Research Methods and Methodology
Research methods and methodology
This domain examines how sociological knowledge is produced: the logic of research design, the strengths and limits of methods, and the ethical and theoretical choices behind them. Exam questions often ask you to choose a method for a topic and justify it against alternatives.
Key distinctions
- Methodology is the theory of method; positivism seeks causal, law-like explanation through measurement, while interpretivism seeks verstehen and meaning.
- Quantitative data (surveys, official statistics) support comparison and statistical generalisation; qualitative data (interviews, ethnography) support depth, meaning and process.
- Validity asks whether the method measures what it claims; reliability asks whether repetition would yield the same result; representativeness asks whether the sample reflects the population.
- Sampling: random, systematic, stratified, cluster, quota, snowball, opportunity. Probability samples allow inferential claims; non-probability samples do not.
- Ethics: informed consent, confidentiality, avoiding harm, right to withdraw, and dilemmas of covert research.
Worked example
Topic: why students truant. A postal questionnaire to all Year 11 students gives broad patterns but may under-represent truants (sampling bias) and invites socially desirable answers (low validity). Unstructured interviews with a snowball sample of truants give depth and trust but are not generalisable. A triangulated design - questionnaire plus attendance records plus interviews - improves validity and reliability. Reasoning: match method to the research question, acknowledge limitations, and justify the combination rather than claiming one method is universally best.
Traps
Confusing reliability with validity; calling a quota sample random; treating official statistics as objective facts without asking how they were produced; ignoring ethical issues in covert observation.
Key points
- Justify method by research question, not by habit.
- Define validity, reliability, representativeness and generalisability.
- Distinguish probability from non-probability sampling.
- Triangulation can offset single-method weaknesses.
- Official statistics are socially produced.