Error Bars Explained: Standard Deviation (SD) vs. Standard Error (SE) vs. 95% CI
Clear rules for choosing the right error bar metric and interpreting overlap correctly in manuscripts.
Switch ±1 SD, ±1 SE, or 95% CI from raw replicates or summary tables.
SD describes data spread; SE and CI describe mean precision
Standard Deviation (SD) quantifies how variable individual subjects are around the sample mean and does not shrink as sample size grows. Standard Error (SE = SD / √n) quantifies uncertainty in the sample mean itself, shrinking toward zero with larger n.
Why 95% confidence intervals are preferred for inference
A 95% confidence interval translates SE directly into a plausible range for the population mean using the Student-t multiplier. Leading biomedical journals increasingly recommend reporting 95% CI or SD rather than bare ±1 SE bars.
Always state the metric and sample size
Never publish unlabeled error bars. Always specify both the error bar definition (±1 SD, ±1 SE, or 95% CI) and per-group n directly in the figure subtitle or caption.