Mathematics · Year 10 · Statistical Investigations and Data Analysis · Term 4

Correlation and Causation

Understanding the difference between correlation and causation in bivariate data.

• ACARA Content Descriptions
• AC9M10ST01

About This Topic

Correlation describes the strength and direction of association between two variables in bivariate data, often shown in scatterplots and measured by coefficients from -1 to 1. Year 10 students distinguish this from causation, where one variable directly causes changes in another. They analyze patterns, such as the link between study hours and test scores, and recognize that correlation alone cannot prove cause and effect.

Aligned with AC9M10ST01, this topic strengthens statistical investigations by introducing confounding variables, like exercise influencing both diet and health outcomes. Students evaluate real-world data from sources like Australian Bureau of Statistics reports on crime rates and ice cream sales, both peaking in summer due to temperature. These exercises build data literacy and skepticism toward oversimplified claims in news or advertising.

Active learning suits this topic well. When students debate causal claims from graphs, hunt for lurking variables in datasets, or generate their own bivariate examples, they internalize the distinction through trial and error. Group analysis uncovers shared errors, while peer teaching reinforces precise reasoning.

Key Questions

1. Explain why correlation does not necessarily imply causation between two variables?
2. Analyze real-world examples where correlation is mistaken for causation.
3. Justify the importance of considering confounding variables in statistical analysis.

Learning Objectives

• Explain why a strong correlation between two variables does not automatically mean one causes the other.
• Analyze real-world scenarios to identify instances where correlation is incorrectly interpreted as causation.
• Evaluate the role of confounding variables in obscuring or creating apparent relationships in bivariate data.
• Critique statistical claims made in media or advertising by distinguishing between correlation and causation.

Before You Start

Bivariate Data and Scatterplots

Why: Students need to be able to interpret scatterplots and understand how to visually represent the relationship between two quantitative variables.

Data Interpretation

Why: Understanding how to read and interpret tables and graphs is essential for analyzing statistical information and identifying patterns.

Key Vocabulary

Correlation	A statistical measure that describes the extent to which two variables change together. It indicates the strength and direction of a linear relationship.
Causation	A relationship where a change in one variable directly produces or brings about a change in another variable.
Confounding Variable	An unmeasured variable that influences both the independent and dependent variables, potentially creating a spurious correlation.
Spurious Correlation	A correlation between two variables that appears to be related but is actually due to coincidence or the influence of a third, unobserved factor.

Watch Out for These Misconceptions

Common Misconception

A strong positive correlation always means one variable causes the other.

What to Teach Instead

Correlation shows association, but causation requires evidence like controlled experiments. Hands-on graphing of confounders, such as temperature in summer sales data, helps students visualize lurking influences. Peer debates expose this flaw through counterexamples.

Common Misconception

No correlation means no causal relationship exists.

What to Teach Instead

Absence of correlation in observed data does not rule out causation, especially with confounders present. Group investigations of datasets reveal hidden patterns, while role-playing scenarios build understanding that correlation is necessary but not sufficient for causation claims.

Common Misconception

Correlation direction determines which variable causes the other.

What to Teach Instead

Direction indicates association type, not causal order. Students clarify this by swapping axes in pair scatterplot activities, prompting discussions on reverse causation or bidirectionality. Collaborative hypothesis testing strengthens analytical precision.

Active Learning Ideas

Jigsaw: Spurious Correlations

Assign small groups one real-world example, such as cheese consumption and bed linen tangles. Groups research data, identify confounders, and create posters. Regroup into expert jigsaws to teach peers, followed by class vote on most convincing case. Conclude with shared scatterplot sketches.

45 min·Small Groups

Scatterplot Debates: Pairs Challenge Claims

Pairs receive a scatterplot with a causal headline, like 'More parks cause lower obesity.' They list evidence for and against causation, then debate with another pair. Switch roles and vote on strongest arguments using correlation coefficient criteria.

30 min·Pairs

Data Detective Hunt: Whole Class Analysis

Project three datasets from Australian sources, such as rainfall and crop yields. Class brainstorms causal hypotheses in a shared digital whiteboard, then identifies confounders via think-pair-share. Tally votes and discuss experimental design needs.

35 min·Whole Class

Simulation Stations: Confounding Variables

Set up stations with props: one for ice cream/shark attacks (weather confounder), another for homework/grades (parental involvement). Groups rotate, model with graphs, and predict coefficient changes if confounder is controlled. Share insights in plenary.

40 min·Small Groups

Real-World Connections

• Public health officials in Sydney must be careful not to assume that increased ice cream sales directly cause higher rates of drowning. Both are correlated with warmer weather, which is the confounding variable.
• Market researchers analyzing sales data for a new smartphone app might observe a correlation between downloads and user engagement. They need to investigate if other factors, like targeted advertising campaigns, are the true cause of engagement, not just the download itself.
• Economists studying the relationship between education levels and income in Australia need to account for factors like socioeconomic background and geographic location, which can influence both education attainment and earning potential.

Assessment Ideas

Discussion Prompt

Present students with a graph showing a strong positive correlation between the number of firefighters at a fire and the amount of damage caused. Ask: 'Does this graph prove that sending more firefighters causes more damage? Why or why not? What other factors might explain this relationship?'

Quick Check

Provide students with three brief statements, each describing a correlation. For example: 'A study shows that students who eat breakfast perform better on tests.' Ask students to write one sentence for each statement explaining if it demonstrates correlation or causation, and to identify a potential confounding variable if applicable.

Exit Ticket

Ask students to define correlation and causation in their own words. Then, have them provide one example of a correlation they have observed (or heard about) and explain why it might not be a causal relationship.

Frequently Asked Questions

What are effective real-world examples for teaching correlation vs causation?

Use Australian data like urban sprawl correlating with stork populations (both rise with development, no causation) or sunscreen sales and ice cream (summer confounder). Students plot these, calculate coefficients, and debate headlines. This grounds abstract ideas in familiar contexts, boosting engagement and retention across diverse learners.

How do I explain confounding variables to Year 10 students?

Describe confounders as hidden third variables linking the two observed ones, like fitness level affecting both running distance and heart health. Draw causal diagrams on boards, then have students add confounders to given scenarios. Real data from ABS health surveys makes it concrete, helping students predict coefficient shifts when controlling for the third variable.

How can active learning help students understand correlation and causation?

Active methods like data hunts and debates make students detectives of their own misconceptions. Collecting school-based bivariate data, such as sleep hours and sports performance, lets them test claims firsthand. Group critiques of causal fallacies build confidence, while simulations reveal confounders dynamically, far surpassing passive lectures for deep comprehension.

Why is distinguishing correlation from causation key in Year 10 Maths?

It fulfills AC9M10ST01 by honing critical analysis of bivariate data, vital for evidence-based decisions in policy, health, and environment. Students learn to question spurious links in media, like vaccine correlations with autism rates debunked by confounders. This skill transfers to senior stats and real life, fostering responsible citizenship.