Fig. 3
From: Using marginal standardisation to estimate relative risk without dichotomising continuous outcomes
Quantile-quantile plot of residuals from linear models applied to simulated datasets when α1 = 0. Panel (a) and (b) plot the residuals from linear models with normal errors. Panel (c) and (d) plot the residuals from linear models with logistic errors. Gray lines plot quantile-quantile lines for 1000 simulations and black line is the 45-degree line. Normal and logistic linear models mean linear model with the error terms assumed to have normal and logistic distribution respectively