How can I tell if a statistic is being used misleadingly?
Answer
Answered 18 October 2026 · Updated 18 October 2026
A statistic can be factually accurate yet misleading because of how it is collected, calculated, presented or interpreted. The key question is not just "Is this number true?" but "What does this number actually allow me to conclude?"
Five common ways statistics mislead
| Technique | What to watch for |
|---|---|
| Relative changes without absolute numbers | A 50% increase may mean a rise from 2 cases per 1,000 to 3 per 1,000. |
| Selective time periods | Choosing dates that make a trend look unusually good or bad. |
| Misleading averages | An average salary can rise even when most employees earn no more, if a few top earners get large pay rises. |
| Confusing correlation with causation | Two variables move together, but one may not cause the other. |
| False precision | A forecast of "37.4%" may look certain despite depending on uncertain assumptions. |
A quick test
A headline claims: "A study app improves exam scores by 20%."
The article states that students using the app scored an average of 60, compared with 50 among non-users. However, students chose for themselves whether to use the app, and the article does not report their scores before using it.
- What the data show: In this comparison, app users scored 20% higher on average.
- What the data suggest: The app might help.
- What the data do not prove: That the app caused the higher scores. Users may already have been more motivated or academically stronger.
A statistic is not necessarily misleading just because it is incomplete or uncertain. The key question is whether the strength of a conclusion matches the strength of the evidence.