of correction once in use. Without a long-term relationship, questions must be ones strangers can ask and answer; personal or sensitive questions may be answered ‘safely’ rather than truthfully. Such faults are non-sampling errors — errors from design or implementation, not from sampling — and some are hard to foresee, so surveys can produce misleading estimates. Ultimately a survey depends on a tightly structured, inflexible questionnaire and on the goodwill and cooperation of respondents.

An interview is a guided conversation between researcher and respondent. Its simplicity is deceptive: it takes practice and skill. It occupies the ground between the structured questionnaire of the survey and the completely open-ended interaction of participant observation. Its chief advantage is the extreme flexibility of the format: questions can be rephrased, their order changed, fruitful topics extended and unwelcome ones cut short or postponed — all during the interview. The same flexibility is its weakness: it is vulnerable to the respondent's changes of mood or the interviewer's lapses of concentration — an unstable, unpredictable format that ‘works very well when it works, and fails miserably when it doesn't’.

Styles differ: some prefer a loose checklist of topics, others specific questions for all respondents. Recording may be by audio or video, detailed notes during the interview, or memory and later writing-up; recorders may make respondents uneasy and formal, while lighter methods may miss information; circumstances often decide. Publication styles also differ: a ‘cleaned up’ continuous narrative or a transcript keeping the asides and digressions. The interview is often used with other methods — long conversations with key informants in participant observation, or intensive interviews adding depth to survey findings. It depends on personalised access and on rapport — mutual trust between respondent and researcher.

The chapter has five activities and several questions in its boxes. Short outlines follow.

The ten questions at the end of the chapter (pages 101–102). Try each first, then open the outline.

Numbers

One-liners

• Sociologists differ from lay people by how they know, not how much.
• Sociology needs both the outsider's and the insider's view.
• Objectivity is a continuous goal, not an achieved result.
• Reflexivity: examine yourself, document everything, declare your background.
• No method is superior; choose the one that fits the question — and combine them.
• Participant observation: live as one of them, for about a year.
• Genealogy and key informants are the anthropologist's early tools.
• A sample is representative if all relevant strata are included and units are chosen by chance.
• How a sample is selected matters more than how big it is.
• Surveys show aggregate patterns invisible in single families.
• The interview is flexible — and therefore unstable.

Traps to avoid

• Methodology is the study of methods, not a synonym for method.
• Statistical stratification is not social stratification.
• Sampling error is unavoidable; non-sampling error comes from faulty design or conduct.
• A sample of villages near the highway is purposive, not random.
• NCERT is wrong that China does not hold a regular census — it held its seventh in 2020.
• Malinowski was an enemy alien as an Austro-Hungarian subject; ‘Poland annexed by Germany’ is incorrect.
• NCERT's 2011 child sex ratios are provisional (India 914; final 918).
• Punjab's and Haryana's child sex ratios rose between 2001 and 2011, despite ‘falling steadily’ in the text.
• An exit poll surveys actual voters after voting; an opinion poll surveys intentions before.
• The interview lies between the survey and participant observation.

Prelims — what UPSC asks

• Prelims asked what an exit poll is — a post-election survey of voters about whom they actually voted for (1994); the wrong options (same as an opinion poll; ‘most exactly predicts’ results) fail on this chapter's ideas of sampling and non-sampling error.
• Know the institutions: the decennial Census (complete enumeration) vs NSO sample surveys; the 2019 merger of NSSO and CSO into the National Statistical Office; the PLFS (from 2017–18) and NFHS. Use final Census figures (child sex ratio 918/919 in 2011).

Mains — where it appears and how to answer

• Essay: ‘What is research, but a blind date with knowledge’ (2021) — field work's correction of first impressions, unsuspected findings, objectivity as a continuous process, reflexivity, the ‘luck’ of random sampling, dialogic research. This is the weakest link: use the chapter for concrete illustrations, not as the whole argument.
• GS answers generally: when citing data, name the source and type (Census = enumeration; NSS/PLFS/NFHS = sample surveys with sampling error); distinguish aggregate trends from individual cases; pair numbers with qualitative evidence (triangulation).
• Methods in society questions: a GS1 question on, say, the joint family or the sex ratio is stronger if it notes which kind of evidence would settle it — the chapter's own example (survey for prevalence, interviews or participant observation for women's status).

Tap an option. NCERT's own exercise questions come first; the rest test the figures and common confusions.

Open any question to attempt it in the Question bank — your attempt is saved to your progress.