Cookie Cutter Review: Implicit Causation

Lucas Carter

I want to quickly discuss a common type of causation argument that LSAC uses.

Here is an example:

Those who wear glasses are more likely than those who do not to have knee problems. To ensure good knee health, ditch the glasses.

We take a correlation and make a recommendation, seems pretty innocuous- maybe this is sound advice.

No! This advice is rooted in making an assumption. This assumption is a really bad reasoning error. It is assuming that wearing glasses is what causes knees to have problems. That is why the advice to stop wearing glasses to prevent knee damage is given. Notice how the argument never comes out and says "Glasses cause Knee problems", that would be too easy. The implicit assumption that the argument makes is inferring causation from correlation.

As we know, when A is correlated with B, there are 4 possibilities :

1. A causes B
2. B causes A
3. 3rd common cause
4. No relationship

For our advice to ditch the glasses to work, we would need A to cause B, or, in other words, glasses to cause knee problems. If it really is the case that knee problems cause people to wear glasses (B causes A), then just stopping wearing glasses will do nothing, the advice would be terrible. Similarly, if genetics causes both knee problems and glasses and that is why we have our correlation, then taking glasses off will do nothing. In short, the only way our advice works is if glasses really do cause knee problems. We cannot say this is the case just based on the existence of a correlation, there are 3 other possibilities which are equally likely.

Boiled down to variables the argument goes like this:

**A is correlated with B

---

If you desire B, just do A.

or

If you want to prevent B, don't to A**

Well, for this advice to make sense, we must assume that A causes B and we cannot do that based on a correlation.

These questions are sometimes tricky because they make intuitive sense. They will really try to make the advice sound good, despite making a correlation causation error. Here is one last example:

People with a lot of sugar in their diets tend to get disease XYZ more often than those who do not. To lower your risk of XYZ, cut out sugar from your diet.

Well, we know sugar is bad for health, so this does not seem bad at all. BUT, this argument commits the error of taking a correlation and jumping to the conclusion that sugar is what is causing XYZ. This is done implicitly (hence to title of the post) and is not ok for the reasons discussed above!

PT 78 S3 Q21 (https://7sage.com/lsat_explanations/lsat-78-section-3-question-21/) is a good example of this form and disguises the flaw with an argument that seems to make sense.

Hope this was helpful!