How to read a statistic in a news headline

Read a statistic in a headline as a claim you can test, not as a fact to accept or reject on sight. Four things determine what it means: the base the number is measured against, the comparison it implies, the period it covers and the wording that describes the change. Headlines are short by design, so they often leave out at least one of these. The missing detail is usually in the article beneath the headline.

You can check all four in under a minute once you know what to look for. The routine is practical rather than mathematical. You do not need to recalculate every figure, only to ask a few questions that expose the most common gaps.

Find the base behind the number

A percentage always has a base, the total it is a share of. A headline that says a share rose sharply tells you little until you know that base. If the group was small, a large percentage may represent only a handful of people. If the group was large, a modest percentage can still mean a great many. Phrases such as "one in five" or "half of people" are shorthand for a base, and the base is what the reader needs to see.

Check whether the base changed between the two numbers being compared. A rate that rises because the group it is measured against shrank tells a different story from one that rises because more events happened. When the article gives counts alongside percentages, the counts often tell you more than the percentage does.

Averages need the same scrutiny. A headline that reports an average may be describing a figure pulled upward by a few very large values, so the typical case looks different. If the article also gives a median, or a range, compare it with the headline figure before accepting the description of a typical result.

Check what the number is compared with

A statistic is meaningful only in relation to something. Identify the comparison: last year, another group, a target or an average across all cases. The headline may not state it, but the article usually does. If it does not, the number may be a level with no comparison at all. That can still be informative, but it cannot by itself support a claim of change.

Watch for the difference between relative and absolute change. Suppose a figure moves from 2 cases per 10,000 people to 4 cases per 10,000. The relative change is a doubling, or a 100 percent increase. The absolute change is 2 cases per 10,000, which is 0.02 percentage points. Both statements are arithmetically correct, but they create very different impressions. A headline that uses only the relative figure often sounds far bigger than the underlying change.

Look at the time frame

Time frames are easy to miss and easy to manipulate. A figure described as a rise since a particular month depends entirely on which month was picked as the start. A statement that something hit a record may refer to a short run of data, or to a series that began only recently. Check whether the period is a single month, a quarter, a year or an unspecified span, and whether both sides of the comparison cover the same length of time.

Seasonal effects belong here too. A comparison between a winter month and a summer month can differ for reasons unrelated to any real change. When a claim compares two different times of year, ask whether the same comparison would look different if it used the same season in both years.

Notice the wording

Small words change what a statistic claims. "Up" and "could rise" are not the same claim. "Linked to" and "associated with" describe an association, while verbs such as "causes", "leads to", "boosts" or "cuts" claim a cause, which only some kinds of study can support. "Nearly", "almost", "more than" and "about" each shift the figure by a little, and phrases such as "experts say" are only useful when the source is named and can be checked.

Check whether percent and percentage points are used correctly. A rise from 30 percent to 33 percent is an increase of 3 percentage points and a relative increase of 10 percent. Headlines sometimes use the two interchangeably, which can make a small absolute change sound like a large relative one. When the wording is loose, the underlying numbers are worth finding.

Adjectives carry judgments of their own. Words such as "record", "surge" or "plunge" tell you how to feel about a figure before you have checked it. Compare the word with the arithmetic. The movement behind a surge may be a small change on a large base, and the number alone may describe something far more modest than the headline suggests.

Go to the original when it matters

When a headline cites a survey, study or official release, go to the original if the claim matters to you. Look for the population measured, the dates of data collection, the definitions used and any statement of uncertainty. A news story is a summary, and a summary can drop the caveats the original report took care to include.

You do not need to do this for every headline. Save the full check for figures you plan to repeat, act on or share. For quick reading, the four questions about base, comparison, time frame and wording are usually enough to tell whether a statistic is solid, exaggerated or missing essential context. That habit costs little and guards against most of the common distortions.