When a Pie Chart Stops Being a Fair Picture
A pie chart represents a whole by dividing a circle into slices. Its strength is also its main limitation: every slice is judged as part of that circle, so the chart is only fair when the slices share one clear total and the visual areas correspond to the displayed proportions. Perspective, inconsistent denominators, or changing circle sizes can break that connection.
A chart can look polished and still make comparisons difficult or misleading. A three-dimensional effect may make a front slice seem larger than an equal slice at the back. Two pies may each fill a circle even though they summarize different sample sizes, hiding the difference between their totals. Small differences between similar angles may also be hard to see. The key question is whether a viewer can recover the intended comparison from the visual encoding.
For example, if 25% of responses fall in one category, its angle should be \(360(0.25)=90\) degrees. For categories that cover the whole, the proportions add to 1, so their angles add to 360 degrees. This common total is what allows a pie chart to work at all.
Perspective and Area Can Distort the Message
A flat pie chart uses angle and area to show a category's share. A three-dimensional pie chart adds depth and perspective, but the display is still viewed on a flat page or screen. The near edge may appear wider or more prominent than the far edge; shading and an exploded slice can add further emphasis. Those effects do not change the data, but they can change the apparent size of a slice. A 3-D pie therefore adds visual complexity without making the proportions more informative.
Area also matters when pie charts are compared. In a circle of radius \(r\), a slice with proportion \(p\) has area \(p\pi r^2\). With the same radius for every pie, slice area tracks the proportion. If the radius changes, the total circle area changes with its square. A slice in the larger circle can then look much larger even when its proportion is unchanged.
This relationship follows from the area of a full circle, \(\pi r^2\). A sector occupying proportion \(p\) of the circle also occupies proportion \(p\) of its area, so its area is \(p(\pi r^2)\). If the radius is multiplied by 2, the full area is multiplied by \(2^2=4\), not 2. This is why changing pie size to suggest a larger total can produce an unexpectedly strong area difference.
Worked Example: Replacing a Three-Dimensional Pie
Suppose a hypothetical campus poll asks 100 students which study location they use most often. The responses are library, residence hall, café, and other, with counts 40, 30, 20, and 10. A designer presents the pie in three dimensions and places the largest slice at the front. What should the slices represent, and how could the results be shown more clearly?
Check the whole. The categories are intended to be mutually exclusive answers to one question, and their counts sum to the 100 responses: \(40+30+20+10=100\). The shares are therefore \(0.40\), \(0.30\), \(0.20\), and \(0.10\).
Calculate the angles. The library slice should be \(360(0.40)=144^\circ\); residence hall, \(360(0.30)=108^\circ\); café, \(360(0.20)=72^\circ\); and other, \(360(0.10)=36^\circ\). Their angles sum to \(144+108+72+36=360^\circ\).
The three-dimensional perspective has no role in these calculations. It can make the front slice appear more prominent than the same angle would appear in a flat pie, so a viewer may overestimate the library share. A clear replacement is a bar chart with category labels and bars of heights 40%, 30%, 20%, and 10% (or the corresponding counts). Bars use a common baseline and make differences easier to compare; the three-dimensional effect should be removed.
One Pie Must Have One Denominator
A pie chart represents one whole, so its slices must be calculated from the same total. A common error is to use the full pooled sample as the denominator for slices in several separate pies, even though each pie represents only one group. The resulting percentages within a pie will not add to 100%, and its angles will not add to 360 degrees. Another common trap is to draw a full pie for each group: each circle necessarily fills the page, so the circles alone do not reveal that the groups have different sizes.
Before comparing pies, decide whether the question concerns composition (the proportions within each group) or counts (how many observations are in each category). A pie can show composition, but a set of equally sized pies does not show the group totals. When both group size and category values matter, label the counts or use a display that makes the requested comparison directly.
Worked Example: Correcting Pies With Mismatched Denominators
In a hypothetical survey, Group A has 40 respondents: 18 choose option X, 12 choose Y, and 10 choose Z. Group B has 60 respondents: 18 choose X, 24 choose Y, and 18 choose Z. A draft chart divides every count by the pooled total of 100, then uses those percentages as slices in a separate full pie for each group. Check the chart and correct the values.
For Group A, the draft percentages are \(18/100=18\%\), \(12/100=12\%\), and \(10/100=10\%\). They sum to 40%, not 100%. The corresponding angles are \(64.8^\circ\), \(43.2^\circ\), and \(36^\circ\), which sum to \(144^\circ\), not \(360^\circ\). They cannot describe the three slices of a full pie for Group A.
For Group B, the draft percentages are 18%, 24%, and 18%, summing to 60%; the angles are \(64.8^\circ\), \(86.4^\circ\), and \(64.8^\circ\), summing to \(216^\circ\). These values are shares of the pooled sample, not shares within Group B.
To show composition, use each group's own total. Group A's shares are \(18/40=45\%\), \(12/40=30\%\), and \(10/40=25\%\); the angles are \(162^\circ\), \(108^\circ\), and \(90^\circ\), totaling \(360^\circ\). Group B's shares are \(18/60=30\%\), \(24/60=40\%\), and \(18/60=30\%\); the angles are \(108^\circ\), \(144^\circ\), and \(108^\circ\), also totaling \(360^\circ\).
These corrected percentages describe composition, but two full-size pies still hide that Group B has 20 more respondents. A 100% stacked bar chart would compare compositions while making the shared 100% scale explicit. If the question is instead how many people chose each option, a grouped bar chart of counts or a table showing both counts and group totals is clearer. In either case, retain the denominators so readers know what each percentage describes.
Similar Angles Can Be Hard to Compare
Even a correctly constructed flat pie can be difficult to read when its slices have similar angles. Viewers must compare angles or areas, often without a common straight baseline. A small difference may be hard to judge, especially among several categories. Labels with percentages can help readers identify the values, but they do not make the visual comparison itself easier.
Worked Example: Showing Small Differences Clearly
A hypothetical survey of 200 people records four mutually exclusive, exhaustive responses. The counts are 52, 50, 48, and 50, for a total of 200. A pie chart is mathematically valid, but would it make the differences easy to compare?
The proportions are \(52/200=26\%\), \(50/200=25\%\), \(48/200=24\%\), and \(50/200=25\%\). Their angles are \(360(0.26)=93.6^\circ\), \(360(0.25)=90^\circ\), \(360(0.24)=86.4^\circ\), and \(360(0.25)=90^\circ\). These angles sum to \(360^\circ\), so the pie is valid.
However, the largest share exceeds the smallest by only 2 percentage points, or \(93.6^\circ-86.4^\circ=7.2^\circ\). The slices will look similar, making that difference difficult to judge by eye. A bar chart with all bars starting from zero gives the values a common baseline; labels can show the exact percentages or counts. The graph should not exaggerate the small difference by cutting off the bar baseline.
A Practical Check Before Using a Pie
A pie chart is most useful when the purpose is to show parts of one whole and the number of categories is small enough for the slices to be distinguishable. Before accepting or making one, check the underlying totals and then inspect how the visual encoding works.
State which observations form the denominator for the pie. Each slice must use that same total.
Categories should be mutually exclusive and cover the whole. Counts should sum to the stated total, proportions should sum to 1 (or 100%), and slice angles should sum to 360 degrees.
Use a flat circle of consistent size. Avoid perspective, 3-D effects, and exploded slices that can make some categories seem more prominent.
Use a bar chart when exact comparisons, small differences, or comparisons across groups are central. Include totals or counts when sample sizes differ.
This check separates two problems that can be confused. A denominator error makes the numerical slices wrong; a perspective or area effect makes the visual impression unreliable even if the underlying arithmetic is correct. Correcting one does not automatically correct the other.
Common Mistakes
- Assuming a complete-looking circle proves the data are valid. A drawing program can turn any set of positive values into a full pie by rescaling them. Check that the reported percentages use the stated denominator and add to 100%.
- Comparing the areas of pies with different radii as though only the proportions changed. Total circle area depends on \(r^2\). Use equal-sized pies for composition, or choose bars and include totals when counts matter.
- Treating a 3-D slice's apparent area as its true share. Perspective changes the appearance of angles and areas. Use a flat pie or, when comparisons are important, a bar chart.
- Expecting viewers to judge close angles precisely. A valid pie can still be an ineffective comparison. Use aligned bars and provide exact values when small differences matter.
Check Your Understanding
Try these questions using the checks and calculations in this tutorial.
- A flat pie has category proportions 0.50, 0.30, and 0.20. What is the angle of each slice, and what should their angles sum to?
- Why might a three-dimensional pie give a misleading impression even when its counts and percentages are correct?
- Separate pies summarize groups of 35 and 65 people. What denominator should be used for percentages within each pie?
- Two pies use radii of 2 cm and 4 cm. By what factor does the larger full circle's area exceed the smaller one's?
- A valid pie has categories with shares of 24%, 25%, 25%, and 26%. Why might a bar chart be a clearer way to compare them?