We have an association between a dependent and independent variable adjusted for 3 covariates. Our dependent variable is measured via self-report, which has a .47 correlation to objective measures. How can we estimate the impact of this reliability issue on our effect estimate?