Non-Random (Non-Probability) Sampling Methods

In non-random (or non-probability) sampling, the chance of any particular unit being selected is not known in advance, and selection depends on the convenience or judgement of the investigator rather than on a defined random procedure. These methods are quicker and cheaper but carry a higher risk of selection bias, and the sampling error cannot be measured mathematically.

1. Convenience Sampling — units are selected simply because they are easy to reach (e.g., interviewing shoppers standing near a store entrance). Quick and low-cost, but the sample may not represent the population at all.
2. Judgement (Purposive) Sampling — the investigator deliberately selects units believed to be typical or expert representatives of the population (e.g., choosing five well-established textile traders to assess the general health of the textile trade). Useful when expert insight matters more than strict statistical representativeness.
3. Quota Sampling — resembles stratified sampling in that the population is divided into groups (quotas) by characteristics such as age or income, and investigators are told how many units to collect from each group — but within each quota, the actual units are chosen by convenience or judgement, not randomly. This within-group selection is the key difference from stratified random sampling.
4. Snowball Sampling — used for populations that are rare or hard to locate directly (e.g., a study of a small, specialised group of exporters); a few initial contacts are surveyed and asked to refer further contacts, and the sample 'snowballs' outward.