How Do I Calculate the Margin of Error?
The Margin of Error is a guide to how precisely a survey can measure a particular metric. It gives an estimate of how much the metric might change if you repeated the survey with a different random sample but the same process.
For simple metrics (such as the percentage of customers who responded "Very Satisfied" on a survey question), the margin of error is usually calculated with the simple formula:
Margin of Error in percentage points = 100 / sqrt(number of responses)
So for example, a survey with 100 responses will have a margin of error of 10 percentage points, and a survey with 400 responses will have a margin of error of 5 percentage points. This is close to the 95% confidence interval, meaning that if you repeated the survey 100 times you would expect 95% of the time the result would be the same to within the margin of error.
This rule of thumb for margin of error is generally close enough for business purposes--realistically we just want a rough idea of how precise our survey is--but it's less accurate when the metric is close to 0% or 100%, or when the total population of customers is close to the number who responded to the survey. In those cases the actual 95% confidence interval will be smaller.
Margin of error only measures one thing: how similar the survey result would be if you repeated exactly the same survey. It does not take into account other sources of inaccuracy or bias in the survey process, like self-selection, poorly worded questions, or intentional efforts to manipulate the survey.