Sociological Research Methodology and Methods
What it is
Unit 2 asks how sociologists know what they claim and how they produce evidence: the philosophy of science, hermeneutics, objectivity and ethics; research design from induction and deduction to hypotheses; ethnographic, survey, historical and comparative methods; and techniques from sampling and scaling to statistics, interviewing and reporting. UGC-NET mixes attribution items with method-to-situation matching and short computations.
Core concepts
Philosophy of science and epistemology. Comte coined "sociology", founded positivism and gave the law of three stages: theological, metaphysical, positive. The Vienna Circle made verifiability the test of meaning; Popper (The Logic of Scientific Discovery) replaced it with falsifiability as the demarcation of science. Kuhn's The Structure of Scientific Revolutions (1962): normal science within a paradigm, anomalies, crisis, revolution, incommensurability. Lakatos offered research programmes (hard core, protective belt); Feyerabend's Against Method (1975) argued for epistemological anarchism, with no binding method. Positivism assumes unity of method with natural science; interpretivism stresses meaning; Bhaskar's critical realism posits unobservable causal mechanisms.
Hermeneutic traditions. Dilthey separated the human sciences (Geisteswissenschaften), which understand (Verstehen), from the natural sciences, which explain (Erklären). Weber folded Verstehen into causal explanation, distinguishing direct observational from explanatory understanding. The hermeneutic circle moves between part and whole; Gadamer's Truth and Method (1960) calls understanding a fusion of horizons; Giddens's double hermeneutic notes that sociology interprets a world its actors have already interpreted.
Objectivity, reflexivity, ethics and politics. Weber demanded value-freedom in analysis but accepted value relevance in choosing problems; Myrdal asked researchers to state value premises openly. Gouldner attacked value-free sociology in "Anti-Minotaur" (1962) and called for a reflexive sociology (The Coming Crisis of Western Sociology, 1970); Becker's "Whose Side Are We On?" (1967) named the hierarchy of credibility; feminist standpoint theory (Dorothy Smith, Sandra Harding) values knowledge from the margins. Ethics: informed consent, no harm, confidentiality, anonymity; Humphreys's Tearoom Trade (1970) and Milgram's obedience studies are classic deception controversies. Politics enters through sponsorship and use of findings, as in the US Army's Project Camelot in the 1960s.
Research design; reading data and documents. A design links questions to evidence: exploratory, descriptive, explanatory (diagnostic) or experimental; cross-sectional or longitudinal (panel: same people; cohort: same birth or event group; trend: fresh samples). Primary data are first-hand; secondary data come from censuses, registers and earlier studies. John Scott judges documents by authenticity, credibility, representativeness and meaning; content analysis counts manifest features of texts. Internal validity asks whether the causal claim holds; external validity whether it generalises.
Induction, deduction; fact, concept, theory; hypotheses. Deduction runs theory → hypothesis → test; induction runs observation → generalisation, as in Glaser and Strauss's grounded theory (1967); Wallace's wheel of science joins the two in a cycle. Goode and Hatt: theory guides fact-finding, and facts initiate, reformulate, reject and clarify theory. Concepts are operationalised into variables (independent, dependent, intervening). A hypothesis is a testable proposition about variables: null H0 (no relation) against alternative H1. Rejecting a true H0 is a Type I error (probability α, the significance level); retaining a false H0 is a Type II error (β). A research question is open; an objective states what the study will do.
Ethnography, survey, historical and comparative methods. Ethnography is long immersion by participant observation, pioneered by Malinowski; Pike's emic (insider) and etic (outsider) views frame it. The survey gathers standardised data from a sample for statistical generalisation. The historical method tests sources by external criticism (authenticity) and internal criticism (credibility). The comparative method sets cases side by side; Durkheim held comparative sociology to be sociology itself and preferred concomitant variation among J.S. Mill's methods (agreement, difference, joint, residues, concomitant variation); Weber compared world religions.
Sampling. Draw from a sampling frame. Probability designs give each unit a known non-zero chance: simple random, systematic (every kth unit, k = N/n), stratified (random within homogeneous strata), cluster and multistage. Non-probability designs: convenience, purposive (judgemental), quota (non-random cousin of stratified) and snowball (hidden populations). Sampling error falls as n rises (standard error of the mean = σ/√n); non-sampling error such as non-response or interviewer bias does not.
Questionnaire, schedule and measurement. The respondent fills a questionnaire, often by post, so literacy is needed; the enumerator fills a schedule face to face, suiting non-literate respondents and raising response. Avoid leading, loaded and double-barrelled questions; pilot first. Stevens (1946): nominal, ordinal, interval (no true zero), ratio (true zero). Scales: Likert (summated ratings), Thurstone (equal-appearing intervals, rated by judges), Guttman (cumulative scalogram), Bogardus (social distance), Osgood (semantic differential). Reliability is consistency (test-retest, split-half, inter-rater); validity is measuring the intended thing (face, content, criterion, construct).
Statistical analysis. Averages follow the level: mode (nominal), median (ordinal), mean (interval, ratio). Dispersion: range, quartile deviation, standard deviation, coefficient of variation = SD/mean × 100. Positive skew: mean > median > mode. Pearson's r (−1 to +1) suits interval data; Spearman's rho = 1 − 6Σd²/[n(n² − 1)] suits ranks; chi-square, Σ(O − E)²/E with df = (r − 1)(c − 1), tests association between nominal variables; the t-test compares two means.
Observation, interview, case study; analysis and reporting. Observation is participant or non-participant, structured or unstructured, overt or covert; the Hawthorne studies showed that being observed changes behaviour. Interviews are structured, semi-structured or unstructured; Merton's focused interview anticipated the focus group. The case study examines one unit in depth, from Le Play's family monographs to Thomas and Znaniecki's The Polish Peasant in Europe and America (1918–20), built on letters and life histories. Analysis codes, tabulates and interprets; Denzin's triangulation combines data, investigators, theories or methods. A report runs problem, review, design, findings, discussion, conclusion, references.
Worked example
A. Describing household size. Eight households (illustrative data) report 2, 3, 4, 4, 5, 6, 7, 9 members.
| x | x − mean | (x − mean)² |
|---|---|---|
| 2 | −3 | 9 |
| 3 | −2 | 4 |
| 4 | −1 | 1 |
| 4 | −1 | 1 |
| 5 | 0 | 0 |
| 6 | 1 | 1 |
| 7 | 2 | 4 |
| 9 | 4 | 16 |
| Σ = 40 | 0 | 36 |
Mean = 40/8 = 5. Median = mean of the 4th and 5th values = (4 + 5)/2 = 4.5. Mode = 4. Variance = 36/8 = 4.5, so SD = √4.5 ≈ 2.12 (with n − 1: 36/7 ≈ 5.14, s ≈ 2.27). Second route: Σx² = 236 and Σx² − n × mean² = 236 − 8 × 25 = 36, the same sum of squares. CV = 2.12/5 × 100 ≈ 42.4%. Mean 5 > median 4.5 > mode 4: positive skew.
B. Testing association. 200 respondents (illustrative), asked whether they approve of inter-caste marriage:
| Group | Approve | Disapprove | Total |
|---|---|---|---|
| Urban | 60 | 40 | 100 |
| Rural | 40 | 60 | 100 |
| Total | 100 | 100 | 200 |
H0: approval is independent of residence. Each expected frequency = row total × column total / N = 100 × 100/200 = 50. Each cell gives (O − E)²/E = 10²/50 = 2, so χ² = 4 × 2 = 8. Second route, the 2×2 shortcut: N(ad − bc)²/[(a + b)(c + d)(a + c)(b + d)] = 200 × (3600 − 1600)²/(100 × 100 × 100 × 100) = 200 × 4,000,000/100,000,000 = 8. df = (2 − 1)(2 − 1) = 1; critical values are 3.841 at 0.05 and 6.635 at 0.01, so reject H0 at both. Yates's correction gives 4 × 9.5²/50 = 7.22, still significant. Phi = √(8/200) = 0.2: significant, yet weak.
Common traps
- Questionnaire versus schedule: the respondent fills the first, the enumerator the second.
- Quota sampling is non-probability even though it mirrors stratified sampling.
- Type I is rejecting a true null (α); Type II is retaining a false one (β).
- Reliability is not validity: a consistently miscalibrated scale is reliable but invalid.
- Falsification is Popper's, paradigm Kuhn's, epistemological anarchism Feyerabend's, the double hermeneutic Giddens's.
- Interval scales lack a true zero, so 30°C is not twice 15°C; ratio scales (income, age) have one.
Speed technique
- Level decides the statistic: nominal → mode, chi-square; ordinal → median, Spearman; interval or ratio → mean, SD, Pearson, t.
- For a 2×2 chi-square, use N(ad − bc)² over the product of the four marginal totals; df is 1.
- Skew: the mean is dragged toward the long tail.
- Situation cues: hidden population → snowball; non-literate respondents → schedule; same people over time → panel.
Check yourself
- Which sampling method takes every kth unit after a random start?
Show answer
Systematic sampling — k = N/n. - Who made falsifiability the demarcation criterion of science?
Show answer
Karl Popper — against verificationism. - Five pairs of ranks give Σd² = 10. What is Spearman's rho?
Show answer
0.5 — 1 − (6 × 10)/(5 × 24) = 1 − 60/120. - How many degrees of freedom has chi-square in a 3 × 4 table?
Show answer
6 — (3 − 1) × (4 − 1). - Which scale is cumulative, so that endorsing a stronger item implies endorsing all weaker ones?
Show answer
Guttman scale — scalogram analysis.
Try it: Sociological Research Methodology and Methods questions
Real questions from the NET Sociology bank on exactly this skill. Pick an answer to see the full solution — the intuition, the worked steps, the faster methods and the traps.
The law of three stages, by which human thought passes from the theological through the metaphysical to the positive stage, was propounded by:
Show the answer and worked solution
Answer: option A
Auguste Comte, who coined the word 'sociology' and founded positivism, held that every branch of knowledge passes through theological, metaphysical and positive (scientific) stages.
Spencer's evolutionism moves from militant to industrial society, and Marx's history moves through modes of production; Durkheim built on Comte's positivism but did not propound this law.
So the law of three stages was propounded by Auguste Comte, option A.
A list of questions that an enumerator reads out and fills in on the respondent's behalf during a face-to-face meeting, suitable for respondents who cannot read, is called a:
Show the answer and worked solution
Answer: option B
A questionnaire is filled in by the respondent, often after receiving it by post, so it requires literacy; a schedule is filled in by the enumerator in a face-to-face meeting.
Because the enumerator records the answers, a schedule suits non-literate respondents and usually yields a higher response rate.
So a set of questions filled in by the enumerator is a schedule, option B.
A researcher studying people who inject drugs, for whom no list exists, asks each participant to refer others they know who might take part. This design is:
Show the answer and worked solution
Answer: option C
When a population is hidden and has no sampling frame, the researcher can begin with a few known members and ask each to refer others, so that the sample grows through referral chains.
This is snowball sampling, a non-probability design used for hidden or hard-to-reach groups.
So recruiting through participants' referrals in a hidden population is snowball sampling, option C.
Match the scales in List – I with their defining logic in List – II and choose the correct answer from the options. List – I (Scale) | List – II (Logic) --- | --- (a) Likert scale | (i) Cumulative scalogram (b) Thurstone scale | (ii) Social distance (c) Guttman scale | (iii) Summated ratings (d) Bogardus scale | (iv) Equal-appearing intervals
Show the answer and worked solution
Answer: option A
Likert's scale adds up respondents' agreement scores on several items, the method of summated ratings (a – iii). Thurstone's scale uses items placed by judges at equal-appearing intervals (b – iv).
Guttman's scale is cumulative, so endorsing a stronger item implies endorsing all weaker ones (c – i). Bogardus measured willingness to accept members of other groups at varying closeness, the social distance scale (d – ii).
So matching List I, from "Likert scale", gives a – iii, b – iv, c – i, d – ii, option A.
Match the thinkers in List – I with the concepts in List – II and choose the correct answer from the options. List – I (Thinker) | List – II (Concept) --- | --- (a) Wilhelm Dilthey | (i) Triangulation (b) Hans-Georg Gadamer | (ii) Epistemological anarchism (c) Paul Feyerabend | (iii) Fusion of horizons (d) Norman Denzin | (iv) Human sciences as sciences of understanding
Show the answer and worked solution
Answer: option C
Dilthey separated the human sciences, which understand, from the natural sciences, which explain (a – iv). Gadamer's Truth and Method treats understanding as a fusion of horizons (b – iii).
Feyerabend's Against Method argued for epistemological anarchism (c – ii). Denzin set out triangulation of data, investigators, theories and methods (d – i).
So matching List I, from "Wilhelm Dilthey", gives a – iv, b – iii, c – ii, d – i, option C.
Answer above — every one shows its working.