High leverage

PrepNuggets

LEVEL II

Leverage measures how far an observation is from the average value of the independent variable. High leverage observations are those that are far away from the average and have a large effect on the regression line.

For a particular independent variable, leverage of a particular data point measures the distance between its value and the mean value of that variable. Leverage is normalised between 0 and 1. High leverage points may have a leverage factor closer to 1, while low leverage points are closer to 0.  The higher the leverage, the more influence it can potentially exert on the estimated regression. A useful rule of thumb is that if the leverage measure exceeds this value, then it is a high leverage point and is potentially influential.

See also: Influence analysis, Studentized residual, Cook’s Distance, Influence plot

« Back to Index