Sampling
A sample is useful when the way people are selected allows researchers to make defensible inferences about a larger population.
← AFDES Civic Academy
Public Opinion · Interactive Simulation
How Public Opinion Polling Works
Experiment with sample size, underlying opinion, and systematic bias to see how researchers estimate what a larger population thinks.
Open the polling lab ↓
Interactive Polling Lab
Change the research conditions.
This simplified simulator separates three ideas: the population’s underlying opinion, random sampling uncertainty, and systematic bias.
Estimated support
52.0%
Approx. 95% margin of error: ±3.1 pts
Population52% support / 48% other
0255075100%
Poll estimate Population value
Margin of error
The familiar margin of error summarizes random sampling uncertainty under assumptions. It does not capture every possible source of survey error.
Weighting & bias
Researchers often weight samples to better reflect the target population, but weighting cannot automatically repair every coverage, nonresponse, or measurement problem.
Interpret Carefully
A precise-looking number is still an estimate.
Poll quality depends on question wording, sampling design, fieldwork, response patterns, weighting, timing, and the population being measured. A large sample can produce a very precise estimate of the wrong thing if the design is systematically biased.
Sources & Further Reading
Learn the standards behind survey research.
AAPORStandards, ethics, and survey research guidance→
Roper Center for Public Opinion ResearchPublic opinion archives and methodology resources→
U.S. Census BureauSurvey research and response resources→
The simulator uses a simplified normal-approximation margin-of-error calculation for educational purposes. Real survey estimates can require design effects, finite-population considerations, weighting adjustments, and other methodological treatment.