ABSTRACT
Every 10 years the United States government is required to conduct a census of the nation’s population. A census attempts to count every person in the population. Although the Census Bureau has used sampling techniques for studies since 1937,1 the decennial census is a complete count. For the 2000 Census, the Census Bureau proposed to interview a sample of the population-rather than carrying out a census-to improve the accuracy of the count and to reduce expenses. Whether you study groups of 100, 1,000, or 1,000,000, drawing a sample is an economical and effective way to learn about their individual members. This applies if your data come from individual respondents, case records, agencies, or computerized datasets. Understanding the principles of sampling would have helped most observers appreciate the Census Bureau’s proposal; however, the proposal was controversial and was ultimately rejected by Congress.2
