Influential Observations in a Scatter Diagram
- David Gurney
Move the seed value to select a new set of data. Move the class slider to choose a different number of classes. If you move point A or point B, up or down, notice how the regression line follows. If you move point C up or down, the regression line does not move so much. This is why points A and B are called influential observations. When you move A and B, they have a large influence on the regression line. Since moving point C has a much smaller effect on the regression line, point C is NOT an influential observation. Notice that there are large horizontal gaps between points A and B and the rest of the data points. These large horizontal gaps are what cause the points A and B to be influential observations.
Influential observations are worth pointing out in a scatter diagram since they have a big influence on where the regression line is located. Because of their influence, the coordinates of influential outliers should be checked. Updated November 13, 2013.