AI generatedStatistics

Correlation, outliers and causation

  • Plot paired observations as separate crosses. Do not join the points.
  • A best-fit line follows the overall trend, with points balanced around it.
  • Predictions are estimates. Extrapolation is less secure, and correlation alone does not establish causation.

Example 1

Describe the direction and strength of the correlation.

Coordinate graph12345678910115101520250
As increases, generally decreases.The points lie close to a straight-line pattern.
Strong negative correlation.

Example 2

A study of 41 observations finds positive correlation between shoe size and reading score. Does this prove that one causes the other?

No. A third factor, such as age in a sample of children, could affect both. Correlation alone does not establish a cause.

Example 3

Identify the outlier and explain your choice.

Coordinate graph12345678910115101520250R

R lies well away from the overall linear pattern. Its -value alone does not make it an outlier.

Point R is an outlier.

Your turn

Question 1

Describe the direction and strength of the correlation.

Coordinate graph123456789101151015202530350
Check answer
Strong positive correlation.
Question 2

Describe the direction and strength of the correlation.

Coordinate graph12345678910115101520250
Check answer
Strong negative correlation.
Question 3

A study of 44 observations finds positive correlation between shoe size and reading score. Does this prove that one causes the other?

Check answer
No. A third factor, such as age in a sample of children, could affect both. Correlation alone does not establish a cause.
Question 4

A study of 45 observations finds positive correlation between umbrella sales and traffic delays. Does this prove that one causes the other?

Check answer
No. A third factor, such as heavy rain, could affect both. Correlation alone does not establish a cause.
Question 5

Identify the outlier and explain your choice.

Coordinate graph123456789101151015202530350R
Check answer
Point R is an outlier.