Margin of Error

---

What is Margin of Error?

Margin of error is a statistical measurement that allows a market research company to gauge how accurate survey results are. The figure shows how much the results from a random sample survey can differ from the actual population. The larger the margin of error, the less confidence one should have in the results of a survey.

Margin of error is displayed as a plus or minus percentage figure. This figure can be used to calculate a range of possible percentages you can expect when comparing your survey results to real-world data. For example, if 48% of surveyed consumers plan to purchase Product A (over product B), this figure may actually range from 43% to 53% as a result of the margin of error. This assumes a margin of error of +/- 5%. The range is calculated by adding 5% to 48% and subtracting 5% from 48%.

In the industry, the margin of error is typically calculated at the 95% confidence level. This means that if a survey was conducted 100 times, the survey results would be within the calculated margin of error 95 times out of 100.

How Many Survey Completes do I Need?

As a starting point, we typically consider 400 completes to be a balance between the costs of surveying individuals and reliability of results. With this number of completes, the margin of error is just below our best-practice and industry standard of +/-5% margin of error, providing you with representative survey results of the population.

The number of survey completes needed is ultimately reliant on how you plan to use the data. For example, if you plan to focus on subsets of respondents (i.e. 25 to 34-year-olds) during analysis or you plan to make the findings available for public consumption, minimizing margin of error will be important.

However, if you’re using the survey as a starting point to identify large differences in consumer preference, or you are more interested in the qualitative findings from the survey (i.e. the open-ended responses) then you may be okay with surveying a fewer number of individuals.

Conducting surveys and finding respondents can often be costly, so it also worth noting that there are diminishing returns with margin of error as sample size increases. For the first few hundred survey responses, the margin of error drops significantly. However, with each additional survey complete, there is less and less of a decrease in margin of error. There is a point of diminishing returns and that is largely concluded to be 400 responses.

The chart below illustrates this point. Note, the margin of error in this chart assumes a large population size. The margin of error calculator above can be used to determine how many survey completes you will need to obtain a +/-5% margin of error with a smaller population.

Margin of Error by Number of Survey Responses