Sample partial autocorrelations are one of the main statistical tools of time series analysis. They are especially useful in identifying the order of an AR(p) process. In this note, we show via a simulation experiment that normalizing each sample partial autocorrelation with Anderson's (1993a)...