A frequentist 95% confidence method captures the fixed population parameter in about 95% of intervals over repeated sampling under its assumptions. Read the interval's direction and width, then check the design and population before generalizing.
Worked example
An effect interval from -1 to 4 units includes zero, so it is compatible with no difference at that level. It does not prove the effect is zero or that 95% of observations fall inside the interval.
Invented Experiment Summary
A randomized study compares two paper designs using a standardized test. The estimated mean difference, B minus A, is 5 load units. Its 95% confidence interval is 2 to 8 load units. Assume the interval method's conditions hold. Only one paper material and one humidity range were tested.