Sampling Error
Reviewed byMallik Gaddam
Definition
Sampling error is the difference between a statistic calculated from a sample and the true parameter of the population from which the sample was drawn.[2][1] It arises because only a portion of the population is studied, and even a well designed random sample will not perfectly replicate the characteristics of the entire population. Sampling error decreases as sample size increases and is quantified by the standard error and confidence intervals.[1][2]
Explore the concept
Scatter around the truth
Repeat a 400 nurse survey many times in your head, change how many nurses each survey asks, and compare the random scatter that shrinks with a bias that does not.
30 imagined repeats of the nurse burnout survey
Margin of error
± 4.8
points, 95%
400 nurses: the surveyed example
The big idea: Sampling error is the random gap between a sample estimate and the population value that comes from studying only part of the population. It is measured by the standard error and confidence interval and shrinks as the sample grows (four times the sample, half the error), while bias moves every estimate the same way and does not shrink.
The 400 nurse survey (42%, 95% CI 37% to 47%, ± 4.8 points) and the 1,600 nurse comparison (about 40% to 44%, ± 2.4) are the article's; ± 9.7 for 100 nurses uses the same formula, 1.96 × √(0.42 × 0.58 ÷ n). The dots are 30 illustrative repeat surveys placed at normal quantiles around 42%, which is treated as the true value for the picture. The 6 point bias from nurses with burnout responding less often is illustrative.
Why it matters
Sampling error is an inherent feature of all sample based research and sets a floor on the precision of any estimate derived from a sample. Understanding sampling error helps researchers interpret confidence intervals, design studies with adequate sample sizes, and communicate uncertainty in their findings. It also helps readers evaluate the reliability of published estimates: a large sampling error (wide confidence interval) indicates that the estimate could be substantially different from the true population value, while a small sampling error (narrow confidence interval) indicates a more precise estimate.
How it works
Sampling error occurs because each sample is one of many possible samples that could have been drawn from the population. Different samples yield different estimates due to the random variation inherent in the selection process. The sampling distribution of a statistic (such as the mean) describes how that statistic would vary across all possible samples of the same size drawn from the same population.[1][2]
The standard error quantifies the expected magnitude of sampling error for a given statistic. For the sample mean, the standard error equals the population standard deviation divided by the square root of the sample size; in practice it is estimated with the sample standard deviation.[1][2] This formula reveals two key relationships: sampling error increases with greater population variability and decreases with larger sample sizes. Quadrupling the sample size halves the standard error, demonstrating the diminishing returns of increasing sample size.
Confidence intervals use the standard error to construct a range of values by a method that, across repeated samples, captures the true population parameter at a specified rate (typically 95%).[1] A wider confidence interval reflects greater sampling error and less precise estimation.
Sampling error should be distinguished from systematic errors (biases), which arise from flaws in the study design, measurement, or execution and do not diminish with larger samples.[2] A biased study with a large sample size will produce a precise but inaccurate estimate.
When to use it
Consider sampling error when interpreting any estimate derived from a sample, including means, proportions, regression coefficients, and odds ratios. Report confidence intervals or standard errors alongside point estimates to convey the precision of the estimates. Account for sampling error in power analyses and sample size calculations. When comparing results across studies, consider that differences in findings may be attributable to sampling error rather than genuine differences in the populations or treatments studied.
Key takeaways
- Sampling error is an inherent feature of all sample based research and sets a floor on the precision of any estimate derived from a sample.
- Different samples yield different estimates due to the random variation inherent in the selection process.
- Consider sampling error when interpreting any estimate derived from a sample, including means, proportions, regression coefficients, and odds ratios.
- A common error is confusing sampling error with bias.
Common mistakes
- A common error is confusing sampling error with bias. Sampling error is random and decreases with larger samples; bias is systematic and persists regardless of sample size.
- Another mistake is interpreting a confidence interval as a range within which the true value must lie with certainty, rather than as the result of a method that, across repeated samples, captures the true value in a specified proportion of them.[1]
- Researchers sometimes focus on statistical significance (whether the confidence interval excludes a null value) while ignoring the width of the interval, which reflects the precision of the estimate.
- Finally, neglecting to report confidence intervals alongside point estimates deprives readers of essential information about the reliability of the findings.
Example
A survey of 400 registered nurses estimates the prevalence of burnout at 42% (95% CI: 37% to 47%). The margin of error of about plus or minus 5 percentage points (4.8 before rounding) reflects the sampling error expected with a sample of this size. If the survey had enrolled 1,600 nurses instead, the same prevalence would have a 95% CI of approximately 40% to 44%, with a margin of error of about plus or minus 2.4 percentage points, cutting the sampling error in half.
Test your understanding
5 questions that adapt as you go: answer two in a row correctly and they get harder.
Question 1 of 5
Scenario:
Two surveys of equal size draw random samples from two populations. In one population the outcome varies much more between individuals than in the other.
Question:
Which survey has the larger sampling error for its mean?
Hardness: Interpret
Sources
- [1]Field A. Discovering Statistics Using IBM SPSS Statistics. 6th ed. SAGE Publications, 2024. ISBN 9781529630008
- [2]Cochran WG. Sampling Techniques. 3rd ed. Wiley, 1977. ISBN 9780471162407
Cite this term
Paperguide. (2026, October 16). Sampling Error. In Paperguide Research Glossary. https://paperguide.ai/research-glossary/sampling-error/(Paperguide, 2026)