What is Margin of Error?
Margin of error is a measure of sampling uncertainty that shows how much a survey result (usually a percentage) might differ from the true population value — commonly expressed as ±X% at a given confidence level (often 95%).
The margin of error (MoE) quantifies the expected random sampling variation in survey estimates. When you survey a sample of employees rather than the whole organisation, the percentage you observe (for example, 62% satisfied) is an estimate of the true population percentage. The MoE gives a range around that estimate (for example, 62% ± 6.9%) where the true value is likely to fall, assuming the sample is representative and the stated confidence level (usually 95%). MoE depends mainly on sample size and the proportion measured (the most conservative case is around 50%), and is larger for small samples. It does not account for systematic errors such as biased sampling, low response quality, or poorly worded questions. For small populations (when your sample is a substantial fraction of the total), a finite population correction reduces the MoE.
Usage example
You run a pulse survey in a 1,000-person organisation and receive 200 responses. For a 95% confidence level and the conservative 50% proportion, the margin of error is about 6.9%. If 62% of respondents say they are satisfied, you would report that satisfaction is 62% ± 6.9% (approximately 55.1% to 68.9%). This means a 4% difference between two groups would not be statistically meaningful because it falls inside the margin of error.
Practical application
Margin of error matters because it helps workplace teams and consultants interpret results correctly and plan surveys appropriately. Use MoE to: - Decide whether observed differences between teams, sites or time periods are likely real or just sampling noise. - Estimate the sample size needed to achieve a desired precision (for example, ~385 responses gives about ±5% at 95% confidence in the conservative case). - Choose when to pool data across pulses or groups to increase precision for small teams. - Balance the cost/effort of additional responses against the benefit of smaller uncertainty. Remember to consider MoE together with response representativeness and potential bias — a low MoE is not useful if responses come from an unrepresentative subgroup.
FAQ
Does margin of error measure survey bias?
No. Margin of error measures random sampling variability only. It assumes your sample is representative. Systematic issues — like nonresponse bias, self-selection, framing effects or coverage errors — are not captured by MoE and can still make results misleading.
How can I reduce the margin of error for a workplace survey?
The most direct way is to increase your sample size (more responses). Other approaches include pooling responses from multiple pulses, stratified sampling (to ensure representation of key groups), and boosting participation among underrepresented groups. Note that improving representativeness is as important as increasing sample size.
What sample sizes give common margins of error at 95% confidence?
Using the conservative 50% proportion: ~100 responses → ≈ ±9.8%; ~200 → ≈ ±6.9%; ~385 → ≈ ±5.0%; ~400 → ≈ ±4.9%; ~1,000 → ≈ ±3.1%. These are useful planning benchmarks for workplace surveys.
My organisation is small — does margin of error change?
Yes. When your sample is a substantial share of the total population, the margin of error decreases because you’ve covered much of the population; this is accounted for by a finite population correction. In very small organisations, fewer responses can still give reasonable precision compared with the population, but always check whether respondents are representative of the whole group.