Mathematics · Year 13

Active learning ideas

Product Moment Correlation Coefficient

Active learning works for this topic because students often overlook how small deviations in data affect the PMCC. Engaging with real datasets helps them connect the abstract formula to tangible patterns they can see and question.

• National Curriculum Attainment Targets
• A-Level: Mathematics - Statistical Hypothesis Testing

• 30–50 min
• Pairs → Whole Class
• 4 activities

Activity 01

• Case Study Analysis
• 45 min · Pairs

Data Collection Challenge: Heights and Shoe Sizes

Pairs measure heights and shoe sizes of classmates, enter data into spreadsheets, plot scatter diagrams, and calculate PMCC. They interpret r and predict changes for new data points. Groups then share findings and compare with class averages.

Explain what a high correlation coefficient indicates about the relationship between two variables.

Facilitation Tip

During the Data Collection Challenge, circulate to ensure students measure height and shoe size accurately and record their data in a shared table for comparison.

What to look for

Provide students with three scatter diagrams showing different types of relationships (positive linear, negative linear, no linear). Ask them to estimate the PMCC for each and write one sentence justifying their estimate based on the visual pattern.

• Analyze
• Evaluate
• Create
• Decision-Making
• Self-Management

Activity 02

• Case Study Analysis
• 35 min · Small Groups

Outlier Investigation: Modified Datasets

Small groups receive identical paired datasets, then alter one by adding an outlier. They recalculate PMCC before and after, plot both scatters, and discuss impact on interpretation. Present changes to the class.

Analyze the difference between correlation and causation.

Facilitation Tip

In the Outlier Investigation, ask students to recalculate r with and without the outlier, then compare results to see how one point changes the strength of association.

What to look for

Present the statement: 'A study found a strong positive correlation between ice cream sales and drowning incidents. Therefore, eating ice cream causes people to drown.' Ask students to explain why this conclusion is flawed, referencing the difference between correlation and causation.

• Analyze
• Evaluate
• Create
• Decision-Making
• Self-Management

Activity 03

• Case Study Analysis
• 50 min · Whole Class

Correlation vs Causation Debate: Real Scenarios

Whole class reviews three scenarios with high PMCC (ice cream sales and drownings, shoe size and reading ability). In small groups, brainstorm causal explanations, then debate as a class why association differs from cause. Vote on strongest arguments.

Evaluate the strength and direction of a linear relationship based on the PMCC value.

Facilitation Tip

For the Correlation vs Causation Debate, provide at least two datasets per group so they practice distinguishing correlation from causation with evidence.

What to look for

In pairs, students calculate the PMCC for a small dataset. They then swap their calculated r-value and interpretation with another pair. The receiving pair must critique the interpretation, checking if it accurately reflects the strength and direction indicated by the r-value and if it avoids causal language.

• Analyze
• Evaluate
• Create
• Decision-Making
• Self-Management

Activity 04

• Case Study Analysis
• 30 min · Individual

Spreadsheet Simulation: Variable Relationships

Individuals use Excel to generate random paired data with varying r strengths (-0.9 to 0.9). They plot scatters, compute PMCC automatically via formula, and adjust data to target specific r values. Share screenshots in a class gallery.

Explain what a high correlation coefficient indicates about the relationship between two variables.

Facilitation Tip

In the Spreadsheet Simulation, demonstrate how changing one variable’s spread affects r before letting students explore independently.

What to look for

Provide students with three scatter diagrams showing different types of relationships (positive linear, negative linear, no linear). Ask them to estimate the PMCC for each and write one sentence justifying their estimate based on the visual pattern.

• Analyze
• Evaluate
• Create
• Decision-Making
• Self-Management

A few notes on teaching this unit

Experienced teachers approach this topic by balancing calculation practice with critical interpretation. They avoid rushing through the formula without addressing its limitations, using activities to reveal why r is not a universal measure. Research shows that students grasp correlation better when they create and manipulate data themselves rather than passively receiving pre-made scatterplots.

Successful learning looks like students confidently calculating r and interpreting its meaning in context. They should also articulate why correlation does not imply causation and recognize when linear measures miss non-linear trends.

---

Watch Out for These Misconceptions

• During the Correlation vs Causation Debate, watch for students who claim a high PMCC proves causation.

  Use the real scenario datasets in the debate to redirect students to look for lurking variables. Ask them to brainstorm alternative explanations and test each one using the data provided.

• During the Data Collection Challenge, watch for students who assume the scatterplot’s pattern matches the PMCC without plotting.

  Have students plot their height and shoe size data first, then calculate r. Ask them to compare the visual trend to their calculated value to see if non-linear patterns appear.

• During the Outlier Investigation, watch for students who think a single outlier cannot change the PMCC much.

  Ask students to recalculate r after removing the outlier and discuss how the value shifts. Use their calculations to emphasize that r is sensitive to extreme points.

---

Methods used in this brief

• Case Study Analysis