Doing Sociology: Research Methods

Written and reviewed by Subhra Gupta, Humanities faculty at UniDrill (M.A. Sociology, University of Delhi · 7 years teaching Humanities). Based on NCERT Sociology Class XI (Introducing Sociology), Chapter 5.

If everyone already knows society, why call sociology a science?

You have lived in a family, a neighbourhood and a school, so you already "know" a lot about what sociologists study. Your aunt can tell you how families work. Your shopkeeper knows how bargaining works. So what makes a sociologist different? Not how much she knows, and not even what she knows, but how she acquires her knowledge. The answer is method: the procedures through which knowledge is gathered. That is the reason method has special importance in sociology.

There is a second reason. Sociologists want to know not only what a bystander can see, but also the opinions and feelings of the people involved: what friendship means in a culture, what a believer thinks he is doing in a ritual, how a shopkeeper and customer read each other's gestures during bargaining. Sociology has to combine the outsider's and the insider's views, and that needs careful method.

Methodology: the study of method

People often use "methodology" to mean "method", but strictly it means the study of method. Methodological issues are the general problems of scientific knowledge gathering that go beyond any one technique. The biggest of these is objectivity.

Objectivity and subjectivity

In everyday language, objective means unbiased, neutral, based on facts alone. Subjective means based on individual values and preferences. Every science aims to be objective, but it is much harder in the social sciences.

Why is it harder? First, bias. A sociologist studying families is herself a member of one. Even when studying a caste or religious community other than her own, she may carry attitudes about it from her own social surroundings. Second, there are many versions of the truth in the social world. A shopkeeper and a customer may differ on what is a "good" price, and a young person and an old person on "good food". Sociology does not try to decide which version is correct, because it is really interested in what people think and why. Third, sociology is multi paradigmatic: competing, incompatible schools of thought, such as the functionalist and conflict views, coexist within the discipline.

How do sociologists guard against bias?

Pause and predict: a researcher from a wealthy family studies street vendors and mentions at the start of her report that her own family is wealthy and that she had never bought from a street stall. Which safeguard is this?

Check your answer: stating one's own background, a part of reflexivity. It alerts readers to a possible bias so they can mentally adjust for it. It does not remove the bias, which is why documentation and reflexivity are used together.

Two views. For the exam: the old notion of an "objective, disinterested" social science is widely considered outdated and unattainable, and objectivity should be seen as the goal of a continuous process. Another reading: Max Weber argued for value neutrality, meaning that a researcher should keep personal values out of the analysis of evidence even though values rightly shape the choice of topic. Some scholars therefore still defend a disciplined ideal of objectivity, and the two positions agree that bias must be openly guarded against.

Many methods, no single road

Because there are multiple truths and perspectives, there are also multiple methods. There is no single unique road to sociological truth. Each method has strengths and weaknesses, so it is futile to argue which is superior. The sensible question is whether the method suits the question. If you want to know whether most Indian families are still joint families, a census or survey is the best method. If you want to compare the status of women in joint and nuclear families, interviews, case studies or participant observation may be better.

Methods can be classified in several ways, but the lines between categories are a matter of convention and are not sharp.

The choice of method depends on the research question, the researcher's preferences, and the limits of time and resources. A modern trend is triangulation: using several methods on the same problem from different vantage points so that they complement each other. This chapter covers three primary data methods: participant observation, survey and interview.

Participant observation (fieldwork)

In participant observation, the researcher lives for a long period, usually about a year or more, among the people being studied, learns their language and takes part in daily life, trying to acquire the knowledge of an "insider". The overall aim is to learn the whole way of life of a community, as a small child learns about the world. The term field work comes from the natural sciences, where botanists, zoologists and geologists had to go out into "the field" and not only work in laboratories.

How fieldwork changed anthropology. Early anthropologists were "armchair scholars" who worked from reports of travellers, missionaries, colonial administrators and soldiers. James Frazer's The Golden Bough and Emile Durkheim's work on primitive religion were based on such second hand accounts. From the late 19th and early 20th centuries researchers began first hand surveys of languages, customs and beliefs, and relying on second hand accounts came to be seen as unscholarly. Since the 1920s, fieldwork has been the principal method of social anthropological training.

What a fieldworker actually does.

Malinowski and the "invention" of fieldwork. Bronislaw Malinowski, a Polish anthropologist settled in Britain, is widely believed to have established fieldwork as the distinctive method of social anthropology, although he was not the first to use it. In 1914, when the First World War broke out, he was visiting Australia, and since Poland had been annexed by Germany, Britain treated him as an "enemy alien" who had to be interned. He asked to serve his internment in the Trobriand Islands and spent about a year and a half there, living in a tent in the villages, learning the language and keeping detailed records and a daily diary. His influential post at the London School of Economics helped make fieldwork a mandatory part of anthropology training.

Fieldwork in sociology and in India

Sociologists use the same techniques, but the difference lies in the context: anthropologists typically went to remote tribal communities, while sociologists did fieldwork among all sorts of communities, and not always by "living in". William Foote Whyte, an American sociologist, spent three and a half years "hanging out" with a street gang of poor unemployed youth in an Italian American slum and wrote Street Corner Society. Michael Burawoy worked for several months as a machinist in a Chicago factory to write about work from the workers' view. One difficulty in studying a modern community is that the people are literate: they may read the report, and even disguised names will be recognised by the people themselves.

Village studies in India. In the 1950s and 1960s, village studies became the main preoccupation of Indian sociology.

Strengths and limits of participant observation.

Some scholars suggest dialogic formats to correct this: translating the work into the community's language, asking for their reactions and recording their responses. This may make research more controversial, but also more democratic.

Beyond the textbook. For the exam: fieldwork gives an insider's detailed picture but covers a small area. In reality: the link between early anthropology and colonialism is a recognised area of debate. Fieldwork sites such as Malinowski's were part of empires ruled by the researchers' own countries, which is one reason Indian scholars turned to villages. Why it matters: the choice of method is shaped by politics as well as by research questions.

Surveys: seeing the whole by studying a part

A survey gives an overview of a subject from information obtained from a carefully chosen representative set of people, called respondents. It is run by teams of researchers and investigators (or research assistants). Questions may be asked orally in person or by telephone, sent as written questionnaires by post, or answered electronically through email or websites. The survey is now common beyond sociology, from predicting election results to market strategies.

Its main advantage is that it lets us generalise results for a large population while studying only a small part of it, with manageable time, effort and money. This works because of sampling theory, and the key is a representative sample. Two principles guide the choice.

Note that this "stratification" is a statistical idea and differs from the sociological concept of stratification from Chapter 2. A chance sample is different from a purposive sample. If a team picks only villages near the main highway, or mostly middle class households, or households it knows, the sample is purposive and is unlikely to be representative. A coin shows what randomness means: heads or tails each have a 50 per cent probability, and which one appears depends on chance alone.

Worked example: choosing a sample. The hypothesis: living in smaller, more intimate communities produces greater harmony between communities than living in larger, impersonal ones. We study the rural sector of one state.

The common wrong step: picking the 10 most accessible villages near the main road. That makes the sample purposive, not random, and a convenient village may not be typical.

Sampling error and the margin of error. A sample will closely resemble its population, but not exactly. The chance of small differences can be specified, and this is the margin of error or sampling error. It arises not from researchers' mistakes but from using a small sample to stand for a large population. Researchers must report the size and design of the sample and the margin of error. The bigger the sample, the better the chance it is representative, and the extreme case is the census, which includes the whole population. But the way a sample is chosen matters even more than its size.

Census and the National Statistical Organisation. India's census, held every ten years, covers every household. Because it is so large, it cannot be held often, and many developed countries rely on sample surveys instead. The NSO conducts sample surveys every year on subjects such as family expenditure and employment. Every five years it conducts a bigger survey of about 1.2 lakh households covering over 6 lakh persons. That is only about 0.06 per cent of the population, yet because it is scientifically selected to be representative it can estimate population characteristics well.

Aggregate statistics. Some social problems appear only at the level of a whole population. The child sex ratio (girls per 1,000 boys in the age group 0 to 6) is such an aggregate or macro variable. You cannot see the problem by looking at individual families, since a family with more boys can be balanced out by others with more girls. Only a census or large survey reveals it. In India the child sex ratio was 945 in 1991 and 927 in 2001.

Pause and predict: a news channel reports that "most Indians prefer brand X" after a poll of 500 viewers who phoned in. What should a sociologist ask?

Check your answer: how was the sample selected, how big was it, and what is the margin of error? Callers choose themselves, so the sample is not random and cannot represent the population. Surveys with unspecified sample size, design and margin of error should not be trusted.

Interviews: a guided conversation

An interview is a guided conversation between researcher and respondent. It looks simple, but takes practice and skill. It lies between the structured questionnaire of the survey and the open ended interaction of participant observation. Its chief advantage is flexibility: questions can be rephrased, the order can change, productive subjects can be extended and unwelcome ones postponed, all during the interview. The same flexibility is a weakness. The interview is vulnerable to the respondent's mood or the interviewer's lapse of concentration, and it can be "an unstable and unpredictable format", working very well when it works and failing miserably when it does not.

Styles differ: a loose checklist of topics versus specific questions for everyone. Recording differs too: audio or video, note taking, or writing up from memory. Recorders can make a respondent uneasy, while fewer records may lose information. Researchers may also present a "cleaned up" narrative or keep the asides and digressions. Interviews are often used with other methods: long conversations with key informants support participant observation, and intensive interviews add depth to a survey. Success depends on access and rapport, the mutual trust between the respondent and the researcher.

Comparing the three methods

Timeline and key names

Exam traps

60-second revision

Where this chapter connects

• Previous: Culture and Socialisation uses fieldwork findings and Malinowski's definition of culture.
• Linked chapters: Sociology and Society first introduced the rule bound, checkable character of sociology, and Indian Sociologists returns to Srinivas and the village studies.
• Class XII link: Suggestions for Project Work puts surveys, interviews and observation to use in a student project.