03

Small Numbers,
Big Uncertainty

A percentage can look precise even when the evidence behind it isn’t.

Survey findings are estimates, not exact measurements of a population. As we divide a sample into smaller groups, uncertainty generally increases—even though the result may still appear as a reassuringly precise percentage.
 

What this demonstrates
Survey uncertainty • Confidence intervals • Subgroup analysis • Responsible interpretation

​

Data: 2025 Greater Boston Jewish Community Study
Created by Daniel Parmer, PhD

About this experiment:
Good analysis communicates not only what the best estimate is, but also how much confidence the data can support. The goal isn’t to avoid using uncertain estimates. It’s to make uncertainty visible so that a 28% finding doesn’t automatically imply that we know the answer is exactly 28%. Choose a measure and an age group to see how the estimate and its range of plausible values change.

​

Want to see another experiment?