Error bars in ggplot2 with geom_errorbar

Sample data

geom_errorbar requires a summary data frame with a central value and lower/upper bounds. Here we compute the mean and standard deviation of Sepal.Length per species from the built-in iris dataset.

# install.packages("dplyr")
library(dplyr)

df <- iris |>
  group_by(Species) |>
  summarise(
    mean_sl = mean(Sepal.Length),
    sd_sl   = sd(Sepal.Length)
  )

`geom_errorbar()`

Map ymin and ymax inside aes() to define the lower and upper bounds of each bar. Here we use mean ± 1 SD.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df, aes(x = Species, y = mean_sl)) +
  geom_errorbar(aes(ymin = mean_sl - sd_sl,
                    ymax = mean_sl + sd_sl))

Basic error bars in ggplot2 with geom_errorbar

`width`

Controlling the cap width of error bars in ggplot2

width controls the size of the horizontal caps. The default is relatively wide — 0.2 is a common choice for a cleaner look.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df, aes(x = Species, y = mean_sl)) +
  geom_errorbar(aes(ymin = mean_sl - sd_sl,
                    ymax = mean_sl + sd_sl),
                width = 0.2)

With a bar chart

The most common use of geom_errorbar is on top of a bar chart. Add geom_col() first so the bars sit behind the error bars.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df, aes(x = Species, y = mean_sl)) +
  geom_col(fill = "steelblue", alpha = 0.7) +
  geom_errorbar(aes(ymin = mean_sl - sd_sl,
                    ymax = mean_sl + sd_sl),
                width = 0.2)

Error bars on top of a bar chart in ggplot2

With points

Error bars combined with points in ggplot2 dot plot

Draw the error bar first and the point on top for a clean dot plot with confidence intervals.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df, aes(x = Species, y = mean_sl)) +
  geom_errorbar(aes(ymin = mean_sl - sd_sl,
                    ymax = mean_sl + sd_sl),
                width = 0.2) +
  geom_point(size = 3, color = "steelblue")

With a line chart

Combine geom_line(), geom_errorbar() and geom_point() to show uncertainty over time.

# install.packages("ggplot2")
library(ggplot2)

df_time <- data.frame(
  year   = 2018:2023,
  mean_y = c(3.2, 3.8, 4.1, 3.9, 4.5, 4.8),
  sd_y   = c(0.5, 0.4, 0.6, 0.5, 0.4, 0.3)
)

ggplot(df_time, aes(x = year, y = mean_y)) +
  geom_line() +
  geom_errorbar(aes(ymin = mean_y - sd_y,
                    ymax = mean_y + sd_y),
                width = 0.2) +
  geom_point(size = 2)

Error bars on a line chart in ggplot2 showing uncertainty over time

`geom_errorbarh()`

Horizontal error bars in ggplot2 with geom_errorbarh

Use geom_errorbarh() for horizontal error bars. Swap x/y in aes() and replace ymin/ymax with xmin/xmax. The cap size is controlled with height instead of width.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df, aes(y = Species, x = mean_sl)) +
  geom_errorbarh(aes(xmin = mean_sl - sd_sl,
                     xmax = mean_sl + sd_sl),
                 height = 0.2) +
  geom_point(size = 3, color = "steelblue")

Grouped error bars

When you have multiple groups per category, use position_dodge() on both geom_errorbar() and geom_point() with the same dodge width so they stay aligned.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df_group, aes(x = group, y = mean_y,
                     color = condition)) +
  geom_errorbar(aes(ymin = mean_y - sd_y,
                    ymax = mean_y + sd_y),
                width = 0.2,
                position = position_dodge(0.5)) +
  geom_point(size = 3,
             position = position_dodge(0.5))

Grouped error bars with position_dodge in ggplot2

Color, linewidth and linetype

Customized error bars with color linewidth and linetype in ggplot2

Pass color, linewidth and linetype outside aes() to apply them to all bars at once.

# install.packages("ggplot2")
library(ggplot2)

ggplot(df, aes(x = Species, y = mean_sl)) +
  geom_errorbar(aes(ymin = mean_sl - sd_sl,
                    ymax = mean_sl + sd_sl),
                width = 0.2,
                color = "steelblue",
                linewidth = 0.8,
                linetype = "dashed")