Glossary of Statistical Terms: Definition Margin of error
The margin of error is the maximum deviation of the sample results from the real values. If there is a large margin of error in a survey, there will be less confidence that it matches the ‘true’ figures for the total population. In many surveys which are referred to as representative, the margin of error is between three and five percent. Therefore at the end of the poll, narrow results give no clear information about the actual state of the target group. With a margin of error of three percent, a PRO CONTRA 48 to 52 percent will have the same probability as PRO 51 (+3%) to CONTRA 49 percent (-3%).
The margin of error is calculated from the size of the sample, as well as the underlying confidence level and the proportion involved (percentage of respondents who have selected the characteristic under consideration). Projections over the course of an evening after an election can show the power of an opinion poll as well as its error limits. In a sample survey, a result can already be claimed by 6 pm, which is usually a few tenths of a percent from the final result. Nevertheless, in the course of many election evenings, there can be deviations of plus or minus two percent between the first prognosis and the official result. This difference shows the deviation of the sample result (referring to voters outside the polling stations) from the actual result of the population (in this case all active voters). The maximum extent of this deviation can be mathematically predicted using the margin of error.
Please note that the definitions in our statistics encyclopedia are simplified explanations of terms. Our goal is to make the definitions accessible for a broad audience; thus it is possible that some definitions do not adhere entirely to scientific standards.