A sample is a smaller portion (or subset) of a population.
Samples save time and money. They are also efficient at providing timely data.
It would be very time consuming and expensive to ask every person in the United States information on their labor force status each month. Instead, BLS asks a subset of the population (about 60,000 households) about their labor force status each month. The households that participate in the survey participate for a set number of months and then are rotated out of the sample. At the same time a new set of households enter into the sample. This rotation allows for breadth of response and lowers sampling error (see the Error measurement topic page).
BLS samples are carefully chosen. Depending on what the survey is trying to measure will influence how the sample is chosen. For example, household surveys are generally chosen such that the racial and ethnic make-up of the sample is similar to that of the nation. Samples should mimic the population value that they are trying to measure. When the characteristics (industry make-up, demographic characteristics) of the sample are similar to those of the population, there is greater confidence that the measures produced by the sample are similar to the population
For more information on how an individual survey’s sample is chosen, please view the Collection and Design sections of the Handbook of Methods.