Standard deviation vs standard error: the difference, in R

Standard deviation describes the spread of data; standard error describes the precision of the mean. An R simulation shows why the standard error shrinks as your sample grows.

Mixing up standard deviation (SD) and standard error (SE) is one of the most common — and most consequential — errors in reporting. They answer different questions. Here’s the distinction, made concrete in R.

The one-line difference

• Standard deviation (SD): how spread out the data are. A property of the population; it does not shrink as you collect more data.
• Standard error (SE): how precise your estimate of the mean is. SE = SD / √n, so it does shrink as the sample grows.

\[SE = \frac{SD}{\sqrt{n}}\]

Watch the SE shrink (and the SD not)

Draw larger and larger samples from the same population and track both:

library(ggplot2)
set.seed(3)
ns <- c(5, 10, 20, 50, 100, 200, 500)
res <- do.call(rbind, lapply(ns, function(n) {
  x <- rnorm(n, mean = 50, sd = 10)
  data.frame(n = n, SD = sd(x), SE = sd(x) / sqrt(n))
}))

ggplot(res, aes(n)) +
  geom_line(aes(y = SD, colour = "SD (spread of data)"), linewidth = 1.2) +
  geom_line(aes(y = SE, colour = "SE (precision of mean)"), linewidth = 1.2) +
  scale_colour_manual(values = c("SD (spread of data)" = "#17a2b8",
                                 "SE (precision of mean)" = "#2f6fed")) +
  labs(title = "SE falls with sample size; SD stays put (~10)",
       x = "Sample size (n)", y = NULL, colour = NULL) +
  theme_minimal(base_family = "sans") +
  theme(legend.position = "bottom",
        plot.title = element_text(face = "bold", colour = "#1b2a4a"))

Figure 1

The SD hovers around the true value of 10 no matter the sample size; the SE keeps falling — more data buys you a more precise estimate of the mean, not less variable data.

Which to report

• Describing how variable your observations are (e.g., patient ages) → SD.
• Describing how precisely you’ve pinned down a mean (e.g., for error bars on a group average, or a confidence interval) → SE.

Error bars are the usual culprit: SE bars look tighter than SD bars, so quietly using SE where readers assume SD overstates precision. Always say which one you plotted.

---

Small reporting choices change the story. We get them right — see our services.