amazon.co.ukResearch Methods
Quantitative and qualitative methods, sampling, positivism versus interpretivism and PET/E evaluation, with the research process and 30-mark exam technique for AQA 7192, OCR H580/02 and Eduqas AC2.
🎯 Learning Objectives
- Classify research data as quantitative/qualitative and primary/secondary, and match methods to positivism or interpretivism (AQA 4.1.3.1-3)
- Outline the practical, ethical and theoretical strengths and limitations of each main method (AQA 4.1.3.1)
- Describe the research process from aim to report, including stratified random sampling, quota sampling and pilot studies (AQA 4.1.3.4)
- Apply theoretical, practical and ethical considerations to the choice of topic, method and conduct of research (AQA 4.1.3.5)
- Evaluate positivism and interpretivism with named studies and methodological pluralism (AO3)
- Deploy methods knowledge in 10-mark, 20-mark and 30-mark answers across AQA Papers 1 and 3 and OCR H580/02
📋 Key Definitions and Core Concepts
Positivism: The view (Comte, Durkheim) that society is an external reality of social facts which can be studied objectively, like nature, to discover cause-and-effect laws; favours reliable, representative, quantitative data.
Interpretivism: The view (Weber, symbolic interactionism) that social reality is constructed through meanings, so sociology needs valid, authentic data on actors' interpretations; favours qualitative methods and verstehen.
Hypothesis: A testable prediction about the relationship between two or more variables, usually used in quantitative, deductive research.
Operationalisation: Turning an abstract concept (e.g. class, religiosity) into a measurable indicator, such as occupational category or church attendance.
Pilot study: A small trial run of a questionnaire or interview schedule to test clarity, timing and wording before the main study.
Stratified random sampling: A probability technique that divides the sampling frame into relevant strata (e.g. class, gender, ethnicity) then selects randomly from each in proportion, maximising representativeness.
Quota sampling: A non-probability technique where interviewers fill quotas of respondents of particular types until targets are met; faster but open to researcher bias.
Snowball sample: A non-probability technique where existing respondents recruit further contacts; used for hard-to-reach or deviant groups but unrepresentative.
Reliability: Whether repeating the research by another researcher would produce the same results; highest for standardised methods like questionnaires.
Validity: Whether the data gives a true, authentic picture of what is being studied; highest for naturalistic methods like participant observation.
Representativeness: Whether the sample reflects the characteristics of the wider population, allowing generalisation; depends on sampling technique and response rate.
Value freedom: The ideal that research should be objective and unbiased; Weber argued values may choose the topic but must not dictate the findings, challenged by Becker's 'whose side are we on?'
📌 Key Points
Positivism seeks reliable, representative data for causal laws (Durkheim); interpretivism seeks valid, authentic data for meanings (Weber).
Questionnaires and official statistics suit positivists; unstructured interviews and participant observation suit interpretivists.
Sampling: random, systematic, stratified random, quota, snowball, opportunity — probability techniques need a sampling frame.
Ethics: informed consent, deception, confidentiality, harm, right to withdraw (BSA ethical guidelines).
PET/E rule: every method evaluation must name a practical constraint, an ethical issue or a theoretical position — bare lists earn little credit.
Positioning research: data types, positivism and interpretivism
AQA 4.1.3.1-3, OCR H580/02 and Eduqas AC2.1/AC2.2 expect you to classify every method along three distinctions before evaluating it. Quantitative vs qualitative: quantitative methods (questionnaires, structured interviews, official statistics, content analysis) produce numbers suitable for measuring patterns and testing hypotheses; qualitative methods (unstructured interviews, participant observation, personal documents) produce words and observation suitable for understanding meanings and motives.
- Primary vs secondary: primary data is collected first-hand by the sociologist (experiments, questionnaires, interviews, observation); secondary data already exists (official statistics, company reports, diaries, media content) — cheap and accessible, but not designed for the researcher's purpose.
- Positivism: associated with Comte and Durkheim — society is an external reality of social facts that constrains individuals, so it can be studied objectively to discover cause-and-effect laws; positivists therefore favour methods producing reliable and representative data, above all official statistics and questionnaires. Durkheim's study of suicide (1897) is the classic model: official statistics on suicide rates correlated with levels of social integration.
- Interpretivism: associated with Weber and symbolic interactionism — social reality is constructed through meanings, so sociology needs valid and authentic data through empathy (verstehen); interpretivists favour unstructured interviews and participant observation. Atkinson's study of coroners' verdicts (1978) shows even suicide statistics rest on the meanings coroners give to a death.
- Synthesis: the A* point is methodological pluralism or triangulation — combining methods to balance reliability and validity, as Eileen Barker did in The Making of a Moonie (1984), using questionnaires, interviews and observation over several years.
Why it matters: 10-mark methods questions usually ask for two advantages or problems of a method; the marks come from development, so each point must be explained in the method's own terms — e.g. questionnaires are reliable because they are standardised, so every respondent answers identical questions and results can be replicated.
The main methods and their practical, ethical and theoretical strengths and limits
AQA 4.1.3.1 expects the strengths and limitations of each main method, evaluated through practical, ethical and theoretical (PET) criteria. Learn the methods in pairs of opposites and attach one classic study to each.
| Method | Key strengths | Key limitations |
|---|---|---|
| Lab experiments | Reliable, replicable, controlled variables | Artificial; Hawthorne effect; small samples |
| Questionnaires | Large, cheap, representative; reliable | Low response rates; imposition problem; shallow |
| Structured interviews | Reliable; quick; trained interviewers | Interviewer bias still possible; little depth |
| Unstructured interviews | Valid, flexible, in-depth rapport | Unrepresentative; costly; hard to replicate |
| Participant observation | Valid, naturalistic, access to hidden worlds | Unrepresentative; going native; ethics if covert |
| Official statistics | Free, large-scale, trends over time | Socially constructed; dark figure; definitions change |
| Documents | Authentic meanings; personal insight | Unrepresentative; credibility problems |
Experiments — classic studies
Lab experiments manipulate variables under controlled conditions: Milgram's obedience experiments (1963) demonstrated the power of authority, but were criticised for deception and psychological harm — a stock AO3 example. Field experiments occur in natural settings: Rosenhan's pseudo-patient study (1973) showed how a psychiatric label overwhelms behaviour, though it was ethically heavy on deception and never repeated. Both illustrate the reliability-validity trade-off: control raises reliability but artificiality cuts validity.
Interviews — classic studies
Ann Oakley (1981) rejected the detached, formal interview in favour of unstructured interviews with expectant mothers, building rapport that produced far richer data on women's experiences of childbirth — the standard evaluative evidence that interviewing style shapes validity. Dobash and Dobash (1980) used unstructured interviews with women in refuges to expose the reality of domestic violence, a sensitive topic where trust is essential.
Observation — classic studies
Participant observation ranges from overt (openly known) to covert (undercover). Humphreys' Tearoom Trade (1970) observed men in public lavatories without consent and later identified them through car number plates — the classic ethical critique of covert research, though it produced data no other method could access. Venkatesh's Gang Leader for a Day (2008) shows how observation accesses hidden subcultures, but also its risks: he witnessed violence and had to negotiate his role with a gang leader. Observation produces high validity but low reliability and representativeness.
Secondary sources — classic studies
Official statistics are cheap, cover whole populations and allow trends over time (e.g. birth rates since 1900), but they are socially constructed: Farrington and Dowds (1986) showed rising recorded crime in one Lancashire area partly reflected police recording practices, not more crime. Victim surveys such as the Crime Survey for England and Wales (developed by Mayhew and colleagues) expose the dark figure of unreported crime. Documents range from public records to personal diaries: Hey (1997) used girls' friendship notes in The Company She Keeps to access authentic meanings, but such sources are unrepresentative and credibility is hard to verify.
The research process: aims, operationalisation, sampling and analysis
AQA 4.1.3.4 lists the main stages of the research process — learn them in order and be ready to explain what can go wrong at each stage.
- Choosing a topic: driven by theoretical interest, practical feasibility, funding bodies and ethical acceptability — sensitive or harmful topics (e.g. child abuse) attract ethical scrutiny from the outset.
- Formulating aims and hypotheses: an aim states the purpose of the study; a hypothesis is a testable prediction, used mainly in quantitative, deductive research. Interpretivists often work with looser research questions instead.
- Operationalising concepts: turning abstract ideas into measurable indicators — e.g. measuring religiosity through attendance, or class through occupational category. Poor operationalisation invalidates findings: if attendance does not capture belief, the study measures the wrong thing.
- Pilot study: a small trial run to check question clarity, timing and response rates before the main study — cheap insurance against wasted fieldwork.
- Sampling: define the population, obtain a sampling frame (a list of the population), then select. Random and systematic sampling need a full frame; stratified random sampling divides the frame into strata (class, gender, ethnicity) and samples proportionally, maximising representativeness; quota sampling fills quotas of respondent types — faster, but interviewer choice biases the sample; snowball and opportunity sampling suit hard-to-reach groups (offenders, drug users) but are unrepresentative.
- Selecting and collecting data: apply the chosen method; response rates decide whether a survey is truly representative.
- Analysing and interpreting data: quantitative data is analysed for patterns, trends and correlations — remember correlation is not causation; qualitative data is coded into themes.
- Reporting and evaluating the research: findings are published and judged on reliability, validity, representativeness and ethics — the same PET criteria that shaped the design.
Examiner tip: 10-mark questions frequently target one stage (e.g. two problems of using quota sampling, or two reasons for using a pilot study). Answer in the stage's own terms with one developed point per reason.
Choosing topic and method: theoretical, practical and ethical considerations
AQA 4.1.3.5 requires you to weigh the theoretical, practical and ethical considerations influencing the choice of topic, method and conduct of research — the PET/E framework that underpins every methods evaluation.
- Theoretical: the positivism-interpretivism divide decides the method: positivists want reliable, representative data (questionnaires, statistics) to test causal laws; interpretivists want valid, authentic data (interviews, observation) to capture meanings. The validity-reliability trade-off means no method maximises both. Value freedom matters here too: Weber argued values inevitably choose the topic but research itself should be objective, while Becker (1967) asked 'whose side are we on?' — arguing sociology should side with the underdog — and Gouldner criticised value-free sociology as a myth that hides the researcher's own assumptions.
- Practical: time and money limit sample size and method choice; access is decisive — schools and prisons require gatekeepers' permission, and criminal subcultures require special negotiation (Venkatesh 2008); the researcher's own characteristics (age, gender, ethnicity) affect rapport; response rates decide whether a large survey is genuinely representative.
- Ethical: informed consent (participants must know they are being studied — parents consent for pupils), deception (covert observation involves it by definition), confidentiality and anonymity (Humphreys 1970 broke both), protection from harm (Milgram's 1963 participants suffered stress), and the right to withdraw. The British Sociological Association's ethical guidelines codify these duties.
The evaluation dilemma: ethical and practical constraints often cut against validity. Covert observation produces the most authentic data on hidden behaviour but breaches consent; large-scale surveys are ethical and representative but shallow. The strongest answers name this trade-off explicitly rather than treating ethics as a bolt-on list.
Exam technique: methods questions, command words and AO balance
On AQA 7192, pure methods appear in the 10-mark 'outline and analyse two...' questions on Papers 1 and 3, and feed the 20-mark Methods in Context question (AQA 4.1.2) and the 30-mark theory-and-methods essays (AQA 4.4.5, e.g. value freedom). OCR assesses methods in H580/02 Section A, where every method point must be applied to researching social inequalities; Eduqas examines quantitative methods (AC2.1) and qualitative methods (AC2.2) separately.
| Command word | What it demands | AO focus |
|---|---|---|
| Outline and analyse two | Two developed points, each explained in the method's own terms | AO1 + AO2 |
| Assess the usefulness of | Sustained weighing of strengths against limits, with a judgement | AO3-heavy |
| Explain what is meant by | Precise definition with an example | AO1 |
30-mark skeleton (assess the usefulness of X for researching Y): (1) define X and its logic; (2) two or three strengths, each named as practical, ethical or theoretical, each with a study; (3) two or three limitations, likewise; (4) a conclusion weighing usefulness for Y specifically — never a vague 'both have merits'.
Common pitfalls: listing strengths without developing them; forgetting to attach the PET dimension to each point; confusing reliability (consistency) with validity (truth); ignoring the positivism-interpretivism pairing examiners reward; omitting named studies — adjectives alone never earn evaluation marks.
Timing: 10 marks ≈ 10 minutes; 30 marks ≈ 30 minutes (plan 4, write 24, check 2). Remember that OCR item-based methods questions reward tight application to the given scenario — quote the item's wording in your answer.
💡 Worked Example
Exam-style: 1. Assess the usefulness of participant observation for investigating crime and deviance. [30 marks]
2. Outline and analyse two advantages of using official statistics in sociological research. [10 marks]
Mark-scheme logic: Q1 (30 marks): AO1 — accurate knowledge of participant observation (overt/covert, gaining trust, verstehen) plus the positivism-interpretivism context. AO2 — apply to crime: access to hidden subcultures and motives no survey reaches (Venkatesh 2008; Winlow 2001), but danger, time and gatekeeper problems in researching criminals. AO3 — strengths: high validity (Humphreys 1970 accessed behaviour impossible to study otherwise); limitations: unrepresentative, unreliable, ethical costs of covert work, researcher effects (going native); balance against positivist alternatives (official statistics and the CSEW expose only part of the dark figure). Judgement: useful for meanings and hidden worlds, inadequate alone for extent and distribution. Q2 (10 marks): two developed advantages with application only — e.g. large-scale and representative, allowing generalisation and trends over time (birth rates since 1900); free and quick to access, enabling comparisons between groups; reliable because standardised, so replicable (Durkheim's suicide study 1897). Development only — no extended evaluation required.
❓ Practice Questions & Mark Schemes
Outline and analyse two advantages of using official statistics in sociological research. [10 marks]
Answer: Mark-scheme logic: two developed points with application. (1) Scale and representativeness: statistics often cover whole populations (births, deaths, crime), so findings can be generalised and trends traced over time — e.g. Durkheim (1897) correlated suicide rates across societies. (2) Reliability and comparability: standardised collection makes statistics replicable and allows comparisons between groups and periods, which is why positivists favour them (Durkheim's social facts). Accept also: free and easily accessed; no Hawthorne effect; useful for detecting changes that trigger further research. Each point developed to 5 marks.
Assess the usefulness of unstructured interviews for investigating domestic violence. [30 marks]
Answer: AO1: unstructured interviews — informal, no fixed questions, rapport, probing, flexibility. AO2: apply to domestic violence — a sensitive, hidden topic where trust decides whether victims disclose; Dobash and Dobash (1980) used unstructured interviews with women in refuges to reveal patterns of male control. AO3 strengths: high validity, victims can explain meanings in their own words, interviewer can adapt to distress (Oakley 1981). AO3 limitations: practical (time-consuming, costly, small unrepresentative samples); ethical (distress, consent, confidentiality, risk if the abuser learns of the interview); theoretical (unreliable to replicate, interviewer bias, interviewees may forget or censor). Conclusion must judge usefulness for this topic specifically: excellent for meanings and motives, weak for measuring the extent of violence.
Outline and analyse two reasons why covert participant observation may be ethically problematic. [10 marks]
Answer: Mark-scheme logic: two developed points. (1) No informed consent: participants are studied without knowing it, denying them the choice to refuse — Humphreys (1970) observed men in tearooms without consent and later identified them through number plates. (2) Deception and harm: the researcher pretends to be something they are not, and exposure could damage participants' reputations, jobs or safety — Humphreys' deception meant men risked public identification as homosexual in 1970. Accept also: no right to withdraw; breach of confidentiality. Each developed to 5 marks.
Assess the claim that sociological research can never be value-free. [30 marks]
Answer: AO1: value freedom — Weber: values choose the topic but research should be objective; the positivists' detached ideal; Becker's 'whose side are we on?'; Gouldner's critique of value-free sociology. AO2: apply to method choice (covert research, topic selection, funding) and interpretation of findings. AO3: for the claim — topic choice is value-laden, funding shapes agendas, interpretation reflects the researcher's framework (Becker 1967; Gouldner); operationalisation embeds values (what counts as 'underachievement' is a judgement). Against — researchers can minimise bias through standardisation, replication and peer review; a wholesale relativist conclusion ('everything is subjective') would make sociology impossible. Conclusion must reach a weighed judgement: values are inevitable at the margins, but research can approach objectivity.
Outline and analyse two reasons why a stratified random sample may improve representativeness. [10 marks]
Answer: Mark-scheme logic: two developed points. (1) Proportional strata: dividing the sampling frame by key characteristics (class, gender, ethnicity) and sampling each in proportion means every subgroup is represented in its correct proportion, reducing sampling error compared with a simple random sample. (2) Random selection within strata: because individuals are chosen randomly from each stratum, researcher bias is removed and the sample can be generalised to the whole population. Accept also: guards against over-representing the majority group; where a full sampling frame exists, using it is itself a strength. Each developed to 5 marks.
Assess the usefulness of participant observation for investigating crime and deviance. [30 marks]
Answer: AO1: participant observation — overt and covert, gaining trust, experiencing the group's world, verstehen. AO2: apply to crime — access to hidden worlds and motives no survey reaches: Venkatesh (2008) joined a Chicago gang; Winlow (2001) gained access to Sunderland doormen's criminal economy; Humphreys (1970) observed behaviour invisible to other methods. AO3 strengths: high validity, naturalistic, reveals motives and subcultural meanings. AO3 limitations: practical (danger, time, access negotiated with gatekeepers); ethical (covert work breaches consent — Humphreys 1970); theoretical (unrepresentative, unreliable, risk of going native, researcher effect changes behaviour). Balance: official statistics and victim surveys measure extent but miss meanings and hide the dark figure. Conclusion must judge usefulness for crime specifically: the best method for meanings, weak for distribution.