Lesson 2 of 6 · 15 min
Reading and interpreting box plots
Read a five-number summary, range and IQR off a box plot, use its four quarters to estimate percentages and counts, describe its shape, and know what a box plot cannot show.
What you will be able to do
- Read the five-number summary, the range and the IQR from a box plot.
- Use the four quarters of a box plot to estimate the percentage or number of values in an interval.
- Describe the shape of a distribution from its box plot.
- Recognise what a box plot hides.
Real-estate agents, weather reports and school data walls all use box plots, because one glance shows the centre, the spread and the shape. To read one, you only need to remember what the five lines mark and one fact about the four sections between them.
Key idea
Four quarters
The minimum, , the median, and the maximum split the data into four sections, and each section holds about 25% of the data, whatever its length.
- The box (from to ) holds the middle 50%.
- Below the median, and above the median: 50% each.
- Below or above : 25% each.
| Interval | Approximate percentage of the data |
|---|---|
| minimum to | 25% |
| to median | 25% |
| median to | 25% |
| to maximum | 25% |
| to (the box) | 50% |
| to maximum, or minimum to | 75% |
In any box plot, approximately what percentage of the data lies between the lower quartile and the maximum?
Common mistake
Longer does not mean more
A long whisker does not contain more data than a short one. Both hold about a quarter of the data; the long one has its values more spread out. The same goes for the two halves of the box.
Because every section holds the same share of the data, the lengths of the sections show where the data are bunched and where they are stretched out. That gives you the shape.
Short sections mean bunched-up data; long sections mean stretched-out data. Skew is named after the long tail.
| Shape | Median within the box | Whiskers |
|---|---|---|
| Approximately symmetric | Near the middle | About the same length |
| Positively skewed | Closer to | Right (upper) whisker longer |
| Negatively skewed | Closer to | Left (lower) whisker longer |
The histogram and the box plot both show a tail to the right. Drag the small values up and the large values down until the box plot looks symmetric. What happens to the histogram?
Worked example
Interpret a box plot of house prices
The box plot shows the sale prices of 80 houses in a Toowoomba suburb last year, in thousands of dollars.
- (a) State the median sale price.
- (b) Calculate the interquartile range.
- (c) What percentage of the houses sold for more than $640 000?
- (d) Approximately how many houses sold for between $520 000 and $800 000?
- (e) Describe the shape of the distribution.
Sale prices of 80 houses ($000).
Step 1: Read the five-number summary
Minimum 420, , median 560, , maximum 800 (thousand dollars).
Step 2: (a) Median
The line in the box is at 560, so the median sale price was $560 000.
Step 3: (b) IQR
, so $120 000.
Step 4: (c) Above
$640 000 is , and about 25% of the data lie above .
Step 5: (d) From to the maximum
From ($520 000) to the maximum ($800 000) covers three quarters of the data, 75%: houses.
Each of the four sections holds about a quarter of the data, so three sections hold about three quarters.
Step 6: (e) Shape
The median is closer to (40 above it) than to (80 below it), and the upper whisker (160) is longer than the lower one (100). The prices are positively skewed: most houses sold near $520 000 to $560 000, with a tail of more expensive sales.
Try it yourself first, then check the working one step at a time.
Answer
(a) $560 000 (b) $120 000 (c) 25% (d) about 60 houses (e) positively skewed.
Method
Read the five-number summary off the scale first, with units.
Range and IQR are differences: max − min and .
Each section holds about 25% of the data; multiply by for a count.
Shape comes from the lengths of the sections, not the number of values in them.
Watch out for
Thinking the long upper whisker holds more houses than the short lower whisker. Both hold about 20 houses.
Forgetting that the scale is in thousands: the IQR is $120 000, not $120.
A data set has the five-number summary 2, 10, 14, 16, 18. Which best describes the shape of its distribution?
Note
What a box plot hides
A box plot keeps five numbers and throws the rest away. It can't show:
- how many values there are (a box plot of 10 values can look just like one of 10 000);
- clusters or gaps, so a bimodal distribution can look symmetric;
- individual values other than the five numbers and any outliers.
The data sets 1, 2, 3, 4, 5, 6, 7, 8, 9 and 1, 2, 3, 5, 5, 5, 7, 8, 9 have identical box plots. When the detail matters, look at a dot plot, stem plot or histogram too.
A box plot summarises the sale prices of 120 houses in a Logan suburb. Its lower quartile is $510 000. Approximately how many of the houses sold for less than $510 000?
Exam tip
Reading values off a box plot
Read each value off the scale carefully, write it with its units, and show the subtraction for a range or IQR: " thousand dollars". For a percentage or count question, say which quarters you are using.
Summary
The five lines of a box plot are the minimum, , median, and maximum. The box length is the IQR; the full width is the range.
Each of the four sections holds about 25% of the data, whatever its length. The box holds the middle 50%.
Multiply the percentage by to estimate a count.
Long sections show stretched-out data. Median nearer and a long upper whisker: positive skew. Median nearer and a long lower whisker: negative skew.
A box plot hides the number of values, clusters, gaps and most individual values.
Finished the lesson?
Try the 3 quick checks above and the lesson completes itself, or mark it complete now.