The same change, two headlines
Imagine a fictional survey in which the share of adults reporting a particular symptom rises from 4 in 100 to 6 in 100. “Reports rose by 50%” sounds dramatic. “Reports rose by two percentage points” sounds quieter. Neither tells the whole story alone.
The relative calculation is (6 minus 4) divided by 4, or 0.5. The percentage-point calculation is simply 6 minus 4. A reader who hears “a 50% increase” might mistakenly picture 50 additional people in every 100. Here, the difference is two.
| Starting share | Later share | Percentage-point change | Relative change |
|---|---|---|---|
| 4% | 6% | +2 | +50% |
| 40% | 42% | +2 | +5% |
| 40% | 60% | +20 | +50% |
The table is arithmetic, not evidence about any particular symptom. It shows why the starting value belongs near the change.
When the population also changes
Suppose the first survey included 100 people and the second included 1,000. Four reports followed by 60 reports would be a 15-fold change in the count, yet the shares would still be 4% and 6%. Counts, percentages, and percentage-point differences cannot be exchanged without changing the meaning.
A useful description includes the population, time period, and measurement: “In these two surveys, the share reporting the symptom increased from 4% to 6%.” Whether that reflects a real population change also depends on sampling and uncertainty.
A sentence that keeps both views
“The share increased by two percentage points, from 4% to 6%, a 50% relative increase” is longer than either headline. It also gives readers enough information to judge scale for themselves. The CDC's discussion of ratios and proportions supplies the underlying distinction between a part of a whole and a comparison between quantities.
