A scatter graph usually shows the relationship between two variables which can also be described as correlation (which means a general trend in the data items, however where there are exceptions to the rules, they are called outliers).
The types of correlation in scatter graphs can be described as:
- positive: one variable increases with the other
- negative: one variable increases and the other decreases
- zero: there is no relationship between the variables
Point to note:
Correlation can be weak or strong therefore no correlation is not the same as having no relationship between variables, but rather no linear relationship.
The line of best fit
A line of best fit is drawn on scatter graphs by eye which also ensures that there are equal number of points above and below the line.
Where a line of best fit has points closer to the line, the correlation is stronger.
Furthermore, if the gradient is positive, the correlation is positive and if the gradient is negative, then the correlation is negative.
The following points should be considered when drawing a line of best fit:
- It does not necessarily have to pass through the origin
- It can be used to predict one variable from another
- It cannot be used for predictions outside the range of data
- The equation of the line of best fit can be worked out using the gradient and intercept. This can then be used to work out specific scenarios.