Scatter graphs, correlation and cautious predictions · Foundation
Introducido| English | Significado |
|---|---|
| correlation/ˌkɒrɪˈleɪʃn/ | correlación |
| line of best fit/laɪn ɒv best fɪt/ | A line chosen to summarise the overall relationship between two variables in a scatter plot. |
| outlier/ˈaʊtlaɪə/ | An observation unusually far from the rest of the data or from the pattern expected under a model. |
Galton's parents and children, 1886
- In 1886 Francis Galton plotted the heights of 928 adult children against their parents' heights. Taller parents tended to have taller children, but the points formed a cloud, not a line. He had drawn one of the first scatter graphs.
- A scatter graph shows whether two measurements are related.
Correlation
- Positive correlation 相关性: as one goes up, the other tends to go up. Negative: as one goes up, the other tends to go down.
- No correlation: the points show no pattern.
Match each pair to its likely correlation.
Dificultad 2/5 · 140 monedas + un pequeño bono en el primer intento de hoy
Think about how they change together.
Line of best fit
- Draw a straight line through the middle of the points, with about as many above as below.
- Study hours 1–5 and scores 44, 49, 53, 58, 61 fit the line $y = 40 + 4.5x$.
A line of best fit must pass through the first and last points.
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
It goes through the middle of the cloud.
Worked example: predicting
Use $y = 40 + 4.5x$ to predict the score for 3 hours and for 10 hours of study.
- 3 hours: $40 + 13.5 = 53.5$. 3 is inside the data (1 to 5), so this is reliable.
- 10 hours: $40 + 45 = 85$. 10 is far outside the data.
- Judgement: the 10-hour prediction may be wrong — scores may level off, and nobody in the data studied that long.
Scatter graphs, correlation and cautious predictions
Plot paired observations as points, with one variable on each axis
Classify the worked-case statements, then explain the units or invariant that justifies each decision.
Use $y = 40 + 4.5x$ to predict the score for 2.5 hours.
Dificultad 2/5 · 140 monedas + un pequeño bono en el primer intento de hoy
40 + 11.25.
Interpolation and extrapolation
- Interpolation: predicting inside the range of the data — fairly reliable.
- Extrapolation: predicting outside it — unreliable.
Data covers 1 to 5 hours. Which prediction is extrapolation?
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
Outside 1–5.
Outliers
- An outlier is a point far from the others. Ignore it when drawing the line of best fit, but think about why it is there.
- A student who studied 5 hours but scored 20 may have been ill.
What should you do with an outlier when drawing a line of best fit?
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
It does not fit the pattern.
Correlation does not prove that one thing causes the other. Ice-cream sales and sunburn rise together, but ice cream does not cause sunburn: hot weather causes both.
A strong positive correlation proves that one variable causes the other.
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
A third factor may cause both.
Describe correlation as positive, negative or none. Draw a line of best fit through the middle of the points, ignoring outliers. Predictions inside the data are more reliable than outside it, and correlation is not causation.
A scatter graph of 20 towns shows a positive correlation between the number of shops and the number of crimes. A councillor says "Shops cause crime." Give a better explanation.
Interpreting correlation.