Charts: one variable | R for Psychological Research

Set up

R code

library(tidyverse)
students <- read_csv("data/students.csv", show_col_types = FALSE) %>%
  mutate(Motivation = factor(Motivation, levels = c("Low", "Medium", "High")))

The recipe

1. Data

ggplot(students,

2. Aesthetics

which column goes where
aes(x = ExamScore))

3. Geom

what to draw
geom_histogram()

Every ggplot chart is built from these three parts, joined with +.

Data and aesthetics alone give an empty canvas:

R code

ggplot(students, aes(x = ExamScore))

An empty chart with exam score on the x axis

Add a geom with +:

R code

ggplot(students, aes(x = ExamScore)) +
  geom_histogram()

Output

`stat_bin()` using `bins = 30`. Pick better value `binwidth`.

Histogram of exam scores in 30 narrow bars

The message suggests choosing a bar width. Let's do that.

Histograms: one number

A histogram shows how values are spread: where most are, and how wide the spread is.

R code

ggplot(students, aes(x = ExamScore)) +
  geom_histogram(binwidth = 5, fill = "steelblue", colour = "white")

Histogram of exam scores in 5-mark bars; most students scored between 55 and 80

binwidth = 5: each bar covers 5 marks.
fill colours the bars, colour their outline.

Colour by group

Put fill inside aes() to colour by a column. facet_wrap() gives each group its own panel:

R code

ggplot(students, aes(x = ExamScore, fill = Gender)) +
  geom_histogram(binwidth = 5, colour = "white") +
  facet_wrap(~ Gender)

Two histograms of exam scores, for female and male students, with similar shapes

Outside aes()

fill = "steelblue"
sets one colour for all

Inside aes()

aes(fill = Gender)
colours by a column, adds a legend

Bar charts: categories

geom_bar() counts rows in each category:

R code

ggplot(students, aes(x = Motivation)) +
  geom_bar()

Bar chart of the number of students in each motivation group; Medium is the largest

Split each bar by another category with fill, side by side with position = "dodge":

R code

ggplot(students, aes(x = Motivation, fill = Gender)) +
  geom_bar(position = "dodge")

Grouped bar chart of motivation for female and male students

Exercises

Exercise 8.1 Dataset: students.csv

Histogram of sleep

Make a histogram of Sleep with bars half an hour wide (binwidth = 0.5). What does the warning mean?

Show answerHide answer

R code

ggplot(students, aes(x = Sleep)) +
  geom_histogram(binwidth = 0.5, fill = "darkseagreen", colour = "white")

Output

Warning message:
Removed 10 rows containing non-finite outside the scale range
(`stat_bin()`).

Histogram of hours of sleep; most students sleep between 5.5 and 7.5 hours

Most students sleep 5.5 to 7.5 hours. The warning says 10 rows were left out: the students with no sleep answer.

Exercise 8.2 Dataset: students.csv

Bar chart

Make a bar chart of Living (Hall or Home), split by Workshop, side by side.

Show answerHide answer

R code

ggplot(students, aes(x = Living, fill = Workshop)) +
  geom_bar(position = "dodge")

Grouped bar chart of where students live, split by whether they took the workshop

Key points

ggplot(data, aes(...)) + geom_...()
Histogram = one number; bar chart = counts of categories.
Inside aes() maps a column; outside sets a fixed look.