Figure 1.

The compound bias in the two predictors due to measurement error in one predictor.

Note. ${\rho }_X=1.00$ in all conditions and is the reliability of $X$. ${\rho }_{\mathrm{XW}}$ is the correlation between the two predictors had they been measured without error. $R^2$ is the proportion of variance in the outcome explained by the two predictors if both had been measured without error.