What Is Margin of Error?
How many percentage points a survey's results could differ from the true value in the full population, at a given confidence level.
When a poll reports “62% approve, ±3 points,” the margin of error is that ±3. It means the true figure across the whole population is likely somewhere between 59% and 65%, not exactly 62% — surveys measure a sample, not everyone, so there’s always some uncertainty band around the result.
Margin of error shrinks as sample size grows, but not in a straight line — cutting the margin of error in half requires roughly quadrupling the sample size, which is why very tight margins (±1%) get expensive fast while going from ±5% to ±3% is a more modest jump.
Margin of error is always reported alongside a confidence level (commonly 95%) — the two travel together. “±3 points at 95% confidence” means: if you ran this survey 100 times, about 95 of those times the true value would fall within 3 points of your result.
Example
- Lowest likely score
- 42 − 4=38
- Highest likely score
- 42 + 4=46
The true score is most likely between 38 and 46, not exactly 42.
What it does not tell you
- It covers sampling error only. Leading wording, a skewed sample and people who never answered are not part of it. See survey bias.
- It applies to a random sample. A sample of whoever chose to respond has no true margin of error.
- A reported ±3 is the widest case, for a share near 50%. A share closer to 0% or 100% has a narrower range, and a subgroup has a wider one.
Sources
- Cochran, W. G. (1977). Sampling Techniques (3rd ed.). John Wiley & Sons.
- Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R. (2009). Survey Methodology (2nd ed.). John Wiley & Sons.
How to Calculate Margin of Error for SurveyRock
Use SurveyRock's free sample size calculator to see how margin of error changes as you adjust sample size and confidence level, before you decide how many responses to collect.