Toxicology basics · Concept 7 of 9 · 3 min

Reading the odds

Risk is a chance with a base. Absolute and relative risk, how to turn percentages into people, and why a big percentage can be a small change.

A number about chance needs three things: how many, out of how many, over what time (my example).

The idea

People understand natural frequencies, like "3 in 100", better than percentages or probabilities Akl and colleagues, 2011. Relative changes look bigger and more persuasive than absolute ones Akl and colleagues, 2011. Studies call this chance risk: absolute risk is the count in one group, and relative risk compares two groups (my example). "50% higher" can mean 2 in 1,000 became 3 in 1,000 (my example). Incidence is the number of new cases in a group over a set time (my example). A link in a study isn't the same as a cause (my example).

Reading the oddsA grid of 1,000 small squares, one per person. Two are gold: the people who get the illness without the habit. One more is white: the extra case with the habit. Two in 1,000 becomes three in 1,000, which is one extra per 1,000, or 50 percent higher.1,000 people2 without the habit1 more with it"50%higher"means1 morein 1,000
A thousand people as dots, with the extra case marked, next to the same change written as a percentage (my example).

Use it on a case

When you read a claim with a number:

  1. Turn a relative risk into an absolute one: how many out of how many.
  2. Ask for the starting point: higher than what?
  3. Ask about time: per year, or over a lifetime?
  4. Ask who was counted: people like you, or a different group?
  5. Ask whether it's a link or a cause.

An example

A report says a habit raises the chance of an illness by 50% (my example). If 2 in 1,000 people get it without the habit, 3 in 1,000 get it with the habit (my example). That's one extra case per 1,000 people (my example). Worth knowing, and smaller than "50%" first sounds (my example).

What it can't tell you

A group's odds don't say what will happen to one person in it (my example).

Connects to

Sources for this page

  • Akl and colleagues (Cochrane) (2011). Using alternative statistical formats for presenting risks and risk reductions.

    Systematic review, 35 studies: pulls together every study it could find on one question.

    What we use: People understood natural frequencies ('3 in 100') better than percentages or probabilities; relative changes looked bigger and more persuasive than absolute ones.

    Checked Oct 11, 2026 · https://doi.org/10.1002/14651858.CD006776.pub2 · DOI 10.1002/14651858.CD006776.pub2 · PubMed 21412897

Every source on the site