Research literacy · uncertainty
Statistical uncertainty: why a result needs a range, not just a headline number
A point estimate can look exact while still carrying uncertainty. Confidence intervals and related measures help show how precise an estimate is—not whether a headline has become a guarantee.
Why a single number is incomplete
Research often reports a point estimate: a percentage, mean difference, rate, or ratio calculated from the observed sample. A point estimate is useful, but it is not the same as a precise description of every future population or event. Samples vary, measurements vary, and study choices affect how stable a result appears.
Confidence intervals provide one way to show that uncertainty. They display a range around an estimate under the method used. A wider interval generally indicates less precision than a narrower one, although width alone does not determine whether a study is well designed or useful.
Statistics Canada’s statistical-inference materials are a helpful reminder that estimates, sampling variation, and data quality need to be interpreted together. A statistic is evidence, not a prophecy.
Common errors to avoid
One common mistake is to treat a confidence interval as a guarantee that the true value is inside the displayed range. Another is to treat a non-significant result as proof that no relationship or difference exists. Both remove the conditional reasoning that statistical methods require.
Readers should also avoid assuming that a dramatic point estimate tells the whole story. The sample size, measurement quality, study design, follow-up, and missing data can all affect the interpretation. Confidence intervals are one part of the evidence picture, not a substitute for reading the methods.
A well-written page gives the figure, names the uncertainty, and links to the source. A weak page offers only the most convenient number, often rounded or stripped of its denominator, context, and range.
A practical precision check
When a source reports a number, ask: what was measured, in whom, over what time, and how precise was the estimate? If a confidence interval is reported, read it as a range that belongs to the specific analysis. If it is not reported, that absence should make a reader more cautious about claims of exactness.
The Institute for Work & Health offers a public explanation of confidence intervals that helps readers move from a vague sense of “margin of error” to a more useful understanding of uncertainty. That literacy is valuable whenever a website uses statistics to create an impression of certainty.
This page is not statistical consulting and does not interpret an individual study for a personal decision. It provides a reader’s checklist for keeping a number tied to its method and uncertainty.