In Statistics, we often look for patterns between two different sets of data. When we see a pattern on a scatter diagram, it is very tempting to say, "Aha! Variable \(x\) is making Variable \(y\) happen!"

However, this is one of the biggest traps in Statistics. Just because two things move together doesn't mean one is causing the other. In this chapter, we will learn how to tell the difference between a simple association and a true cause, and why some patterns are actually "fake" or spurious.

1. Correlation vs. Causation

Before we go further, let's define our two main terms clearly:

Correlation: This simply means there is a mathematical relationship or association between two variables. If Variable \(x\) goes up and Variable \(y\) also goes up, they are correlated. (For a deep dive into the types of correlation, see the chapter on "Correlation vocabulary and interpretation").

Causation: This means that a change in one variable directly causes a change in the other. One variable is the "engine" driving the change in the other.

The Golden Rule of Statistics

Correlation does not imply causation.

Don't worry if this seems tricky at first! Even professional scientists have to be careful with this. It means that just because you see a strong positive or negative correlation on a scatter diagram, you cannot automatically claim that one thing caused the other.

Example of Causation: There is a strong correlation between the amount of fuel put into a car and the distance it can travel. This is causation because the fuel is physically required to make the car move.

Example of Correlation (but NOT necessarily causation): There is a correlation between people wearing sunglasses and people buying ice cream. Does wearing sunglasses cause you to feel hungry for ice cream? No! Both are being caused by a third factor: sunny weather.

2. Spurious Correlation

A spurious correlation is a mathematical relationship where two variables appear to be related, but they actually have no direct causal link.

These correlations usually happen because of an underlying factor (sometimes called an extraneous variable) that is affecting both of them at the same time.

How to identify a Spurious Correlation:

1. Look at the two variables: \(x\) and \(y\).
2. Ask yourself: "Does it make logical sense for \(x\) to change \(y\)?
3. If the answer is "no," look for a "hidden" third variable that might be influencing both.

A Classic Example: Ice Cream and Sharks

In some seaside towns, data shows a strong positive correlation between ice cream sales and shark attacks. Does eating ice cream make sharks want to bite you? Of course not!

The underlying factor is warm weather. When it is hot, more people buy ice cream, and more people go swimming in the sea. The warm weather causes both variables to increase, creating a "spurious" link between them.

Quick Review:
If you see a scatter diagram showing that as the number of fire trucks at a fire increases, the amount of damage also increases, is this causation? No! The underlying factor is the size of the fire. Bigger fires need more trucks and cause more damage.

3. Interacting Factors (Higher Tier Only)

At the Higher Tier, you need to recognize that real-life situations are rarely as simple as one thing causing another. Often, multiple factors may interact to affect a result.

For example, a student’s exam result (the response variable) isn't just "caused" by the time spent revising. It is influenced by an interaction of several factors:
• Time spent revising.
• Attendance in class.
• Access to resources (textbooks, internet).
• Difficulty of the exam paper.

When analyzing data, Higher Tier students should be aware that a correlation between "Revision Time" and "Exam Score" might be strong, but it doesn't tell the whole story because these other factors are also at work.

4. Avoiding Common Mistakes

When you are answering exam questions about correlation and causation, keep these tips in mind:

  • Don't jump to conclusions: If a question asks you to "interpret the correlation in context," start by describing the relationship (e.g., "There is a strong positive correlation"). Then, if asked about causation, explain that while they are related, one might not cause the other.
  • Use the term "Underlying Factor": If you suspect a correlation is spurious, use this phrase to explain why. Mention a specific third variable if you can think of one (like "weather," "population growth," or "time").
  • Context is key: Pearson Edexcel exams use real-life data. Use your common sense! If the data shows a correlation between "Number of TVs per household" and "Life Expectancy," think about whether a TV can actually make you live longer, or if "Wealth of the Country" is the true underlying factor.
"Did you know?"

There is a famous spurious correlation between the number of pirates in the world and global warming. As the number of pirates has decreased over the centuries, global temperatures have increased. This is a perfect example of two things happening at the same time by coincidence (or due to the passage of time) without one causing the other!

Summary Key Takeaways

1. Correlation means there is a pattern or link between two variables.
2. Causation means one variable makes the other happen.
3. Correlation does NOT prove causation.
4. Spurious correlations are "false" relationships caused by a hidden underlying factor.
5. (Higher Tier) Most outcomes in the real world are caused by multiple interacting factors rather than just one.