Breaking down the question
The question has two clear demands. First, it asks you to define sampling and locate it within the wider logic of social research — why researchers rarely study whole populations and what a sample is meant to achieve. Second, it asks for a discussion of different forms of sampling, each weighed by its relative advantages and disadvantages.
The verb discuss signals that a bare list will not suffice. The examiner wants you to compare the two great families — probability and non-probability sampling — and to show judgement about when each is appropriate. The reference to social research is a cue to keep examples sociological rather than purely statistical.
A strong answer therefore rests on three pillars:
- A crisp conceptual definition anchored in the ideas of population, sampling frame and representativeness.
- A structured typology of sampling methods with honest trade-offs.
- An evaluative close that connects choice of sample to the aims of a study — whether it seeks generalisation or depth of meaning.
For the underlying vocabulary of variables, populations and representativeness, see the note on variables, sampling, reliability and validity.
How to approach it
Open with a tight definition and the rationale for sampling — economy, feasibility and, paradoxically, often greater accuracy than a poorly conducted census. Then split the answer into probability and non-probability sampling, treating the major types within each. For every method, give one line on its logic, one on its strength and one on its weakness, so that the advantages and disadvantages demand is met visibly.
Sequence the methods from simple to complex within each family. Weave in short sociological illustrations — a survey of caste attitudes, a study of pavement dwellers, snowball recruitment among a hidden group — so the discussion breathes.
Close by resisting the temptation to crown one method as best. The examiner rewards the recognition that sampling choice is problem-driven: quantitative, generalising research leans on probability designs, while exploratory or qualitative work often prefers purposive selection.
Model answer
Sampling is the process of selecting a subset of units — individuals, households, groups or events — from a larger population so that conclusions drawn from the subset can be extended, with a known margin of confidence, to the whole. As Goode and Hatt observed, the sample is a miniature of the population; its scientific value lies in representativeness.
Researchers sample for compelling reasons. Studying an entire population is usually costly, time-consuming and often impossible — one cannot interview every voter or every migrant worker. A well-drawn sample yields results faster, cheaper and, when field control is tighter, sometimes more accurately than an unwieldy census. The key preconditions are a clearly defined population and, ideally, a sampling frame — a list of all units from which the sample is drawn.
Sampling methods fall into two broad families.
Probability sampling gives every unit a known, non-zero chance of selection, allowing calculation of sampling error and statistical generalisation.
- Simple random sampling selects units by lottery or random numbers. Its advantage is freedom from bias and firm statistical grounding; its disadvantage is the need for a complete frame and the risk that a scattered sample becomes expensive to reach.
- Systematic sampling picks every nth unit after a random start. It is simple and quick, but a hidden periodicity in the list can distort the sample.
- Stratified sampling divides the population into homogeneous strata — say by caste, income or region — and samples within each. It guarantees representation of small but important groups and improves precision, yet demands prior knowledge of the stratifying variable.
- Cluster and multi-stage sampling select natural groups such as villages or wards, then sample within them. This slashes cost and needs no full frame, but clustering raises sampling error because units within a cluster tend to resemble one another.
Non-probability sampling selects units without random assignment; probabilities are unknown, so statistical generalisation is weakened but other virtues emerge.
- Convenience sampling takes whoever is available. It is cheap and fast — useful for a pilot study — but highly prone to bias.
- Purposive or judgement sampling selects units the researcher deems typical or information-rich. It suits qualitative and case-study work, though it depends heavily on the researcher's judgement and can smuggle in bias.
- Quota sampling fills fixed quotas for chosen categories, mimicking stratification without randomness. It is economical and widely used in market and opinion research, yet the final choice of respondents remains subjective.
- Snowball sampling builds the sample through referrals, invaluable for reaching hidden populations such as sex workers, drug users or undocumented migrants. Its weakness is that it circulates within a network and cannot claim representativeness.
The two families embody a deeper methodological divide. Probability designs serve the positivist aim of measurement and generalisation; non-probability designs serve the interpretive aim of understanding meaning in depth. The reliability of a survey and the validity of a case study rest on getting this choice right.
In practice, researchers often combine approaches — a stratified sample for a national survey, purposive selection for follow-up interviews. The soundness of any social finding, from voting behaviour to attitudes towards untouchability, ultimately turns on the defensibility of its sample.
Examiner's perspective
Examiners look first for conceptual clarity: candidates who blur population, sample and sampling frame lose ground early. The single most common failing is presenting sampling as a statistics topic divorced from sociology — memorised definitions with no field example.
The phrase relative advantages and disadvantages is doing real work in the question. A script that lists methods but omits their trade-offs answers only half the demand. The best answers pair each method with a crisp merit and limitation, and, crucially, connect sampling choice to research purpose.
High-scoring responses show methodological maturity — the recognition that no single method is universally superior, and that probability and non-probability designs reflect competing epistemologies. A closing line on mixed-methods practice signals exactly the judgement the examiner rewards.