+- The Lumin Archive (https://theluminarchive.co.uk)
+-- Forum: The Lumin Archive — Core Forums (https://theluminarchive.co.uk/forumdisplay.php?fid=3)
+--- Forum: Courses — Structured Learning (https://theluminarchive.co.uk/forumdisplay.php?fid=69)
+---- Forum: Probability & Statistics: From Intuition to Mastery (https://theluminarchive.co.uk/forumdisplay.php?fid=71)
+---- Thread: CHAPTER 18 — MISLEADING GRAPHS & STATISTICS (/showthread.php?tid=215)
CHAPTER 18 — MISLEADING GRAPHS & STATISTICS - Leejohnston - 11-15-2025
Chapter 18 — Misleading Graphs & Statistics
Most people trust graphs automatically.
That’s why graphs are one of the most commonly abused tools in:
• news
• politics
• advertising
• social media
• product marketing
• scientific misunderstandings
This chapter teaches you how to spot when graphs (or statistics) are designed to mislead.
Once you know these tricks, you see them EVERYWHERE.
---
18.1 Why People Manipulate Graphs
Graphs are manipulated to:
• exaggerate a difference
• hide a difference
• make results look more impressive
• make results look less dangerous
• push a narrative
• sell a product
• influence public opinion
A misleading graph doesn’t lie outright —
it simply presents the truth in a *distorted* way.
---
18.2 Trick #1 — Using a Broken or Truncated Y-Axis
The most common trick.
Example:
A company shows sales rising from:
• £998,000
• £1,000,000
A normal axis starting at 0 would show almost no change.
A misleading graph starts the Y-axis at £990,000 —
making the tiny increase look gigantic.
Rule: Always check where the Y-axis starts.
If it's not zero, the graph is probably exaggerating.
---
18.3 Trick #2 — Changing the Scale
Scales can be stretched or compressed.
Two graphs might show identical data but appear totally different.
Examples:
• stretching the axis makes changes look tiny
• compressing the axis makes changes look extreme
• uneven spacing hides trends
Always check the spacing between the numbers on the axis.
---
18.4 Trick #3 — Oversized or Odd-Shaped Bar Charts
Sometimes bars are:
• extra wide
• 3D shaded
• decorated with images
• positioned unevenly
This can make:
• small differences look huge
• data harder to compare
• your brain focus on colour/shape instead of size
Exams LOVE asking about this trick.
---
18.5 Trick #4 — Cherry-Picking the Time Frame
This is a powerful deception.
Example:
• A stock drops for 6 months, then rises for 2 weeks.
The company shows only the 2-week rise.
Or:
• A decline looks extreme only because the graph starts at a high point.
Data without context can be highly misleading.
---
18.6 Trick #5 — Using Percentages Without Base Numbers
A classic trick in advertising.
Example:
“Product X kills 50% more germs!”
But:
• Original effectiveness: 2%
• New effectiveness: 3%
Yes, that’s technically “50% more”,
but the real change is meaningless.
Percentages mean nothing without raw numbers.
---
18.7 Trick #6 — Using Averages Incorrectly
Sometimes “average” is chosen to hide the truth.
Three types of average:
• mean
• median
• mode
Example:
Income in a city:
• 90% of people earn £20,000
• 10% earn £1,000,000
Mean income is very high —
but the “typical person” earns far less.
Median is usually better for real life.*
---
18.8 Trick #7 — Misleading Pie Charts
Common mistakes:
• slices not proportional
• colour confusion
• 3D perspective distorting sizes
• labels missing
• too many categories squeezed in
A poorly made pie chart can be almost impossible to interpret accurately.
---
18.9 Trick #8 — Using Cumulative Graphs Without Warning
A cumulative graph always goes up.
Sometimes people show cumulative data and pretend it represents:
• growth
• popularity
• success
• performance
When in reality:
It just means the total number collected so far.
Example:
Cumulative deaths vs daily deaths
→ VERY different story.
---
18.10 Trick #9 — Selective Data Removal
Removing:
• outliers
• inconvenient categories
• early years
• extreme values
Can totally distort the narrative.
Example:
Removing “bad months” from a business graph
→ makes performance look stable.
---
18.11 Trick #10 — Correlation Presented as Causation
A graph might show:
• ice cream sales go up
• drowning incidents go up
That does NOT mean ice cream causes drowning.
A third factor (hot weather) explains both.
If a graph claims cause, be suspicious.
---
18.12 How to Defend Yourself Against Misleading Data
Ask yourself:
• Does the axis start at zero?
• Is the scale consistent?
• Is the time frame complete?
• Are percentages hiding real numbers?
• Which average is being used?
• Are any categories missing?
• Is correlation being mistaken for causation?
• Is the graph exaggerating or minimising something?
If something feels “off”, you’re probably right.
---
18.13 Exam-Style Questions
1. A bar graph starts the Y-axis at 50 instead of 0.
Explain why this might be misleading.
2. A pie chart shows “40%” but the slice looks like 25%.
Describe what is wrong.
3. A company shows its last 3 months of profits, ignoring previous years.
What trick is being used?
4. Ice cream sales and shark attacks rise together.
Why is this NOT evidence that one causes the other?
5. A line graph changes scale halfway through the axis.
Explain why this is dangerous.
---
18.14 Chapter Summary
• Graphs can distort the truth
• Axis manipulation is the most common trick
• Time frame selection can hide important information
• Percentages are meaningless without raw numbers
• Averages can be used dishonestly
• Correlation ≠ causation
• Always check scale, labels, and context
You now have the skills to detect *data manipulation* in the real world.
This makes you more informed, more logical, and harder to deceive.
---
Written and Compiled by Lee Johnston — Founder of The Lumin Archive