Skip to content
ponko
Create a quiz

Shared quiz

Charts That Lie

Ten ways a graph can be technically accurate and still mislead you. Works as a lesson starter and as a meeting warm-up -- everyone has seen these and most people cannot name them.

Questions
10
Time to play
about 6 min
Category
Education
Language
English
  • data
  • statistics
  • critical-thinking
  • media-literacy

How to use this quiz

  1. Make a copy — it lands in your own quizzes, where you can edit every question. The original is untouched.
  2. Start a game from your copy. Ponko puts a join code and QR code on your screen.
  3. Players join from any phone browser, answer on their own device, and the scores update live.

Hosting and playing are free, and players don't need an account. More in the FAQ

The questions

Correct answers are shown below, so you can check the quiz before you use it.

Question 1

A bar chart's vertical axis starts at 95 instead of 0. What does that do?

  • Makes small differences look enormouscorrect
  • Makes the chart more precise
  • Hides outliers at the top
  • Nothing -- it is standard practice

Why

Bar lengths are read as the size of the value. Starting the axis at 95 cuts off almost all of each bar, so a gap of 1 or 2 can look like double.

20 seconds to answer

Question 2

A chart puts two different measurements on two different vertical axes. What is the risk?

  • The two lines can be made to appear related by choosing the scalescorrect
  • The chart cannot be printed in black and white
  • One line will always be hidden
  • The axes must use the same units

Why

Separate vertical scales can make changes look similar in size and create misleading overlaps or crossings, suggesting a relationship the data do not establish.

25 seconds to answer

Question 3

The slices of a pie chart should always add up to 100%.

  • Truecorrect
  • False

Why

A pie chart shows how one whole divides into parts, so the slices must add up to 100%. If they do not, a pie chart is the wrong chart.

15 seconds to answer

Question 4

On a bubble chart, a designer doubles a bubble's radius to show a doubled value. What has actually happened?

  • The area is four times bigger, so the value looks four times biggercorrect
  • Nothing -- radius and area rise together
  • The area is halved
  • It depends on the colour

Why

A circle's area grows with the square of its radius. Double the radius and the area quadruples, and readers judge bubbles by area.

25 seconds to answer

Question 5

Support rises from 2% to 3%. Which headline is true but most misleading?

  • Support jumps 50%correct
  • Support rises one point
  • Support rises from 2% to 3%
  • Support remains in the low single digits

Why

Going from 2% to 3% is a rise of one percentage point but a 50% relative increase. Both are true; reporting only the 50% makes a small change sound dramatic.

25 seconds to answer

Question 6

A graph shows only the years where the line went the way the author wanted. What is that called?

Accepted answers

cherry picking · cherry-picking

Why

Cherry picking means showing only the data that supports a conclusion and leaving out the rest, for example by choosing convenient start and end dates.

30 seconds to answer

Question 7

Ice cream sales and drowning deaths rise together. What is the most likely explanation?

  • Both rise in hot weathercorrect
  • Ice cream causes drowning
  • Drowning causes ice cream sales
  • The data must be wrong

Why

Both go up in summer: hot weather sells ice cream and sends more people swimming. A hidden third factor like this is called a confounder, and correlation alone does not show causation.

25 seconds to answer

Question 8

A survey result is given with no sample size. Why does that matter?

  • Six people and six thousand people give the same-looking percentagecorrect
  • The percentages will not add to 100
  • It cannot be turned into a chart
  • Sample size only matters for medical studies

Why

Both 3 out of 6 and 3,000 out of 6,000 are 50%. With comparable random sampling, the larger sample usually has less sampling uncertainty, but reliability also depends on representativeness, question wording and survey design.

25 seconds to answer

Question 9

A logarithmic scale is used to deceive the reader.

  • True
  • Falsecorrect

Why

A logarithmic scale shows equal ratios as equal distances. It is useful for exponential growth or data spanning many orders of magnitude and is not inherently deceptive. It should be clearly labelled and suited to the data and audience.

20 seconds to answer

Question 10

Which of these have you been fooled by most recently?

Poll — no right answer, players just vote.

  • A truncated axis
  • A percentage of a percentage
  • A cherry-picked date range
  • A 3D pie chart

20 seconds to answer

Browse more public quizzes