Sampling: probability and non-probability methods
Sampling means studying a small, chosen part of a population to learn about the whole. In probability sampling every member has a known chance of selection.
Methods with merits and demerits
| Method | How it works | Merits | Demerits |
|---|---|---|---|
| Simple random (probability) | Every unit has an equal chance; lottery or random number table | Free from selection bias; results can be generalised | Needs a full list; sample may be scattered; small groups may be missed |
| Systematic (probability) | Every kth unit from a list after a random start (k = N ÷ n) | Quick and easy; spreads evenly over the list | Biased if the list has a repeating pattern |
| Stratified random (probability) | Divide into strata (e.g. year of study), then draw randomly from each | Every subgroup is represented; more precise | Needs information on strata; more work |
| Cluster (probability) | Randomly pick whole groups (e.g. schools) and study everyone in them | Cheap for spread-out populations; no list of individuals needed | Less precise; people within a cluster tend to be alike |
| Multi-stage (probability) | Sample in stages: district → schools → classes → pupils | Practical for large surveys | Complex; error adds up at each stage |
| Convenience (non-probability) | Take whoever is easily available | Quick and cheap | Highly biased; cannot generalise |
| Purposive or judgement (non-probability) | Researcher picks units that suit the purpose | Reaches the right people, e.g. hospital dietitians | Depends on the researcher's judgement; bias |
| Quota (non-probability) | Fill a fixed number from each group, chosen by convenience | Fast; every group appears | Choice within each quota is biased |
| Snowball (non-probability) | Each respondent refers the next | Reaches hidden or rare groups | Respondents tend to be alike; cannot generalise |
Probability vs non-probability
Authors differ: Kothari calls non-probability sampling as a whole 'deliberate' or 'purposive' sampling, and counts systematic sampling as random. Ranjit Kumar calls systematic sampling a 'mixed' design, partly random and partly not.
Steps in drawing a sample
Drawing a systematic sample
A hostel register lists 600 students. You need a sample of 60 for a breakfast survey. Draw a systematic sample.
- Sampling interval k = N ÷ n = 600 ÷ 60 = 10
- Pick a random start between 1 and 10 by lottery: say 7
- Take every 10th name after it: 7, 17, 27, 37 …
- Last number = 7 + (59 × 10) = 597
60 students: register numbers 7, 17, 27 … 597.
NFHS uses a stratified, two-stage sample to estimate anaemia across India. A student studying one college can use stratified random sampling by year of study, or purposive sampling when only diabetic patients or dietitians are needed.
2 marks: “What is a sampling frame?” A sampling frame is the complete list of units in the population from which the sample is drawn, e.g. the hostel register.
5 marks: “Distinguish between probability and non-probability sampling.” Compare chance of selection, bias, generalisation, cost and examples (simple random, stratified, cluster vs convenience, purposive, quota, snowball).
10 marks: “Explain the probability and non-probability methods of sampling with their merits and demerits.” Define population, sample and frame; draw a chart of methods; explain each with merits and demerits; add a nutrition example.
Probability: "Some Students Skip Classes" = Simple random, Systematic, Stratified, Cluster. Non-probability: "Can People Quit Sugar?" = Convenience, Purposive, Quota, Snowball.