What is it? Sampling has different meanings in quantitative and qualitative research. In quant, it refers to the people or data that is researched or sampled. The sample is a subset of the whole intended to be extrapolated to represent the whole. Characteristics of the subset should be representative of the whole population so any inferences or extrapolations ensure that the findings can be generalized. 

In contrast, qualitative research does not attempt to derive representative samples. It seeks to include people or situations within a research study that will prove the most relevant, given the nature of the research question. The sample is the people interviewed, or observed for a study, who are selected to be broadly representative of the target. That said, qualitative samples are not statistically scalable/generalizable to a larger population.

When is it best used? Samples can be used to model or benchmark a larger population or dataset in market research, user research and other sciences. Sampling is best used when studying the full dataset, or conducting a survey of the entire population, is not feasible or necessary. Advantages of sampling over measuring the whole include lower cost, faster data collection and streamlined analysis.

What does it entail? Sampling, similar to recruitment, starts with defining the ideal data set or participant profile based on the target audience and objective of the study. Demographic, behavioral, attitudinal, psychographic, etc. characteristics are often combined to define the subset to be used as the model. 

Researchers typically work with analysts to collect or pull samples, or they make selections from a participant pool / sample panel. Sometimes they analyze the subset of data that exists, and other times they gather additional input in the form of surveys/questionnaires, polls, and other quantitative methods. This smaller dataset can then be extrapolated to the broader population or dataset. 

Interchangeable terms: Subset, sub-section

Use in a sentence: The researchers surveyed a representative sample of the population because polling the entire  population would take too much time and be too costly.

Related Terms: Simple random sampling, systematic sampling, stratified sampling, probability proportional to size sampling, cluster or multistage sampling, convenience sampling, quota sampling and purposive sampling, survey., sample size, representative sample. 

Visual: No

Additional Resources

Articles

• How many test users in a usability study? By Jakob Nielsen

Share this page