Sample

It is often impossible or too expensive to collect data from every member of a population, such as every customer of a bank or every company in an economy. Instead, researchers study a sample and use the results to estimate characteristics of the whole group.

For example, a bank with 2 million customers might survey 1,000 of them to estimate overall customer satisfaction. If 82% of the sample are satisfied, the bank estimates that roughly 82% of all customers are satisfied, with some margin of error.

The quality of conclusions depends heavily on how the sample is chosen. A random sample, where every member of the population has an equal chance of being selected, tends to be representative. Samples chosen in other ways, such as only surveying customers who visit a branch, may be biased and give misleading results.

Larger samples generally give more precise estimates, but a large biased sample can still be very wrong. In finance, samples are used in market research, auditing, economic statistics, credit risk models and investment research.