Random sampling
Random sampling is a way of selecting participants in which every member of the population has a known, usually equal, chance of being chosen. It is the foundation of probability sampling, and what allows a sample's results to be generalized to the population.
How it works
In simple random sampling, every member of the sampling frame has the same chance of selection, typically by drawing random numbers from a complete list. Other probability methods build on it: stratified sampling draws randomly within subgroups, cluster sampling draws whole groups at random, and systematic sampling takes every nth member from a random start.
Randomness removes the researcher's choices from selection, which is what makes the statistics work: confidence intervals and margins of error assume a random sample. In practice the sample is only as good as the frame it is drawn from, and nonresponse can undo randomness, so researchers check the achieved sample against the population.
A worked example
An airline wants to survey its 480,000 loyalty members. It assigns each a number and draws 1,500 at random. The sample's mix of membership tiers matches the full list within 1 point on each tier, which gives the team confidence that results reflect members as a whole.
Related
Automate this in Pulse
A random draw from a customer database can be uploaded to Pulse as the invitation list, and comparing who completed with who was drawn shows where nonresponse has bent the sample.