How Samples are Selected in National Surveys

The first four links below describe the sampling designs of four major surveys. The first and most elementary discussion describes sample selection in the Gallup poll; the second describes the sampling procedure in the New York Times/CBS News Poll. Based on telephone surveys, both of these polls use random-digit dialing. The third and fourth links are more technical, describing complex multistage sampling designs for in-person surveys conducted by the National Opinion Research Center and the Bureau of the Census.

How Polls are Conducted (http://media.gallup.com/PDF/FAQ/HowArePolls.pdf)

An excerpt from the book Where America Stands, this essay provides an overview of how the Gallup Organization conducts public opinion telephone surveys.  Gallup poll editors describe the selection of a random sample and the selection of respondents within households.

A Sample of a Sample: How the “Typical” Respondent is Found (http://www.nytimes.com/library/national/110499poll-watch.html)

This brief non-technical article describes the random-digit dialing sampling procedure of the New York Times/CBS News Poll.

Sampling Design and Weighting in the General Social Survey (http://publicdata.norc.org:41000/gss/documents//BOOK/GSS_Codebook_AppendixA.pdf)

At the GSS Documentation Web site, click on “Appendix A.” The paper provides a detailed description, including the sampling frame and interviewer instructions, of the multi-stage, stratified random sampling design of the General Social Survey.

Sample Design of the Current Population Survey (https://www.bls.gov/opub/hom/pdf/homch1.pdf)

This link contains a description of the multi-stage sampling design of the Current Population Survey (CPS), a monthly survey of about 60,000 households conducted by the Bureau of the Census for the Bureau of Labor Statistics. The CPS is the primary source of information on the labor force characteristics of the U.S. population, including estimates of employment, unemployment, earnings, hours of work, and other indicators.

Research Randomizer (http://www.randomizer.org/)

This site contains a program that assists users in drawing a random sample. Lesson 1 in the Tutorial shows how to draw a random sample of 50 from a population of 643; Lesson 4 shows how to select a random sample of 100 telephone numbers from the same telephone exchange or prefix.

Determining Sample Size (http://edis.ifas.ufl.edu/pd006)

Though this document was prepared for agricultural researchers, it contains a good overview of the criteria that must be specified to determine the appropriate sample size and the strategies for determining sample size.

Sample Size Calculator (http://www.surveysystem.com/sscalc.htm)

You can use the calculator at this site to determine how many cases you need to select to get results with a given level of precision or to find the level of precision in an existing sample.