Techniques of Data Collection

Questionnaire and Schedule: The Same Questions, a Different Hand Holding the Pen

Sociology Optional Notes {PDF}

Complete Notes Package (Paper I & II) | 1000+ Pages | Instant Download

Discounted Price | Buy Now @ ₹600 →

• A questionnaire is a document containing a set of questions, the answers to which the respondent provides personally.
  • Usually sent by post, though sometimes delivered by hand, often with a self-addressed stamped envelope and a covering letter explaining the survey’s purpose.
  • Suits research where very large samples are wanted, costs must stay low, and a moderate response rate is acceptable.
• A schedule is functionally the same set of questions, but filled in by a trained investigator or enumerator in the respondent’s presence rather than by the respondent alone.
  • The single distinction to hold onto, because everything else follows from it: in a questionnaire, the respondent fills the form; in a schedule, the investigator fills it.

• This distinction is not a pedantic technicality in the Indian context — it is decisive.
  • Given India’s literacy distribution and linguistic diversity, a form arriving by post carries little authority in most households.
  • An enumerator-administered schedule is therefore the default choice for any nationally representative survey: the Census of India and the National Sample Survey both rely on trained enumerators rather than mailed forms.

Framing Good Questions

• Clear and unambiguous — a question presuming the respondent already understands a specific policy proposal will be misread by anyone who does not.
• Relevant to something the respondent has actually thought about.
• Short, so a respondent can read, understand, and answer without strain.
• Free of negation, since a negative word buried in a statement is easy to miss and reverses the intended meaning.
• Free of loaded or biased terms, which visibly nudge respondents toward a particular answer.
• Pitched to respondents who are actually competent and willing to answer — asking daily-wage labourers to comment on macroeconomic policy produces answers that are not worth collecting.

Types of Questions

• By relevance to the topic:
  • Primary questions bear directly on the research question.
  • Secondary questions are indirect, but useful for checking a respondent’s truthfulness.
  • Tertiary questions are neither, but help build rapport and structure the instrument.
• By response format:
  • Closed-ended questions offer a fixed set of response options — easy to code and compare, but the researcher may have omitted an option that actually fits the respondent.
  • Open-ended questions invite free responses in the respondent’s own words — richer, and capable of surfacing answers the researcher never anticipated, but harder to classify and more time-consuming to analyse.
• By directness:
  • Direct questions ask about the respondent themselves.
  • Indirect questions ask about “people like them,” which can surface more honest answers on sensitive topics.
• By level of measurement:
  • Nominal questions sort responses into unordered categories (male/female).
  • Ordinal questions rank responses (smoking: regularly/occasionally/never).
  • Interval questions place responses into equally-spaced numerical bands (age, income brackets).

Constructing and Piloting a Questionnaire

• Constructing a questionnaire proceeds through a systematic sequence: preparation, drafting, self-evaluation, external evaluation by colleagues or experts, revision, a pilot study, a further revision based on what the pilot reveals, and only then a final draft.
• A pilot study is a small-scale trial run conducted before the main research, on people drawn from the relevant population but excluded from the eventual sample.
• It exists to catch problems the researcher cannot see from their desk: whether questions actually make sense to real respondents, and whether rapport-building strategies work.
• For funded research, it can also determine whether the whole project should proceed at all — an unsuccessful pilot has, in practice, led to entire studies being abandoned before real money was spent on a flawed design.
• Even careful wording is not immune to distortion. In a well-known 1939 American survey, respondents asked to place themselves in the “upper,” “middle,” or “lower” class overwhelmingly chose “middle” (88 per cent).
• When “lower” was relabelled “working class,” the distribution of answers shifted dramatically — a single word choice in a closed-ended question can materially change what a survey appears to find.

Administering a Questionnaire

• A structured interview, where a trained interviewer asks the fixed questions aloud, reduces misunderstanding but reintroduces interviewer bias.
• Postal delivery is cheap and wide-reaching, but the response rate is low and non-response skews toward the more educated and more interested.
• Group administration is efficient, but requires preventing respondents from discussing answers among themselves before submitting.
• Telephone administration is fast, but makes rapport and sensitive questions harder, and permits no visual aids.
• Email/online delivery is very cheap, but the sample can be skewed by unequal digital access — a genuinely material limitation in a country with an uneven digital divide.

Advantages of the Questionnaire

• Low cost, minimal staffing, and fast turnaround relative to face-to-face methods.
• Reach across geographically dispersed populations.
• No interviewer bias, since no interviewer is present to influence answers by tone or expression.
• Greater anonymity, which encourages honesty on sensitive or socially undesirable topics.
• Respondent’s convenience — it can be completed at leisure.
• Standardized wording across every respondent, which supports comparison, and stable, uniform administration.

Disadvantages of the Questionnaire

• Demands literacy, which rules out large sections of a population like India’s.
• Response rates are typically low (commonly 30-40 per cent) and biased in who bothers to respond.
• Different respondents can interpret the same wording differently, with no way to correct the misunderstanding.
• No way to verify who actually completed the form, or whether they consulted others before answering.
• Unanswered items go unchased, and there is no opportunity for follow-up probing.

Interview

• Lindzey Gardner defines a research interview as a two-person conversation initiated by the interviewer for the specific purpose of obtaining research-relevant information, focused on content specified by the study’s own objectives. This is deliberately distinguished from ordinary conversation by its systematic preparation, controlled execution, and tie to a specific research question.
• Black and Champion identify its core characteristics: face-to-face personal communication, formally equal status between interviewer and respondent, verbal questions and responses, information recorded by the interviewer, a relationship that is typically transitory between strangers, and considerable flexibility of format.

Types of Interview, By Structure

• A structured interview follows fixed questions, fixed order, and often fixed response categories — high reliability and easy comparison, at the cost of no probing.
• An unstructured interview is closer to a guided conversation — high validity and depth, but low comparability and low reliability.
• A semi-structured interview, which most real research actually uses, follows a topic guide while leaving room to probe.

Types of Interview, By Style

• A non-directive interviewer withholds their own opinions and any sign of approval or disapproval, to avoid steering the respondent’s answer.
• Howard Becker argued this passive style can actually suppress information, and used a more active, even confrontational approach with Chicago schoolteachers to draw out attitudes about students’ class and ethnic background that teachers would ordinarily keep hidden.
• Ann Oakley, by contrast, favours an empathetic approach: a collaborative, non-hierarchical relationship with respondents — even answering their questions and offering help when asked — produced richer, more honest data than either strict neutrality or confrontation (see the article on qualitative and quantitative methods for the fuller feminist-methodology context).

The Focused Interview and the Focus Group

• Robert Merton, with Marjorie Fiske and Patricia Kendall, developed the focused interview for interviewing people known to have shared a specific situation, concentrating on their subjective response to that shared exposure.
• The modern focus group is this technique’s direct group-level descendant — a lineage worth naming, since it is easy to mistake focus groups for a purely commercial invention with no roots in sociological method.
• A focus group typically brings together 6 to 12 participants with a moderator around a carefully specified topic.
• Its distinctive value: the data are produced by interaction — participants challenge, refine, and build on each other’s answers, revealing how opinion is socially negotiated rather than simply held in isolation.
• Its distinctive risks: group conformity, one or two dominant speakers crowding out the rest, and the moderator’s own influence on the discussion.

Reducing the Moderator’s Influence in a Focus Group

• Keep the moderator’s role facilitative rather than interrogative, using open, non-leading prompts.
• Avoid signalling approval or disapproval through tone or expression.
• Actively draw in silent participants while managing dominant ones.
• Use reasonably homogeneous groups so status differences do not silence people, in a neutral venue.
• Record sessions verbatim, run multiple groups to expose moderator-specific effects, and have a second, independent analyst code the transcripts.

Interviewer Bias and Its Named Varieties

• Social desirability bias: respondents report the answer that flatters them rather than the accurate one — why self-reported caste prejudice, alcohol consumption, and voter turnout are systematically over- or under-stated.
• The interviewer effect: an interviewer’s own visible caste, gender, age, or class can change what a respondent is willing to say at all.
• Example: interviewing destitute men about the causes of their situation, those interviewed by a known supporter of Prohibition tended to blame alcohol, while those interviewed by a committed socialist tended to blame the industrial system — the interviewer’s own visible position shaped the explanation offered.
• Example: Puerto Rican respondents in a study of mental health reported more symptoms than Jewish, Irish, or Black respondents asked the identical list — not because they experienced more distress, but because they were more willing to admit to it.
• An interviewer must also understand the social conventions of those being interviewed to interpret responses correctly, since identical answers can carry different meanings across groups.

Advantages and Disadvantages of the Interview

• A high response rate, and the ability to probe in depth.
• The ability to build rapport and trust with groups (professional criminals, for instance) that would never consent to a questionnaire.
• The interviewer’s ability to explain difficult terms on the spot, which eases administration with less-educated respondents.
• The chance to observe non-verbal behaviour, and greater completeness overall.
• Possible concealment or distortion out of fear of exposure, and higher cost and time than a questionnaire.
• Sensitivity to the respondent’s mood, and variability across interviewers, especially in unstructured formats.
• Reduced anonymity, which can suppress honesty on sensitive questions.
• Hammersley and Gomm’s general critique: internal feelings and motives can never be fully expressed through verbal responses alone.

Observation

• Lindzey Gardner defines observation as the selection, provocation, recording, and encoding of behaviour and settings in their natural surroundings, consistent with a study’s empirical aims.
• Scientific observation differs from other data-collection methods in several ways: it is always direct, typically conducted in a natural setting, usually less structured, and — in its purest form — qualitative, since its core purpose is discovering how subjects experience and make sense of their own world.

Non-Participant Observation

• The researcher watches without joining the group’s own activities.
• Less reactive in principle, permits systematic recording against a fixed schedule, and suits publicly visible behaviour.
• Surface conduct is recorded without access to the meaning behind it, and true non-participation is something of a fiction, since a visible stranger watching a group changes how that group behaves.

Participant Observation

• The researcher takes part in the group’s own life over an extended period while recording it.
• Bronislaw Malinowski established the method as a discipline through his Trobriand-islands fieldwork, setting its enduring aim as grasping “the native’s point of view, his relation to life, his vision of his world.”
• In India, M. N. Srinivas used it to study sanskritisation in Mysore, while André Béteille used it to study social inequality by class, status, and power in a Tanjore village.

Four Roles for a Participant Observer

• Complete participant — identity fully concealed, fully covert.
• Participant as observer — a group member whose research role is openly known.
• Observer as participant — mainly observing, with only brief, openly acknowledged involvement.
• Complete observer — fully detached and unnoticed.

Covert Versus Overt Observation

• Covert work avoids the observer effect and can access settings that would otherwise refuse a researcher entirely, but raises much sharper ethical questions.
• Leon Festinger’s covert infiltration of a doomsday cult produced the theory of cognitive dissonance.
• Laud Humphreys’ covert observation of anonymous sexual encounters in public restrooms, followed by tracing participants to their homes to interview them under a false pretext, remains the standard case study in research ethics — the data were genuinely valuable and the method genuinely indefensible.
• William Foote Whyte, studying an Italian-American slum community, adopted a middle path: introducing himself honestly as a writer without volunteering the full scope of his research. He described starting as a non-participant observer and, as he was accepted, gradually becoming “almost a non-observing participant.”

The Observer Effect and Going Native

• The observer effect, and its most famous named instance, the Hawthorne effect: at the Western Electric Hawthorne Works studies (1924-1932), associated with Elton Mayo, worker productivity rose regardless of how lighting and conditions were varied — apparently simply because workers knew they were being studied.
• Going native and its associated failure of over-rapport, described by S. M. Miller: a participant observer identifies so completely with the group that the analytic distance needed to study it collapses, turning the researcher into an advocate rather than an observer. The mirror-image failure is staying so detached that no one ever confides anything real.
• Access and gatekeepers: Whyte’s slum study entered the community through a well-placed insider he called “Doc” — a gatekeeper who opens one door while simultaneously and quietly closing others, handing the researcher their own enemies, blind spots, and position within the group’s internal politics.

Strengths and Weaknesses of Observation

• Unmatched validity, since conduct is recorded in its natural setting rather than reported after the fact.
• Access to the unspoken and the routine, which no respondent could ever report because they have never consciously noticed it.
• The capacity to discover what the researcher did not think to ask in advance, and the possibility of studying groups that are hostile, illiterate, or would never willingly complete a form.
• Time- and cost-intensive, and a single setting cannot support statistical generalization.
• Difficult or impossible to replicate, since the “instrument” is a particular researcher with a particular biography.
• Recording is selective and memory-dependent, and the observer’s own presence can alter the field.
• Entry and access can be difficult to secure, particularly for illegal or high-status activities, and covert work carries severe ethical exposure.

Ethnography

• Ethnography is the broader research tradition observation belongs to: a systematic, long-term study of how people in a social setting actually lead their lives.
• Most commonly combines in-depth interviews, participant observation, and qualitative documents, sometimes with quantitative data.
• Entered the social sciences through anthropologists studying small-scale, pre-industrial societies before spreading across sociology more broadly.
• Critical ethnography, its politically engaged variant, uses the same tools to develop and test theories about social structure and to understand how oppression is experienced from within — Paul Willis’s study of working-class boys and schooling is a standard example.
• It faces a specific circularity charge: the same body of description is used both to build a theory and to confirm it, so experience risks being interpreted only in terms the theory already expects.

Digital Ethnography

• Digital ethnography relocates participant observation online: sustained immersion in forums, gaming worlds, social media communities, and messaging groups, together with analysis of the visual culture these spaces generate.
• What participation means when the researcher is only a screen name, and what counts as “the field” when a community has no physical location.
• Where public ends and private begins, when a forum post is technically visible but written for a small audience.
• The possibility of identity play, where the researcher can never be certain who is typing, and the fact that a platform’s own architecture, not the researcher, often decides what is even visible to study.
• Because these standards remain unsettled, researchers increasingly look to the Association of Internet Researchers’ ethical guidelines as an emerging point of reference — including treating the public/private boundary as something to actively manage, and making a researcher’s presence known wherever practically possible rather than defaulting to silent observation.

Case Study and Life History

• A case study is the intensive, holistic examination of a single unit — an individual, an institution, a community, an event, or an entire culture — usually studied in its full context and often over an extended period.
  • Yin defines it as an empirical inquiry into a contemporary phenomenon within its real-life context, where the boundary between phenomenon and context is not sharply defined, using multiple sources of evidence.
  • It is a research strategy, not merely a data-collection technique, since it typically draws on interviews, observation, and documents together.

Five Types of Case Study

• The critical case — chosen because it is especially useful for testing a hypothesis (a religious cult’s members’ reaction when a prophesied apocalypse fails to occur).
• The extreme or unique case — the only known example of something.
• The revelatory case — granting access to an aspect of social life previously closed to research.
• The longitudinal case — studied at two or more points in time to observe the effects of change directly.
• The representative or typical case — chosen because it illustrates a common, widespread phenomenon.

The Logic of Generalization From a Case Study

• Its logic of generalization is theoretical, not statistical — the point most often missed in criticizing it.
• A case study cannot establish how common something is, but it can establish that something is possible, refuting a claimed universal generalization with a single counter-example.
• It can trace how a causal mechanism actually unfolds, rather than merely showing that two variables correlate, and it can generate the very concepts a later, larger survey will go on to measure.
• Running several case studies of the same phenomenon — ideally by the same researcher, over the same period — can strengthen validity without claiming statistical representativeness.

Strengths and Weaknesses of the Case Study

• Strengths: depth, holism, sensitivity to context and process, high validity, and the ability to discover the genuinely unexpected.
• Weaknesses: no statistical generalization to a wider population, heavy dependence on the researcher’s own interpretation, difficulty of replication, and — where a case is chosen simply because it is dramatic — a real risk of selection bias.

Life History

• Life history (or biography) is a specific, extended type of case study, usually gathered through multiple unstructured interviews cross-checked against personal documents.
• Its unique payoff is showing the intersection of an individual biography with wider historical change — how large structural shifts are actually lived and made sense of by one person.
• It has been especially valuable for recovering the voices of people written out of official records, central to Dalit, feminist, and labour-history research.
• Thomas and Znaniecki’s classic study of Polish peasants migrating to America, built substantially from 764 personal letters, remains the field’s founding exemplar.
• Its dangers: the unreliability of memory, retrospective rationalization (imposing a coherent narrative on a life not actually lived coherently), self-serving editing, and the impossibility of generalizing from a single life.

Content Analysis and Other Secondary Sources

• Content analysis works on communication rather than on people directly: newspapers, textbooks, films, advertisements, speeches, and social media posts.
  • Bernard Berelson’s classic definition — a technique for the objective, systematic, and quantitative description of the manifest content of communication — also marks the technique’s central limitation.
  • Klaus Krippendorff’s later, wider definition extends it to making replicable and valid inferences from texts to the contexts of their use.
• The manifest/latent distinction is the crux of the whole method.
  • Manifest content is what is literally present and countable — how many times a word appears, how many female characters appear on screen — and counting it is reliable because it requires no interpretation.
  • Latent content is the underlying meaning, ideology, and framing beneath the surface — richer, but less reliable, so inter-coder agreement must be reported whenever latent coding is used.

Four Approaches to Content Analysis

• Ray Pawson identifies four distinct approaches within the technique:
• Formal content analysis — objective counting of pre-specified features, as simple as a word count.
• Thematic analysis — tracing the intentions and ideological biases behind how content was produced.
• Textual analysis — examining the specific linguistic devices a text uses to steer interpretation (calling the same event a “riot” or an “uprising” is a standard illustration).
• Audience analysis — studying how audiences actively interpret content, rather than treating them as passive recipients.

Strengths and Weaknesses of Content Analysis

• Unobtrusive, since a text cannot react to being studied, and it permits studying the past and the dead.
• Cheap, replicable, and easy to repeat over time, which suits tracing trends, and can handle very large volumes, a capacity computational and AI-assisted text analysis has now amplified.
• Confined to what was actually recorded, which already reflects someone’s prior selection.
• Establishes what was communicated, never how it was received, and strips a text from the context of its own production.

AI and Qualitative Coding

• Recent scholarship (2025) has begun examining whether large language models can perform qualitative coding at a standard comparable to trained human coders, particularly for deductive coding against a clear codebook.
• Early results are promising for well-structured tasks, but researchers flag real concerns about reproducibility (a closed, proprietary model can change or disappear, so another researcher cannot verify the original coding) and about bias embedded in the model itself.
• Best treated as a genuinely emerging technique rather than a settled addition to the standard toolkit.

Other Secondary Sources

• Official statistics (census data, crime and unemployment figures) divide researchers by camp: positivists like Durkheim accepted them largely at face value; phenomenologists and ethnomethodologists reject this, arguing the statistics are themselves the product of officials’ own taken-for-granted assumptions; a conflict perspective treats them as shaped by the needs and interests of the state agencies that produce them.
• Historical sources are essential for studying change over periods no live fieldwork could observe — Weber’s study of Protestant belief and the rise of capitalism would have been impossible without them. Their survival is always partial, so generalizing from what happens to have survived carries real risk.
• Life documents — diaries, letters, memoirs, photographs — offer direct insight into subjective states, but are hard to obtain in a representative spread and open to multiple interpretations. A private diary never meant for other eyes carries correspondingly greater validity than one written with an audience in mind.
• Visual material can be examined through content analysis (counting depicted features), semiotic analysis (how visual signs generate meaning through their relationship to other images), or discourse analysis (relating an image to a society’s dominant ways of talking about its subject).
• John Scott’s four criteria for evaluating any secondary source: authenticity (is the document genuine, and its authorship what it claims), credibility (how much distortion has crept in), representativeness (how typical the surviving material is), and meaning (whether the document can actually be correctly understood).

Social Survey

• A social survey collects standardized data from a large number of people, most practically through questionnaires or schedules.
  • Descriptive — establishing what currently exists, such as a government census.
  • Explanatory — identifying the causes behind an observed pattern (the element most sociological surveys carry to some degree).
  • Predictive — forecasting the likely effects of a future policy.
  • Evaluative — assessing the actual results of a policy already implemented.
• One of the earliest systematic social surveys is Charles Booth’s investigation into the life and labour of London’s poor, published across seventeen volumes between 1889 and 1902.
  • It established the empirical documentation of poverty as a legitimate object of social research well before sociology had settled methodological conventions of its own.
• Not every social survey relies on self-report alone. Field experiments embedded in survey-style research can measure discrimination directly rather than asking about attitudes toward it.
  • Example: in one classic British study, matched job, housing, and insurance applications were submitted by an Englishman, a Hungarian, and a West Indian applicant, each with identical qualifications. The Englishman succeeded most consistently, the West Indian least — directly demonstrating discrimination that a simple attitude survey would likely have understated.

Measuring Attitudes: Scales and Projective Techniques

• The Likert scale presents a statement and asks respondents to indicate their degree of agreement along a fixed ordinal range (commonly five points) — simple to construct and easy to score.
• Bogardus’s social distance scale measures the degree of intimacy or distance a person will accept with members of another social, ethnic, or religious group, from close kin by marriage down to exclusion altogether — well suited to measuring prejudice and social exclusion.
• The Thurstone scale has a panel of judges first rate a large pool of statements for their relative favourability, retaining only those the judges agree on, so a respondent’s score reflects an already-calibrated scale of intensity.
• Projective techniques — the Rorschach inkblot test, the Thematic Apperception Test, draw-a-person tasks — ask a respondent to interpret or produce an ambiguous stimulus, on the reasoning that people project their own attitudes and anxieties onto a meaningless prompt, surfacing sensitive attitudes a direct question would not.