I explain the difficulty of interpreting the standardized partial coefficient in multiple regression and instead advocate for the use of the semi-partial correlation as an effect size.