# Thread: Correlation does not imply causality question

1. ## Correlation does not imply causality question

Hello,

Could someone please explain to me why a significant linear correlation does not imply causality?

I understand that there are lurking variables, but unsure as to how this relates to a significant linear correlation.

Thanks in advance for any help!

2. I read a story once in a textbook, I will share it with you:

Basically, it boiled down to this:

In the summertime in (fill in city here), there was a strong correlation between the purchase of ice cream cones and armed robbery. Does this mean that ice cream cone purchases caused the robberies? No, it just means there was a strong linear correlation.

It's a very tricky concept to learn. Generally when we think of x increasing and y increasing we are tempted to read into it too much and say "x causes y". Sometimes, x does cause y, or y causes x, but not always.

Does this help?

3. Yes that helps significantly.
I think I just needed an example.

Thanks!

4. Originally Posted by elizsimca
I read a story once in a textbook, I will share it with you:

Basically, it boiled down to this:

In the summertime in (fill in city here), there was a strong correlation between the purchase of ice cream cones and armed robbery. Does this mean that ice cream cone purchases caused the robberies? No, it just means there was a strong linear correlation.

It's a very tricky concept to learn. Generally when we think of x increasing and y increasing we are tempted to read into it too much and say "x causes y". Sometimes, x does cause y, or y causes x, but not always.

Does this help?
Obviously the robbers are using their ill-gotten gains to buy lots of ice cream.