a set of techniques that allows one to determine the degree to which specific data points affect the overall result of a statistical procedure. For example, in regression analysis it is used to examine which observations have a disproportionate influence on the proposed regression equation.