statistics
P-value, Standard Deviation and Standard Error: What They Mean and How Not to Misuse Them
Standard deviation (SD), standard error (SE) and the p-value appear in almost every medical paper, yet they are often confused. SD describes how spread out individual values are in your sample. SE describes how precisely your sample mean estimates the true population mean;
Resora Research Team · 14 Nov 2026 · 7 min read
Open any results table and you will see "mean ± SD", "mean ± SE" and "p = 0.03". They sound technical and similar, and many authors use them interchangeably — sometimes on purpose, because SE always looks smaller and "tidier" than SD.
They answer three different questions:
| Measure | Question it answers |
|---|---|
| Standard deviation | How much do individual patients differ from each other? |
| Standard error | How precisely does my sample estimate the true population value? |
| P-value | If there were no real difference, how surprising would my result be? |
Keep these three questions in mind and the rest of this blog falls into place.
Standard deviation: how spread out are the values?
The standard deviation (SD) is the typical distance of individual observations from the mean. A small SD means values cluster tightly around the mean; a large SD means they are widely scattered.
SD = \sqrt{\frac{\sum (x_i - \bar{x})^2}{n - 1}}We divide by n − 1 rather than n because the sample mean is itself estimated from the same data, and this corrects the slight underestimate that would otherwise result.
Worked example. Haemoglobin (g/dL) in 10 antenatal women:
9.8, 10.5, 11.2, 12.0, 10.9, 11.5, 9.5, 12.4, 11.0, 10.2
- Mean = 109.0 / 10 = 10.9 g/dL
- Sum of squared deviations from the mean = 7.74
- Variance = 7.74 / 9 = 0.86
- SD = √0.86 = 0.93 g/dL
Report this as 10.9 ± 0.93 g/dL (mean ± SD).
The normal-curve rule
When data are roughly normally distributed:
- about 68% of values lie within mean ± 1 SD
- about 95% lie within mean ± 2 SD (more exactly, 1.96 SD)
- about 99.7% lie within mean ± 3 SD
In our example, roughly 95% of similar women would have haemoglobin between 9.0 and 12.8 g/dL. This is how many laboratory reference ranges are built.
Key point: SD is a property of the population you are studying. Collecting more data gives a better estimate of the SD, but does not make it smaller. If the SD is close to or bigger than the mean for a measure that cannot be negative (hospital stay, CRP), the data are skewed — report the median and interquartile range instead.
Standard error: how precise is the estimate?
If you drew many samples of 10 women from the same population, each would give a slightly different mean. The standard error of the mean (SE or SEM) is the standard deviation of those sample means. It tells you how much your mean would vary from sample to sample — in other words, how precise it is.
SE = \frac{SD}{\sqrt{n}}For the haemoglobin example: SE = 0.93 / √10 = 0.29 g/dL.
SE shrinks as the sample grows
Because n sits under a square root, you need four times the sample to halve the SE. With an SD of 12 mmHg for systolic BP:
| Sample size | SE (mmHg) |
|---|---|
| 25 | 2.4 |
| 100 | 1.2 |
| 400 | 0.6 |
The spread of BP among patients (SD) stays around 12 whatever the sample size; only the precision of the mean (SE) improves.
From SE to the 95% confidence interval
The main use of SE is to build a 95% confidence interval (CI):
95\%\ CI = \bar{x} \pm t \times SE \quad (t \approx 1.96 \text{ for large samples})For our 10 women (small sample, so t = 2.26 with 9 degrees of freedom):
10.9 ± 2.26 × 0.29 = 10.2 to 11.6 g/dL
This means: if we repeated the study many times, about 95% of intervals built this way would contain the true mean haemoglobin of the population. Compare this narrow interval with the much wider range (9.0 to 12.8) that holds 95% of individual women — that difference is exactly the difference between SE and SD.
SE is not only for means. Proportions, differences between means, odds ratios and regression coefficients all have their own standard errors, and each is used the same way to build a CI or run a test.
SD vs SE: which one to report
| Standard deviation (SD) | Standard error (SE) | |
|---|---|---|
| Describes | Spread of individual values | Precision of the sample mean |
| Formula | √[Σ(x − x̄)² / (n − 1)] | SD / √n |
| Effect of larger sample | Stays about the same | Gets smaller |
| Used for | Describing data, reference ranges | Confidence intervals, hypothesis tests |
| Report as | Mean ± SD in baseline tables | Usually as a 95% CI, not ± SE |
| Example (Hb, n = 10) | 0.93 g/dL | 0.29 g/dL |
The rule: use SD to describe your participants (Table 1 in any paper) and 95% CI to describe your estimates (the results). Reporting mean ± SE in a descriptive table makes the data look far less variable than they are — a practice many journals and statistical reviewers specifically reject.
If a paper reports only SE and you need the SD (for a meta-analysis or a sample size calculation), recover it with SD = SE × √n.
The p-value: how surprising is the result?
Every hypothesis test starts with a null hypothesis (H₀) of no difference or no association. The p-value is the probability of getting a result at least as extreme as the one observed, assuming the null hypothesis is true.
A small p-value means your data would be unusual if there were truly no effect, so you reject H₀. By convention, p < 0.05 is called "statistically significant" — but 0.05 is an arbitrary line, not a law of nature.
Worked example
A trial compares systolic BP after treatment in two groups of 50 patients each:
| Group | n | Mean SBP (mmHg) | SD |
|---|---|---|---|
| New drug | 50 | 126 | 12 |
| Standard drug | 50 | 132 | 12 |
- Difference in means = 6 mmHg
- SE of the difference = √(12²/50 + 12²/50) = 2.4 mmHg
- t = 6 / 2.4 = 2.5, with 98 degrees of freedom
- p = 0.014; 95% CI for the difference = 1.2 to 10.8 mmHg
Notice that the test is simply the difference divided by its standard error. SE is the bridge between your data and the p-value.
The same difference, different p-values
| Patients per group | Mean difference | p-value |
|---|---|---|
| 15 | 6 mmHg | 0.18 |
| 50 | 6 mmHg | 0.014 |
The effect is identical; only the sample size changed. With 5,000 per group, even a 1 mmHg difference — clinically meaningless — gives p < 0.001. A p-value mixes effect size and sample size, so it cannot tell you how important a result is.
What the p-value is NOT
The American Statistical Association's 2016 statement on p-values set out these principles:
- It is not the probability that the null hypothesis is true.
- It is not the probability that the result happened by chance alone.
- It does not measure the size or clinical importance of an effect.
- p > 0.05 does not prove there is no difference — the study may simply be too small.
- Scientific conclusions and decisions should not rest on whether p crosses 0.05 alone.
Better practice: report the exact p-value (p = 0.014, not "p < 0.05") alongside the effect size and its 95% confidence interval, and judge clinical importance from the CI.
How the three connect
They form a chain:
- SD measures variation among individuals.
- SE = SD / √n turns that into the precision of your estimate.
- The test statistic = effect / SE, and the p-value comes from that statistic.
- The 95% CI = estimate ± about 2 SE carries the same information as the p-value, plus the size and direction of the effect. If a 95% CI for a difference excludes 0 (or for a ratio excludes 1), p < 0.05.
Common mistakes
- Reporting mean ± SE in descriptive tables to make data look less variable.
- Not saying which one the ± refers to. Always write "mean ± SD" or state it in the table footnote.
- Using mean ± SD for skewed data. Use median (IQR).
- Treating p < 0.05 as proof and p > 0.05 as proof of no effect.
- Reporting "p = 0.000" from software output. Write p < 0.001.
- Equating statistical with clinical significance.
- Leaving out confidence intervals, so readers cannot judge the size or precision of the effect.
- Judging overlap of SE error bars as a test of significance. Error bars that do not overlap are not a formal test, and ±1 SE bars can overlap even when p < 0.05 or fail to overlap when p > 0.05.
Conclusion
SD describes your patients, SE describes your precision, and the p-value describes how surprising your result would be if nothing were going on. None of them alone tells you whether a finding matters. Report SD to describe, confidence intervals to estimate, exact p-values to test — and let clinical judgement decide what is important.
Keywords
p-value, standard deviation, standard error, standard deviation vs standard error, SD vs SE, what is p-value, p-value meaning, p-value interpretation, standard error of mean, SEM, how to calculate standard deviation, standard error formula, 95% confidence interval, statistical significance, clinical significance, mean ± SD, p < 0.05, null hypothesis, biostatistics for medical students, statistics for thesis
References
- Wasserstein RL, Lazar NA. The ASA statement on p-values: context, process, and purpose. The American Statistician. 2016;70(2):129–133.
- Altman DG, Bland JM. Standard deviations and standard errors. BMJ. 2005;331:903.
- Kirkwood BR, Sterne JAC. Essential Medical Statistics. 2nd ed. Blackwell Science; 2003.
- Greenland S, et al. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. European Journal of Epidemiology. 2016;31:337–350.
- Mahajan BK. Methods in Biostatistics for Medical Students and Research Workers. Jaypee Brothers.
Related services
- Research Topic Selection & RefinementRefinement of a research topic based on feasibility, relevance, available literature and study setting. Final topic approval remains with the researcher/institution.
- Research Question, Objectives & HypothesisSupport to formulate a focused research question, primary/secondary objectives and testable hypothesis where applicable.
- Literature Search StrategySearch-term development, database search strategy and inclusion/exclusion framework for a literature search.
- Preliminary Literature ReviewStructured summary of relevant literature to help identify context and research gaps; scope and number of papers agreed in advance.
- Synopsis Preparation SupportStructured synopsis support based on the researcher's own study idea, institutional format and supplied references; includes one revision.
- Study Design ConsultationConsultation to compare appropriate observational, experimental or other study designs for a defined research question.
- Sample Size CalculationSample size estimate with stated assumptions, formula/software approach and brief justification; final inputs must be confirmed by the researcher/supervisor.
- Protocol ReviewExpert review of an existing protocol for clarity, design, feasibility, outcomes, eligibility criteria and methodological gaps; written feedback.