Data presentation
In a minute
Raw data is just a list until you organise it. This lesson turns lists into and , reads medians, quartiles and off a graph, and uses to compare data sets.
You'll be able to
- Put discrete and continuous data into a frequency table, with written as inequalities with no gaps.
- Draw and read a frequency histogram with equal class widths, and describe its shape.
- Use a cumulative frequency graph to estimate the median, quartiles, percentiles, range and IQR.
- Draw and read a box and whisker diagram, with outliers as crosses.
- Compare two distributions from their box plots, and judge from symmetry whether data may be normal.
Before you start3 quick questions on what this builds on
Revise first: 4.1 Sampling, data quality and outliers
Frequency tables
A list of numbers is hard to take in. Count how often each value appears, and the pattern shows at once.
For with only a few different values, give each value its own column. Tally the list once, then count the marks.
Try it first, then reveal one step at a time.
- The values run from to , so the table needs a column for each of .
- Go through the list once, making a tally mark under each value. Cross off each number as you go, so none is counted twice.
- Check the total: , one for each day.
is different. Measured values are rarely equal, so you group them into classes instead.
Classes must leave no gaps. With classes – and –, a time of has nowhere to go. With and , every time lands in exactly one class.
A value on a boundary goes in the class whose inequality it satisfies. Here goes in . Some tables use the form instead, so read the signs each time.
Try it first, then reveal one step at a time.
- Sort out the boundary values first. goes in , in and in .
- is less than , so it goes in .
- Tally the rest, then count.
- Check: runners.
Histograms
A table gives you numbers. A picture shows where the data bunches up, and whether it trails off to one side.
In this course every class in a histogram has the same width. So a taller bar always means more values.
Try it first, then reveal one step at a time.
- Horizontal axis: time, minutes, from to , with a mark at each class boundary.
- Vertical axis: frequency, from to , with an even scale. The largest frequency, , fits.
- Draw each bar from its lower to its upper boundary, as tall as its frequency: , , and . The bars touch.
Now look at the shape. If the two sides roughly mirror each other, the data is symmetric. If one side trails off in a long tail, the data is .
A long tail to the right is positive skew, and a long tail to the left is negative skew. Symmetric, bell-shaped data may follow the normal distribution of 4.9.
Try it first, then reveal one step at a time.
- (a) “At least ” means the three bars from to . Read their heights: , and .
- Add them: patients.
- (b) The bars rise to a peak at –, then fall away slowly to . The long tail is on the right, so the waiting times are positively skewed.
Cumulative frequency graphs
A takeaway records the delivery times of orders. How many arrived within minutes? Add the frequencies of every class up to .
1. Add a cumulative frequency column to the table.
2. Plot each cumulative frequency above the upper boundary of its class. Only there has every value in the class been counted.
3. Plot a first point at the lower boundary of the first class, with cumulative frequency .
4. Join the points with a smooth curve or with straight lines. This lesson uses straight lines. A curve gives slightly different readings.
Try it first, then reveal one step at a time.
- Keep a running total, adding the next class each time.
- The last total is , the number of orders. That is your check.
- Plot , , , , and , then join them in order.
Now read the graph the other way. Half the data lies below the median. A quarter lies below the lower quartile , and three quarters below the upper quartile . So, with values, read across from these heights:
The is the spread of the middle half of the data. It is in the formula booklet.
The range is the largest value minus the smallest. Grouped data hides the exact values, so estimate it from where the graph starts and ends: here minutes.
Try it first, then reveal one step at a time.
- (a) The median is at . Read across and down: minutes. As a check, is halfway from to , so the time is halfway from to .
- is at , giving . is at , giving . So minutes.
- (b) The slowest lie above the th percentile, at . Read across and down: minutes.
- (c) Read up from and across: orders took minutes or less.
- The graph counts orders at or below a time, so subtract from the total: .
Box and whisker diagrams
A box and whisker diagram, or box plot, draws five numbers to scale. It shows where the data starts and ends, and where its middle half sits.
1. Draw a scale, then the box from to with a line at the median.
2. Find the fences and from 4.1. Any value beyond them is an .
3. Draw each whisker to the most extreme value that is not an outlier, and mark each outlier with a cross.
In the fish diagram, and , so and . The fences are and .
No fish is shorter than cm, so the left whisker reaches the shortest fish, cm. A cm fish is not beyond the upper fence, so the right whisker ends there. The cm fish is beyond it, so it is a cross.
Try it first, then reveal one step at a time.
- (a) Enter the items in one list and the frequencies in a second. Run one-variable statistics with the second list as the frequency list.
- Write down what you read: smallest , , median , , largest .
- (b) . The fences are and .
- Nothing is below . Only , so the order of items is an outlier.
- (c) The box runs from to with a line at . The left whisker reaches . The right whisker reaches , the largest value that is not an outlier, and a cross marks .
Box plots are best for comparing two data sets on one scale. Make two comparisons, each in context: the medians for centre, and the IQRs or ranges for spread. A box plot doesn't show the mean, so don't compare means.
Shape matters too. If the median sits in the middle of the box and the whiskers are about the same length, the data is symmetric, so it may be normally distributed (4.9). A lopsided box or one long whisker means skewed data, which is unlikely to be normal.
Try it first, then reveal one step at a time.
- (a) and .
- Check for outliers. Class A's fences are and , and class B's are and . Every mark lies inside its fences, so each whisker reaches the smallest or largest mark.
- (b) Centre: class B's median, , is higher than class A's, . So class B generally scored higher.
- Spread: class B's IQR, , is smaller than class A's, . So class B's marks were more consistent.
- (c) Class A: the median is from each quartile and both whiskers are long. It is symmetric, so its marks may be normal. Class B's whiskers are and long, so its marks are skewed.
Common mistakes
- Plotting cumulative frequency at class midpoints. Plot it at each upper class boundary.
- Reading “more than” straight off a cumulative frequency graph. The graph counts values at or below, so subtract the reading from . For the top , read across from of .
- Drawing a whisker out to an outlier, or calling a value equal to a fence an outlier. Whiskers stop at the most extreme value that is not an outlier.
- Leaving out the frequency list when entering a frequency table on the GDC. Each value then counts once, and the quartiles are wrong.
- Comparing means from box plots. A box plot shows medians and quartiles, so compare those, in context.
Exam tips
- Use the graph to estimate: write down the height you read across from, such as for the median, and draw the reading lines. That earns the method mark.
- Draw a histogram or box plot: label the axes, use an even scale and plot accurately. Mark outliers with crosses.
- Compare two distributions: one sentence on the medians and one on the IQRs, each with numbers and in context.
Key takeaways
Practise this
3 past-paper questions on 4.2 in the AA HL bank, each with a mark scheme and a worked explanation.