Research Methods in Sociology
The logic of inquiry
Understand the research process: problem, literature review, hypothesis, design, data collection, analysis, conclusion. Distinguish deductive (theory to data) from inductive (data to theory) reasoning, and quantitative from qualitative approaches. Know validity (does it measure what it claims?) and reliability (would it give consistent results?), and the tension between them.
Methods
Know the strengths and limits of surveys, experiments, ethnography/participant observation, interviews (structured, semi-structured, unstructured), case studies and documentary/secondary analysis. Understand sampling: random, systematic, stratified, cluster, quota, purposive and snowball. Distinguish probability from non-probability sampling, and know that generalisability depends on representativeness.
Ethics
Informed consent, confidentiality, anonymity, no harm, and voluntary participation. Note the special difficulties of covert observation.
Traps
Do not confuse reliability with validity. Do not claim that qualitative work is unscientific; it uses different criteria (credibility, transferability). Remember that a hypothesis is tested, not proved. Note that correlation is not causation, and that spurious relationships arise from third variables.
Worked example
A researcher wants to study drop-out in rural Bangladeshi schools. A survey of a random sample gives generalisability but limited depth; ethnography of one district gives depth but not representativeness. A mixed-methods design combining both exploits the strengths of each - this is triangulation.
Rule of thumb
Match method to question: what, how many, how much suggest quantitative designs; why, how, in what context suggest qualitative ones.