IB Mathematics: Analysis and Approaches · Topic 4 · Statistics and probability
4.4 Correlation and regression
Interpret scatter diagrams, classify correlation, and find the equation of a regression line from data.
1/5A scatter diagram shows how two variables relate
- A scatter diagram shows how two variables relate
- Correlation describes the direction and strength of a linear link
- The regression line is the best-fitting straight line
- The mean point and gradient give the equation
- Predictions are reliable only within the range of the data
- Check your recall
- Practice: 10 questions
Part 1 of 5: A scatter diagram shows how two variables relate
A teacher records the hours each student spent revising and the student’s test mark. Every student gives a pair of values. Data in pairs like this are bivariate data.
Plot each pair as a point, with the explaining variable on the horizontal axis. The result is a scatter diagram. It shows a pattern before you calculate anything.
Here, revision time is the variable we use to explain the mark . Look for the overall direction, how tightly the points cluster, and any point far from the rest.
Key terms · tap one for its definition
bivariate data
Bivariate data are data in which two variables are measured for each item, giving pairs of values .
scatter diagram
A scatter diagram is a graph in which each pair of values is plotted as one point.
Part 2 of 5: Correlation describes the direction and strength of a linear link
Points that rise from left to right show that tends to increase as increases. Points that fall show the opposite. This kind of linear association between two variables is called correlation.
The correlation coefficient (Pearson’s product-moment) puts a number on it, with . The sign gives the direction. Values near mean the points lie close to a line.
Reading the pattern and the value of r
- Pattern
- Points rise, close to a line
- Value of r
- close to
- Name
- strong positive
- Pattern
- Points fall, close to a line
- Value of r
- close to -1
- Name
- strong negative
- Pattern
- Points loosely follow a trend
- Value of r
- nearer 0
- Name
- weak
- Pattern
- No visible pattern
- Value of r
- about 0
- Name
- none
The coefficient measures only linear association. A clear curve can still give close to 0.
Key terms · tap one for its definition
correlation
Correlation is the strength and direction of the linear association between two variables.
correlation coefficient
The correlation coefficient is a number from -1 to measuring the strength and direction of linear association.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: A · Worked explanation
Part 3 of 5: The regression line is the best-fitting straight line
To make predictions, we draw one straight line through the points. The regression line of on is the line that fits best. It minimises the sum of the squared vertical distances from the points to the line.
The line always passes through the mean point . Its gradient uses the covariance and the variance of the -values, .
The equation is , where and . A calculator’s linear regression function gives the same line. The vertical distances are the gaps between each observed and the value the line predicts.
Key terms · tap one for its definition
regression line
The regression line of on is the line that minimises the sum of the squared vertical distances to the points.
covariance
The covariance is the mean of the products minus the product of the means, .
Part 4 of 5: The mean point and gradient give the equation
Now apply the formula to data. Each student has a revision time in hours and a mark . Work out the two means first. Then find the sums that the covariance and variance need.
Worked example
Finding a regression line
- Means: and .
- Sums: and .
- Covariance: .
- Variance of : .
- Gradient: , so .
Answer: The regression line is .
For these data , a strong positive correlation. The gradient 0.8 means each extra hour of revision adds 0.8 marks on average. The intercept 1.8 is the predicted mark for no revision. The line rises from left to right, matching the positive correlation. Points above the line scored more than predicted, and points below scored less.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: B · Worked explanation
Part 5 of 5: Predictions are reliable only within the range of the data
A regression line predicts for a given . Predicting inside the range of the data is interpolation. It is usually reliable when is close to 1.
Predicting outside that range is extrapolation. The pattern may not continue, so the result is less reliable. For in the example, the line gives marks.
The line of on predicts from . Do not use it to predict from a given . A different line is needed for that.
Key terms · tap one for its definition
interpolation
Interpolation is using a regression line to predict for a value of inside the range of the data.
extrapolation
Extrapolation is using a regression line to predict for a value of outside the range of the data.
Check your recall[15 marks]
6 written questions. Answer each one in your head or on paper, then open the mark scheme.
- 1.Distinguish[3 marks]
Distinguish between positive, negative and zero correlation.
Show mark schemeHide mark scheme
Model answer
Positive correlation means tends to increase as increases, with . Negative correlation means tends to decrease as increases, with . Zero correlation means no linear pattern, with near 0.
- 1 mark for describing positive correlation as increasing with
- 1 mark for describing negative correlation as decreasing as increases
- 1 mark for describing zero correlation as no linear pattern
- 2.State[1 mark]
State the range of possible values of the correlation coefficient .
Show mark schemeHide mark scheme
Model answer
.
- 1 mark for stating
- 3.Explain[2 marks]
Explain why a strong correlation between two variables does not prove that one causes the other.
Show mark schemeHide mark scheme
Model answer
Correlation measures only association between the variables. A third factor may drive both, or the link may be a coincidence.
- 1 mark for stating that correlation shows association and not cause
- 1 mark for giving a reason such as a third factor affecting both variables
- 4.Determine[4 marks]
Determine the equation of the regression line of on for the points (1, 3), (2, 5), (3, 6) and (4, 10).
Show mark schemeHide mark scheme
Model answer
and . and . The gradient is . Then , so .
- 1 mark for finding the means and
- 1 mark for finding
- 1 mark for finding and the gradient 2.2
- 1 mark for writing the equation
- 5.Calculate[3 marks]
Calculate the predicted test mark for hours of revision using the regression line , which was found from data with from 2 to 10. Comment on using the line for .
Show mark schemeHide mark scheme
Model answer
For , . The value lies outside the data range, so using it is extrapolation and the prediction is unreliable.
- 1 mark for substituting to obtain 13
- 1 mark for stating that is outside the range of the data
- 1 mark for concluding that the extrapolated prediction is unreliable
- 6.Describe[2 marks]
Describe what the gradient tells you about the regression line , where is a test mark and is revision time in hours.
Show mark schemeHide mark scheme
Model answer
For each extra hour of revision, the predicted mark rises by 1.5 marks on average.
- 1 mark for referring to one extra hour of revision
- 1 mark for stating that the predicted mark increases by 1.5 on average

Practice this topic
10 quick-fire multiple-choice questions, easy to hard. Tap an answer to see if you're right and the worked explanation. (The written questions with mark schemes are under “Check your recall” above.)
0 right0 of 10 answered
Includes the 2 quick checks in the note above.
EasyState[1 mark]
Show answer and explanation
Answer: A · Worked explanation
MediumCalculate[2 marks]
Show answer and explanation
Answer: A · Worked explanation
MediumDetermine[2 marks]
Show answer and explanation
Answer: A · Worked explanation
MediumCalculate[2 marks]
Show answer and explanation
Answer: A · Worked explanation
MediumIdentify[2 marks]
Show answer and explanation
Answer: A · Worked explanation
Hard[2 marks]
Show answer and explanation
Answer: A · Worked explanation
HardDetermine[2 marks]
Show answer and explanation
Answer: A · Worked explanation
Hard[2 marks]
Show answer and explanation
Answer: A · Worked explanation