Introduction to Regression Lines

In previous chapters, you learned how to draw a line of best fit by eye. While that is useful, it can be a bit "guess-worky." The Linear Regression Line is the "mathematically perfect" version of a line of best fit. It is a straight line that best represents the data on a scatter diagram.

For the Higher tier, you need to understand the equation of this line and use it to make accurate predictions. Think of it like a mathematical bridge: if you know one value (like how much you study), the regression line helps you predict the other (like your exam score).

The Variables: \(x\) and \(y\)

Before using a regression line, you must identify your variables correctly:

  • Explanatory Variable (\(x\)): Also called the independent variable. This is the variable you think might be causing a change. It always goes on the horizontal \(x\)-axis.
  • Response Variable (\(y\)): Also called the dependent variable. This is the variable that "responds" to the change. It always goes on the vertical \(y\)-axis.

Example: If you are looking at how temperature affects ice cream sales, Temperature is the explanatory (\(x\)) variable, and Sales is the speedy response (\(y\)) variable.

The Equation of the Regression Line

Since the regression line is a straight line, it follows the same formula you use in GCSE Maths:

\(y = mx + c\)

  • \(y\) is the predicted value of the response variable.
  • \(m\) is the gradient (the slope). It tells you how much \(y\) increases or decreases for every 1 unit increase in \(x\).
  • \(x\) is the value of the explanatory variable you are using.
  • \(c\) is the \(y\)-intercept. This is the value of \(y\) when \(x = 0\).

Quick Review: If a regression line is \(y = 2x + 5\), it means for every 1 unit \(x\) goes up, \(y\) goes up by 2. When \(x\) is 0, \(y\) starts at 5.

Drawing the Line: The Double Mean Point

To draw a regression line accurately, it must pass through the double mean point. This is a point made from the average of all \(x\) values and the average of all \(y\) values.

Step 1: Calculate the mean of \(x\), written as \(\bar{x}\).
Step 2: Calculate the mean of \(y\), written as \(\bar{y}\).
Step 3: Plot the point \((\bar{x}, \bar{y})\) on your scatter diagram.
Step 4: Your regression line must pass exactly through this point.

Key Takeaway: The double mean point is the "anchor" for your regression line. If your line doesn't go through it, it's not a correct regression line!

Making Predictions

One of the main reasons we use regression lines is to predict values we don't have in our raw data. You can do this in two ways:

1. Using the Graph

  1. Find the value of \(x\) you are interested in on the horizontal axis.
  2. Move vertically up to hit the regression line.
  3. Move horizontally to the left to read the corresponding \(y\) value on the vertical axis.

2. Using the Equation (More Accurate)

If you are given an equation like \(y = 0.5x + 10\), simply substitute your \(x\) value into the formula.

Example: Predict the value of \(y\) when \(x = 40\).
\(y = 0.5(40) + 10\)
\(y = 20 + 10\)
\(y = 30\)

Interpolation vs. Extrapolation

Not all predictions are equally reliable! You need to know the difference between these two terms:

Interpolation (The "Safe" Zone)

This is making a prediction inside the range of the data you already have. If your data collected goes from \(x = 10\) to \(x = 100\), predicting a value for \(x = 50\) is interpolation.
Reliability: Usually very reliable.

Extrapolation (The "Danger" Zone)

This is making a prediction outside the range of your data. If your data ends at \(x = 100\), trying to predict what happens at \(x = 200\) is extrapolation.
Reliability: Unreliable and dangerous. We cannot be sure the trend continues in the same way outside our observed data. (For example, a person's height increases with age, but if you use a regression line to predict their height at age 80, the math might suggest they will be 4 meters tall!)

Did you know? Spurious correlations (things that seem related but aren't) make extrapolation even riskier. Just because two things look related now doesn't mean they will stay that way forever.

Common Mistakes to Avoid

  • Swapping Axis: Always check that the explanatory variable (\(x\)) is on the bottom. If you swap them, your prediction will be wrong.
  • Ignoring the "Dangers of Extrapolation": In exam questions, if you are asked to comment on the reliability of a prediction outside the data range, you must state that it is "extrapolation and therefore unreliable."
  • Rounding too early: If you are using a calculated equation, keep your numbers accurate until the very end.

Chapter Summary

1. The regression line is a mathematical line of best fit in the form \(y = mx + c\).
2. It must pass through the double mean point \((\bar{x}, \bar{y})\).
3. Interpolation is predicting within the data range (reliable).
4. Extrapolation is predicting outside the data range (unreliable/dangerous).
5. Use the \(x\) variable (explanatory) to predict the \(y\) variable (call it the response).