Imagine 100 drivers in total get into accidents in Dole County.
SCENARIO A: Most people wear seat belts (Say, 90 out of 100)
Total drivers in accidents: 90 belted, 10 unbelted.
Now look at who gets severely injured:
Suppose 10 people get severely injured.
The study says 80% of them were unbelted = 8 unbelted people got severely injured.
That means 8 out of 10 unbelted drivers got severely injured (80% injury rate).
Only 2 out of 90 belted drivers got severely injured (2.2% injury rate).
Conclusion: Seat belts clearly protect you (2.2% vs 80%). The argument works!
SCENERIO B: Almost nobody wears seat belts (Say, only 20 out of 100)
Total drivers in accidents: 20 belted, 80 unbelted.
Now look at who gets severely injured:
Suppose 10 people get severely injured.
The study says 80% of them were unbelted = 8 unbelted people got severely injured.
That leaves 2 belted people who got severely injured.
Calculate the injury rates:
Unbelted injury rate: 8 out of 80 = 10%
Belted injury rate: 2 out of 20 = 10%
Conclusion: Belted and unbelted drivers get injured at the exact same rate (10%). Wearing a seat belt made zero difference!
Why is this an assumption?
The argument assumes Scenario A is true.
If Scenario B is true instead, the stat ("80% of injured people were unbelted") is meaningless because 80% of all drivers were unbelted anyway.
Therefore, for the argument to work, the overall percentage of drivers who wear seat belts in Dole County MUST be much higher than 20%.