Sampling methods
Sampling Methods
Research rarely studies every member of the target population — surveying every UK homeowner aged 35–55 would be prohibitively expensive and time-consuming. Instead, researchers select a sample — a subset of the population — and use the findings from that sample to draw conclusions about the population as a whole. The method used to select the sample has a direct impact on how representative it is and therefore how reliable the research conclusions are. Five sampling methods are required in the IB specification: quota, random, stratified, cluster, and convenience.
Quota sampling
In quota sampling, the researcher defines how many respondents are needed from specific subgroups (or quotas) — for example, 200 men and 200 women, or 150 respondents from each of three age groups. Interviewers then fill these quotas by selecting respondents who meet the criteria, but with no requirement for random selection within each quota. The researcher has control over the composition of the sample, ensuring that key subgroups are adequately represented. However, because selection within each quota is not random, quota sampling can produce biased results — interviewers may unconsciously favour more approachable or accessible respondents, even within the defined quota.
Random sampling
In random sampling (also called probability sampling), every member of the target population has an equal and known chance of being selected. Respondents are chosen using a random process — for example, using a computer-generated list of random numbers to select names from a customer database. Random sampling is the most statistically rigorous method because it eliminates selection bias, enabling the findings to be generalised to the broader population with a calculable margin of error. Its limitations are practical: it requires a complete sampling frame (a list of every member of the population, which may not exist), and it can be expensive if selected respondents are geographically dispersed.
Stratified sampling
In stratified sampling, the population is first divided into distinct subgroups (strata) based on a relevant characteristic — age, income, geography, or purchasing frequency. A random sample is then drawn from each stratum in proportion to its representation in the population. For example, if 40% of the target population is aged 35–44, then 40% of the sample is randomly selected from that age group. Stratified sampling combines the representativeness of random sampling with the assurance that all important subgroups are adequately included — it tends to produce more reliable results than simple random sampling because it reduces sampling variance within each stratum.
Cluster sampling
In cluster sampling, the population is divided into clusters (typically geographic areas) and a random sample of clusters is selected. All members within the chosen clusters are then surveyed. For example, to survey UK retail customers, a researcher might randomly select ten postal districts and then survey all relevant customers within those districts. Cluster sampling is significantly cheaper and more operationally feasible than national random sampling because fieldwork is concentrated in specific areas. The limitation is reduced precision: if the selected clusters are not representative of the wider population, the findings will be biased.
Convenience sampling
Convenience sampling involves selecting respondents who are most easily accessible — for example, interviewing shoppers at a particular location, surveying students in a class, or using an online panel of volunteers. It is the simplest and cheapest method but produces the least representative sample, because the selection is determined by accessibility rather than any systematic attempt to reflect the population. Convenience sampling is appropriate for exploratory or qualitative research where statistical representativeness is not the goal, but it is inappropriate for quantitative research intended to produce generalisable findings.
| Method | Selection basis | Key advantage | Key limitation | Best suited to |
|---|---|---|---|---|
| Quota | Non-random selection within defined subgroup quotas | Ensures subgroup representation; relatively fast and cheap | Selection within quotas not random; interviewer bias possible | Quantitative research where key subgroups must be represented |
| Random | Every member of population has equal chance of selection | Eliminates selection bias; findings statistically generalisable | Requires complete sampling frame; expensive if dispersed | Large-scale surveys requiring statistical precision |
| Stratified | Random selection within predefined population subgroups | Representative of all subgroups; more precise than simple random | Requires knowledge of population characteristics to define strata | Surveys of diverse populations where subgroup differences matter |
| Cluster | Random selection of geographic clusters; all within clusters surveyed | Cheap and operationally feasible for geographically dispersed populations | Clusters may not be representative; within-cluster homogeneity increases error | Large-scale national surveys with limited budget |
| Convenience | Whoever is most accessible to the researcher | Very cheap and fast; no sampling frame required | Highly unrepresentative; findings cannot be generalised | Exploratory qualitative research; pilot testing survey instruments |
For the post-launch satisfaction survey of smart camera owners, Wrenfield's research team selected a stratified sampling approach. The customer database was divided into strata by purchase channel (direct online, major electronics retailer, specialist security dealer) and by household income band. A proportionate random sample was drawn from each stratum, ensuring that customers who purchased through each channel were represented in the same proportions as the overall customer base. This approach was chosen over simple random sampling because Wrenfield's marketing director suspected that satisfaction might vary significantly by purchase channel — customers who received expert advice from a specialist dealer might have different satisfaction profiles than those who bought online without guidance. Stratification enabled meaningful comparison between these subgroups, which a simple random sample of the same size might have missed if one channel was disproportionately represented by chance.
Key Takeaways
- A sample is a subset of the target population selected for research; the sampling method determines how representative it is of the whole population.
- Random sampling is the most statistically rigorous; it requires a complete sampling frame and gives every member an equal chance of selection.
- Stratified sampling improves on random sampling by ensuring all key subgroups are proportionately represented.
- Quota sampling controls subgroup composition without random selection within quotas — cheaper but subject to interviewer bias.
- Cluster sampling concentrates fieldwork geographically, reducing cost but potentially reducing representativeness.
- Convenience sampling is the cheapest and least representative — appropriate for exploratory research, not for generalisable quantitative findings.