Investigating the relationship between two kinds of data
Marking pairs of values (xᵢ, yᵢ) of two quantities as points in the coordinate plane gives a scatter plot. If the points rise from left to right the relationship is positive (as x grows, y grows too), if they fall it is negative, and if they are scattered without pattern there is no relationship. The closer the points lie to a straight line, the stronger the linear relationship. The correlation coefficient measures this with a number: r = Σ(x − x̄)(y − ȳ) / √(Σ(x − x̄)² · Σ(y − ȳ)²). Its properties: −1 ≤ r ≤ 1; its sign gives the direction; |r| = 1 means all points lie exactly on one line, and r = 0 means no linear relationship. Roughly, |r| ≥ 0.87 is strong, 0.5 ≤ |r| < 0.87 moderate and |r| < 0.5 weak. Note: a relationship does not mean cause and effect — both quantities may depend on a third factor.
| Points | Relationship |
|---|---|
| rise | positive (r > 0) |
| fall | negative (r < 0) |
| scattered | none (r ≈ 0) |
Air temperature falls as height increases. A rounded table: at heights of 0, 1, 2 and 3 km the temperatures are 15, 9, 2 and −4 °C. The points lie almost on a straight line with r ≈ −0.9995: a strong negative relationship. But correlation alone does not prove a cause; here the cause lies in how the air is heated and how thin it becomes with height.
Meteorologists look for relationships in observations, medical researchers study the link between a factor and an illness, and economists analyse whether two indicators change together.
Take six books or notebooks, measure the number of pages and the thickness of each, and plot the points. Say in words which way the points slope and how strong the relationship looks.
“Height and arm span”: measure the height and the arm span of 8 classmates, draw a scatter plot, state the direction and strength of the relationship, and then compute r with a table (use a calculator if needed).