Historical Context & Motivation
People have always tried to figure out why things happen. For thousands of years, humans noticed patterns in the world around them. If it rained after a rooster crowed, did the rooster cause the rain? Of course not! But figuring out the difference between a pattern (things happening together) and a cause (one thing making another happen) took scientists a long time.
So here's the big question this lesson answers: When two things seem connected, how do we know if one actually causes the other? This is one of the most important ideas in all of statistics.
Core Principles & Definitions
Before we dive in, let's get clear on the key vocabulary. These words will come up again and again.
Correlation
Causation
Lurking Variable
Coincidence
Visual Explanation — Types of Correlation
A scatter plot (a graph with dots) is the best way to see correlation. Each dot represents one piece of data. The pattern the dots make tells you the type of correlation.
Look at the dots in each graph. In the left graph, as hours studied increases, test scores also increase. That's a positive correlation. In the middle graph, as TV hours increase, exercise time decreases. That's a negative correlation. In the right graph, shoe size and favorite number have nothing to do with each other. The dots are scattered everywhere with no correlation.
How Correlation Tricks Us
Correlation does not tell you why two things are connected. There are actually three possible explanations when you see a correlation.
Explanation 1: Direct Causation
Sometimes A really does cause B. Practicing piano more causes you to play better. Exercising causes your heart rate to go up. These are real cause-and-effect relationships.
Explanation 2: A Lurking Variable
Here's a famous example: ice cream sales and drowning rates both go up at the same time. Does ice cream cause drowning? No way! The lurking variable is hot weather. When it's hot, more people buy ice cream AND more people go swimming. Hot weather is the hidden cause behind both.
Explanation 3: Pure Coincidence
Sometimes two things just happen to move together by random chance. For example, the number of movies Nicolas Cage appeared in correlated with the number of people who drowned in swimming pools in certain years. That's obviously a coincidence — there's no connection at all!
Describing the Strength of Correlation
Not all correlations are equally strong. Some are very tight patterns, and others are loose. We describe correlation using two things: its direction (positive or negative) and its strength (strong, moderate, or weak).
| Strength | What the Scatter Plot Looks Like | Example |
|---|---|---|
| Strong | Dots are packed tightly around a line | Age and height in kids (strong positive) |
| Moderate | Dots follow a general trend but are more spread out | Outdoor temperature and hot cocoa sales (moderate negative) |
| Weak | Dots barely show a pattern | Number of pets and math grade (weak or none) |
| None | Dots are scattered randomly with no trend | Shoe size and favorite color (no correlation) |
A key point: even a very strong correlation does not prove causation. Remember the ice cream and drowning example? That correlation is quite strong, but ice cream still doesn't cause drowning!
Worked Example — Analyzing a Scatter Plot
Let's walk through an example step by step. A student collects data on the number of hours classmates spend on social media per day and their sleep hours per night.
| Student | Social Media (hrs) | Sleep (hrs) |
|---|---|---|
| A | 1 | 9.5 |
| B | 2 | 9.0 |
| C | 3 | 8.0 |
| D | 4 | 7.5 |
| E | 5 | 7.0 |
| F | 6 | 6.5 |
Common Mistakes & How to Avoid Them
| ❌ Common Mistake | ✅ What to Do Instead |
|---|---|
| "Ice cream causes drowning because the data shows they go up together." | "Ice cream sales and drowning are correlated, possibly because both increase in hot weather." |
| "There is no connection between these two things" (just because causation isn't proven). | "There is a correlation, but more research is needed to determine if one causes the other." |
| Confusing the direction: saying positive when the trend goes down. | Check: when X goes up, does Y go up (positive) or down (negative)? |
| Saying a weak correlation means the variables are definitely not related. | A weak correlation could still be meaningful — it just means the pattern isn't strong. |
Connection to Advanced Statistics
Right now you're learning to describe correlation using words like "strong," "weak," "positive," and "negative." In future math and science courses, you'll learn to put a number on it!
| What You Learn Now | What Comes Next |
|---|---|
| Describe correlation as positive, negative, or none | Calculate the correlation coefficient (r), a number between −1 and +1 |
| Describe correlation as strong, moderate, or weak | Use r² to measure how much one variable explains the other |
| Explain why correlation ≠ causation | Design controlled experiments and studies to test for causation |
| Identify lurking variables in examples | Use regression analysis to model relationships mathematically |
The good news is that the big idea you're learning today — correlation does not imply causation — stays important forever. Even scientists with advanced degrees remind themselves of this rule every day.
Practice Problems
Lesson Summary
Correlation means two variables change together in a pattern. A positive correlation means both go up together. A negative correlation means one goes up while the other goes down. We describe the strength as strong, moderate, or weak based on how tightly the data follows the pattern.
Causation means one thing directly makes another happen. Correlation does not imply causation — two things can be correlated because of a lurking variable (a hidden factor affecting both) or pure coincidence. To prove causation, scientists use controlled experiments that isolate one variable at a time.