2.6 - Correlation vs. Causation

Like sufficient and necessary, the LSAT frequently tests our understanding of correlation and causation. This is another core concept to your LSAT success.

In this lesson, we'll:

• Define correlation and causation
• Learn to distinguish between them
• Review how we'll encounter this flaw on the LSAT
• Learn to counter this flaw once we spot it

Let's dive in.

What are Correlation and Causation?

First, let's start with some definitions.

Correlation describes a statistical relationship between two or more things. When one thing changes, the other thing tends to change in a seemingly predictable way.

Causation describes a logical cause-and-effect relationship between two or more things. If one thing (the cause) leads to another with certainty (the effect), we have causation.

These definitions might seem trivial at first, but it's easy to confuse the two. While causation necessarily implies correlation, correlation does not imply causation. There are tons of ways data can be correlated without having any causal relationship whatsoever. ‍

Check out this hilarious (and well-researched) site about spurious correlations to see how odd some correlations can be.

Distinguishing Between Correlation and Causation

Let's distinguish between the two using some examples.

Correlation: Ice Cream and Swimming Injuries

Imagine that we discover ice cream sales and swimming injuries are remarkably correlated—as ice cream sales increase, more people get hurt while swimming.

An LSAT question might try to convince you that we should ban selling ice cream because of its clear impact on swimming-related accidents.

But does eating ice ream cause swimming injuries? Of course not. 

The real culprit is much more likely to be something like the temperature outside. When it's hotter, people both swim more frequently and eat more ice cream.

Causation: Flipping a Switch

Now, imagine another scenario: you're flipping a light switch on and off.

On goes the switch, on goes a nearby light. Off goes the switch, off goes the light.

We could reasonably call this a causal relationship. When we flip the switch, the same predictable events occur over and over again.

Take note of the difference from our ice-cream-and-swimming example. Before, we had correlation at best, but with our light switch scenario, we have both correlation and causation.

That is, causation always implies correlation—flipping light switches on is correlated with lights turning on because we've established causality.

However, correlation does not necessarily imply causation—increased ice cream sales do not necessarily cause swimming injuries.

---

Let's recap:

• Correlation is about statistical relationships—two or more things happening in a seemingly related pattern.
• Causation is about logical cause-and-effect. One thing occurs, and then another thing occurs as a result.
• Causation has a higher threshold than correlation. It requires a logical connection between events.
• Correlation can mislead us, suggesting connections where none actually exist.

Correlation / Causation Flaws

You'll see lots of arguments confuse correlation for causation. Here are some common ways you'll be tested on it.

After-Therefore

This common fallacy is known in formal logic as post hoc ergo propter hoc (after this, therefore because of this). I've abbreviated it After-Therefore. But don't get caught up in the lingo.

All this means is that an author will try to convince you that because one event occurs after another, the first event must be the cause.

Consider this argument:

I wore my lucky socks.
I aced the LSAT.
Therefore, my lucky socks caused my high LSAT score.

Bull, right? We can all agree that lucky-sock-wearing and high-LSAT-scoring are correlated at best. They just happened to occur in tandem. That doesn't mean there's a causal connection.

Together-Therefore

This fallacy boils down to the error that two things happening simultaneously suggests causation. It doesn't.

In formal logic, they call this cum hoc ergo propter hoc (with this, therefore because of this). I've shortened it to Together-Therefore.

Consider this example:

Each time I study with music, I get better grades.
Therefore, studying with music causes me to get better grades.

Nah, not necessarily. It's possible, but there are lots of plausible alternative causes.

Which brings us to how we attack this flaw.

Two Ways to Combat Correlation / Causation Flaws

There are two tried and true methods for teasing apart this flaw: swapping the suggested cause and effect and providing alternative explanations.

Swapping Cause and Effect

This morning, I've got a headache. And I didn't sleep particularly well. It must be that my sleeping poorly caused my headache. Right? Not so fast.

Imagine I'd tried to convince you of this on the LSAT.

I slept poorly.
I have a headache.
My poor sleep caused my headache.

I chose this example purposefully. It makes a certain kind of intuitive sense—poor sleep can adversely affect our health.

You'll run into arguments like this all the time on the LSAT. They won't be completely baseless ideas. They'll just be trying to convince you of something beyond a threshold they haven't met yet.

In this example, what if my headache actually caused my poor sleep? What if, in the middle of the night, something in my brain chemistry changed just enough to create what would manifest as a headache upon waking?

These what-ifs are your job as a future attorney. You're not trying to disprove the possibility that poor sleep can lead to headaches. You're trying to disprove this particular claim of causality.

Alternative Explanations

Think back to our sleep-and-headache example. We'll use this same argument but attack it with alternative explanations instead.

What are some alternatives that might lead to both poor sleep and morning headaches?

What if I went out drinking with my buddies last night? Alcohol adversely affects sleep and leads to hangover headaches.

What if I was up all night with a screaming infant? I'm sure my sleep would suffer and my head would be pounding.

What if I suffered head trauma yesterday? I'd probably have sleep troubles and a sore noggin.

You get the picture.

Again, these what-ifs are your responsibility. Arm yourself with alternative explanations and you'll tear this flaw to pieces.

---

What did I miss? How can I improve this article for you and future test-takers? Leave me a comment below.

---

Next, we'll wrap our heads around how the LSAT tests our knowledge of numbers, proportions, and overlapping groups.