What is the standard deviation of the sample mean?+
The standard deviation of the sample mean is the standard error (SE), which measures how much sample means vary across repeated samples. SE = sigma divided by sqrt(n). It describes the precision of the sample mean as an estimate of the population mean. A smaller SE means the sample mean is more reliable. SE is always smaller than the population SD for any sample size greater than 1.
What is the difference between standard deviation and standard error?+
Standard deviation (sigma or s) measures the spread of individual data values around their mean. Standard error (SE) measures the spread of sample means across repeated samples. SD is a property of the population or dataset and does not depend on sample size. SE equals SD divided by sqrt(n) and decreases as sample size increases. Reporting SE instead of SD makes results appear more precise than they are, so always label which one you are using.
Why does the standard error decrease as sample size increases?+
SE = sigma divided by sqrt(n). As n grows, sqrt(n) grows, making SE smaller. This reflects the Law of Large Numbers: with more data, the sample mean becomes a more accurate and stable estimate of the population mean. The square-root relationship means improvements in precision are subject to diminishing returns. Going from n = 25 to n = 100 halves SE; going from n = 100 to n = 400 halves it again.
How do I calculate the margin of error from standard error?+
Margin of error = z times SE, where z depends on the confidence level: 1.645 for 90%, 1.960 for 95%, and 2.576 for 99%. For example, if SE = 3 and you want a 95% confidence interval, MOE = 1.960 times 3 = 5.88. The confidence interval around the sample mean is then: sample mean minus 5.88 to sample mean plus 5.88. A narrower MOE requires either a larger sample or accepting a lower confidence level.
What sample size do I need for a given margin of error?+
Rearrange MOE = z times sigma divided by sqrt(n) to get n = (z times sigma divided by E) squared, then round up. For example, with sigma = 10, E = 2, and 95% confidence (z = 1.960): n = (1.960 times 10 divided by 2) squared = 9.8 squared = 96.04, so n = 97. Always round up, never round down, so the actual MOE stays at or below the target. Use the Find Sample Size mode for instant results.
Can I use sample standard deviation instead of population standard deviation?+
Yes. When the population SD sigma is unknown (the usual case in practice) substitute the sample standard deviation s. The estimated SE = s divided by sqrt(n). For small samples (n less than 30) use the t-distribution with n minus 1 degrees of freedom when constructing confidence intervals or running hypothesis tests. For large samples (n at least 30) the t and z distributions are nearly identical and the normal approximation is acceptable.
How does the Central Limit Theorem justify using SE?+
The Central Limit Theorem (CLT) states that the distribution of sample means approaches a normal distribution as sample size grows, regardless of the shape of the underlying population. The mean of that sampling distribution equals the population mean, and its standard deviation equals SE = sigma divided by sqrt(n). The CLT is why confidence intervals and z-tests based on SE are valid even when the raw data is not normally distributed, provided n is large enough (typically at least 30).
What does "SE as a percentage of sigma" mean?+
SE as a percentage of sigma equals (SE divided by sigma) times 100, which equals 100 divided by sqrt(n). This tells you what fraction of the population SD has been reduced by averaging. A sample of n = 100 gives SE = 10% of sigma. A sample of n = 400 gives SE = 5% of sigma. This metric is useful for quickly comparing how much precision different sample sizes deliver relative to the underlying variability in the data.
How is SE used in hypothesis testing?+
In a one-sample z-test or t-test about a mean, the test statistic is z (or t) = (sample mean minus hypothesised mean) divided by SE. If this ratio is large in absolute value, the sample mean is far from the hypothesised value relative to the sampling noise, and you reject the null hypothesis. The p-value is then read from the normal or t-distribution. SE is the denominator of every test statistic involving a mean, making it central to inferential statistics.
Is SE the same as the standard deviation of a sampling distribution?+
Yes, exactly. When you imagine drawing many random samples of size n from a population and computing the mean of each, those means form a sampling distribution. The standard deviation of that sampling distribution is the standard error: SE = sigma divided by sqrt(n). SE is not a property of any single sample; it is a property of the process of sampling. This is why SE quantifies precision rather than variability of raw data.
Why do I need to round up when calculating minimum sample size?+
The formula n = (z times sigma divided by E) squared usually produces a non-integer. Rounding down would give a sample size slightly below the minimum needed, causing the actual margin of error to exceed the target E. Rounding up guarantees that the actual MOE is at or below E. For example, if the formula gives 96.04, using n = 96 gives MOE just above the target; using n = 97 delivers MOE at or below it. Always use ceiling (round up) for sample size calculations.
What are common mistakes when reporting standard error?+
The most frequent mistakes are: (1) labelling SE as SD or vice versa, which misrepresents precision; (2) using SE for descriptive statistics when SD is more appropriate; (3) forgetting to specify the confidence level when reporting a margin of error; (4) applying z-critical values when a t-distribution should be used for small samples; and (5) reporting SE from a sample with unknown population SD without noting that it is an estimate. Always state clearly whether you are reporting SE or SD and the sample size used.