Margin of error

A measure of sampling uncertainty, commonly the half-width of a confidence interval around a sample estimate. It describes how far the estimate may lie from the population value under stated assumptions.

Confidence interval: A range of values constructed by a statistical procedure to capture an unknown population parameter at a stated confidence level. The margin of error is the distance from the estimate to either endpoint of a symmetric confidence interval.
Point estimate: A single sample-based value used to estimate an unknown population parameter. The margin is reported around this central estimate.
Opinion poll: A survey designed to measure public attitudes, preferences, or intentions. Poll results commonly pair an estimated percentage with a margin of error.
Credible interval: A Bayesian interval containing a specified posterior probability of a parameter, given a model and observed data. Unlike a frequentist confidence interval, it supports a direct probability statement about the parameter under the posterior.
Bootstrap (statistics): A resampling method that estimates a statistic's sampling distribution by repeatedly sampling with replacement from observed data. Bootstrap intervals offer alternatives when standard formulas poorly approximate the sampling distribution.
Standard error: The standard deviation of a statistic's sampling distribution, measuring how much estimates vary across repeated samples. The margin is often calculated by multiplying a standard error by a critical value.
Population parameter: A numerical characteristic of an entire population, such as its mean or proportion. The sample estimate and its margin are intended to locate this unknown quantity.
Proportion confidence interval: An interval estimate for a population proportion based on sample data. It is the common setting for interpreting margins reported for survey percentages.
Prediction interval: An interval estimating where a future individual observation is likely to fall under a statistical model. It covers variation in future observations, not merely uncertainty about a population estimate.
Cluster sampling: A probability sampling design that selects groups of units and then observes units within selected groups. Within-cluster similarity can make conventional margins too small if the design is ignored.

Linked from 9 pages