Fig. 2From: Using marginal standardisation to estimate relative risk without dichotomising continuous outcomesSimulation results for type 1 error and power when α1 = − 0.15 and α1 = − 0.3. Panel (a) and (c) plot the simulation results for \( \hat{\beta_1} \) and panel (b) and (d) plot the simulation results for \( \hat{\mathrm{RR}} \). Dashed lines are plotted at 0.05 and 1 for Type 1 error and Power, corresponding to 0.05 type 1 error and 100% power respectively. Normal and logistic linear model means linear model with the error terms assumed to have normal and logistic distribution respectivelyBack to article page