under the classical errors-in-variables (cev) assumption, the measurement error means that the estimated ols effect, on average, is closer to zero than the true effect.