Whites have the highest crime rate
Me: Gemini, how do I explain to an obstinate individual why 13% of the population committing 37% of crime is a far worse statistic than 67% of the population committing 60% of crime even though 60% is more than 37%?
Keep in mind, he's a little dense, lies a lot, and is full of the self confidence of the ignorant.
Gemini: Here's a way to explain this, breaking it down into simple concepts:
"Okay, so let's think about this using a really simple example, like two groups of people, apples and oranges, and how many times they 'borrow' a pen without asking.
Imagine Group A (the 'apples') makes up a small part of our classroom – let's say they're only 13% of all the students. But out of all the times pens go missing, 37% of those missing pens were last seen with an 'apple.'
Now, Group B (the 'oranges') makes up most of the classroom – 67% of the students. And they were last seen with 60% of all the missing pens.
At first glance, it looks like the 'oranges' are responsible for more missing pens overall (60% vs. 37%).
But here's the key: There are way fewer 'apples' than 'oranges.'
Think of it this way:
- If there are only 13 apples, and they're involved in 37% of the pen disappearances, each individual 'apple' is involved in a lot of pen disappearances on average.
- If there are 67 oranges, and they're involved in 60% of the pen disappearances, each individual 'orange' is involved in fewer pen disappearances on average, simply because there are so many more of them to spread those 60% across.
So, even though the 'oranges' might be linked to more missing pens in total, the rate at which 'apples' are involved in missing pens, per person, is much, much higher. It's like if one person robs 10 banks, and 100 people rob 1 bank each. The one person is a much more prolific bank robber, even though the 100 people collectively robbed more banks."
To make it even more concrete, you could use simplified numbers if your friend is good with them:
"Let's imagine a town of 100 people to make the math easy.
- Group 1 (13% of the population): That's 13 people.
- Group 2 (67% of the population): That's 67 people.
Now for the crimes:
- If there are 100 violent crimes total:
- Group 1 commits 37%: So, 37 crimes are committed by Group 1 members.
- Group 2 commits 60%: So, 60 crimes are committed by Group 2 members.
Now, let's look at crimes per person:
- For Group 1: 37 crimes / 13 people = roughly 2.85 crimes per person.
- For Group 2: 60 crimes / 67 people = roughly 0.89 crimes per person.
See? Even though Group 2 commits more total crimes (60 vs. 37), each person in Group 1 commits almost three times as many violent crimes on average compared to each person in Group 2. That's what 'per person by a huge margin' means."
Key takeaways to emphasize:
- Total numbers vs. per capita/per person rates: These are different ways of looking at data and can tell very different stories.
-
Proportion matters: When a smaller group accounts for a disproportionately large share of something, it means the individuals within that smaller group are engaging in that activity at a much higher rate."