What is Confidence Interval?
A confidence interval is a range of values, calculated from survey data, that likely contains the true value for the population measure you care about (for example, the true percentage who support a policy). It expresses the uncertainty around a point estimate using a chosen confidence level (commonly 95%).
When you measure something in a sample — say the percent of respondents who found a service helpful — that sample result is only an estimate of the true value for the whole population. A confidence interval gives a lower and upper bound around the estimate which reflects the uncertainty due to using a sample instead of surveying everyone. The width of the interval depends mainly on the sample size, how much responses vary, and how confident you want to be (e.g., 90% vs 95%). Note: the standard interpretation is that if you repeated the same sampling process many times, the chosen percentage of those intervals (e.g., 95%) would contain the true population value.
Usage example
You run a parent survey and find 60% of respondents support a proposed change. With a 95% confidence interval of 55% to 65%, report: 60% (95% CI: 55%–65%).
This communicates that, given the sample, the true support rate for the whole parent population is plausibly between 55% and 65%.
Practical application
Confidence intervals help teams avoid over-interpreting noisy survey results. They show how precise an estimate is, guide decisions about whether differences between groups (for example, respondents answering in different languages) are likely real, and inform how large a sample you need to reach a desired precision. In multilingual surveys, small language-specific sample sizes often produce wide intervals — a useful signal that you need either more responses or to be cautious drawing conclusions about that subgroup.
FAQ
Does a 95% confidence interval mean there's a 95% chance the true value lies inside it?
Not exactly. For a fixed interval from a single survey, you can't assign a probability to the true value. The correct idea is: if you repeated the same survey process many times and built a 95% interval each time, about 95% of those intervals would contain the true population value.
What makes a confidence interval wider or narrower?
Three main factors: smaller sample sizes produce wider intervals; more variability in responses produces wider intervals; and choosing a higher confidence level (e.g., 99% vs 95%) makes the interval wider. Increasing your sample size is the most direct way to narrow the interval.
Can I use confidence intervals for percentages and for averages?
Yes. Confidence intervals apply to many kinds of estimates, including proportions (percentages) and means (averages). The calculations differ slightly, but the idea — showing a plausible range for the true value — is the same.
Are confidence intervals reliable if my sample isn’t random or representative?
They become less reliable. Confidence intervals quantify sampling uncertainty, assuming the survey sample reasonably represents the target population. If the sample is biased (for example, if some language groups are under-represented), the interval may not capture the true population value and you should interpret results with caution.