Year 9AC9M9ST04
Scatter Plots and Correlation
Choose appropriate forms of display or visualisation for a given type of data; justify selections and interpret displays for a given context.
A scatter plot shows two things about each person or item at once. The pattern tells you whether they are linked.
Method
Positive correlation
as x↑, y↑ Points rise from left to right.
Negative correlation
as x↑, y↓ Points fall from left to right.
No correlation
no pattern Points scattered with no clear direction.
Line of best fit
A straight line drawn through the middle of the pattern.
Worked example
e.g. hours studied against test score for 8 people - Plot each pair as one point(1;42), (2;48), …
- Look at the direction. These rise from left to rightpositive correlation
- Draw a line of best fit through the middle of the points. Roughly as many above as belowy≈6x+36
- Use the line to predict. At 4.5 hours, read off about 63y≈63
- Say how strong it is. Points close to the line means a strong linkstrong positive correlation
Common mistakes
WrongCorrelation means one thing causes the other RightCorrelation only means they move together Ice cream sales and sunburn rise together, but ice cream does not cause sunburn. Hot weather causes both. Never claim cause from a scatter plot alone.
WrongJoining the points dot to dot RightOne straight line through the middle The line of best fit is a single straight line showing the trend, not a zigzag connecting every point.
WrongPredicting far outside the data RightOnly predict inside the range you measured The line may not continue the same way. Predicting a score for 40 hours of study from data that stops at 8 is not reliable.
Practice
Describe the correlation you would expect:
1
Hours of exercise and resting heart rate AnswerNegative 2
Height and shoe size AnswerPositive 3
Shoe size and test score AnswerNone 4
Age of a car and its value AnswerNegative 5
Number of pages and weight of a book AnswerPositive 6
Using y=6x+36, predict the score for x=5 7
Using y=6x+36, predict the score for x=2 8
A plot shows points falling steeply, all close to the line. Describe it AnswerStrong negative correlation 9
Why can you not say more study CAUSES higher marks from this plot? AnswerCorrelation is not causation — something else may explain both 10
Would you use this line to predict a score for 30 hours? AnswerNo — that is far outside the measured range