Today is rung 5 of 6 on this topic
Scatter plots, correlation and the line of best fit
Can do: Plots bivariate data, describes the association, and knows correlation isn't causation.
Why it matters
It's the statistical literacy that matters most in adult life, and the causation error is everywhere in the news.
What it looks like when it's wobbly
Reads any correlation as proof of cause, and draws a line of best fit through the first and last points only.
⏱️ Two-minute check
"Ice cream sales correlate with drownings. Does ice cream cause drowning? What's really going on?"
Solid looks like: No, and identifies a third factor — hot weather.
The full activity for this topic · 30 minutes
Collect and Plot Real Data
- 1Collect fifteen paired data points from real people or real records.
- 2Plot them and describe the pattern in words before drawing anything.
- 3Draw a line of best fit by eye, with roughly equal points either side.
- 4Estimate the slope and say what it means in the situation.
- 5Propose a third factor that could explain the association and discuss how you'd test it.
If it's too hard
Ten points, describe the direction and strength only.
If it's too easy
Use the line to predict, then discuss the danger of extrapolating beyond the data.
Say this
"They move together — fine. What third thing could be causing both?"