Linear combinations of all of the original Variables that maximally explain variability. Useful when the analysis considers effects of many variables at once. By combining the variables into groups, you can reduce the total number in any one analysis.