DataTerminology.comNEW
A range computed from sample data that likely contains a population parameter at a stated confidence level.
DESCRIPTION
A confidence interval quantifies uncertainty around an estimated population parameter, such as a mean or proportion. It is constructed using sample statistics and an assumed sampling distribution, yielding a lower and upper bound. The confidence level (e.g., 95%) reflects long-run coverage properties across repeated samples, not the probability that one realized interval contains the parameter. Wider intervals indicate greater uncertainty. They are essential for reporting effect sizes and assessing practical significance alongside p-values.
ALTERNATIVE DEFINITIONS
Wikipedia
In statistics, a confidence interval is a type of interval estimate, computed from the statistics of the observed data, that might contain the true value of an unknown population parameter.
Clarifies that the interval is derived from sample statistics and targets an unknown parameter.
NIST Engineering Statistics Handbook
A confidence interval for a parameter is an interval computed from sample data by a method that, in repeated sampling, has a specified probability of producing an interval containing the parameter.
Emphasizes the frequentist long-run coverage interpretation.
Penn State Eberly College of Science
A confidence interval is a range of values that is likely to contain a population parameter with a certain level of confidence.
Provides a concise, instructional description highlighting likelihood and population parameter.
RELATED TERMS
Hypothesis Testing p-value Standard Error Margin of Error Bootstrap Normal Distribution A/B Testing
COMPETING TERMS
Credible Interval Prediction Interval
KNOWLEDGE AREAS
Mathematics Data Science Testing & Validation
TARGET AUDIENCE
Data Scientist Data Analyst Information Analyst BI Specialist
EXTERNAL SOURCES
- Wikipedia - Confidence interval
- NIST Engineering Statistics Handbook - Authoritative CI definition and methods
- Penn State - Introductory confidence intervals lesson
- Khan Academy - Confidence intervals videos and practice
TOP QUESTIONS
What does a 95% Confidence Interval mean?
It means that the method will produce intervals containing the true parameter in 95% of repeated samples, not that a specific interval has 95% probability.
Category: Misunderstanding
How is a Confidence Interval constructed?
Using an estimate (e.g., sample mean), its standard error, and a quantile from a sampling distribution (e.g., t or z) to form lower and upper bounds.
Category: Explanation
What affects Confidence Interval width?
Sample size, variability, and chosen confidence level; larger samples and lower confidence levels produce narrower intervals.
Category: Explanation
How do Confidence and Credible Intervals differ?
Confidence intervals have frequentist long-run coverage; credible intervals represent Bayesian posterior probability about parameter values.
Category: Alternatives
Can I use Confidence Intervals in A/B testing?
Yes. They provide uncertainty bounds on lifts or differences and complement hypothesis tests and p-values.
Category: Application
What is a common mistake with Confidence Intervals?
Interpreting the interval as a probability statement about the parameter for the realized data.
Category: Mistakes
How to compute a Confidence Interval without normality?
Use bootstrapping or distribution-free methods when assumptions are violated or sample sizes are small.
Category: Guidance
Are two non-overlapping Confidence Intervals equivalent to significance?
Non-overlap suggests significance, but overlap rules are conservative; formal tests or intervals on differences are preferred.
Category: Testing
Why report Confidence Intervals instead of only p-values?
Intervals convey effect sizes and uncertainty, improving interpretability and decision-making.
Category: Relevancy