Runback ice accretions present a unique problem in iced-airfoil aerodynamics in that the airfoil typically has a clean leading edge prior to the ice shape itself. In order to investigate the aerodynamic effects of runback ice accretions simulated ice shapes were scaled, from accretions obtained in testing at the NASA Glenn Icing Research Tunnel, for testing in the UIUC 3’x4’ subsonic wind tunnel. Simple geometric scaling, based on airfoil chord, as well as boundary-layer scaling, based on estimated boundary-layer thickness, was used. The NACA 3415 and the NACA 23102 were tested at a Reynolds number of 1.8x10⁶ and Mach number of 0.18 with and without the simulated ice shapes attached. Simple 2-D simulations were constructed for the test as well as 3-D simulations that used multiple substrate layers as well as roughness to simulate the features of the full-scale accretion. Significant penalties due to runback accretions were identified. The NACA 3415 experienced a 20% reduction in C_lmax and no loss in stalling angle of attack due to the 3-D warm hold simulation while the NACA 23012 experienced a 25% reduction in C_lmax and a loss of two degrees in stalling angle of attack. The 3-D cold hold simulation caused a 40% loss in C_lmax and a two degree reduction in stalling angle of attack for the NACA 3415 while the same ice shape caused a 55% reduction in C_lmax and a seven degree reduction in stalling angle of attack for the NACA 23012. Geometrically-scaled 2-D simulations of the warm hold accretions were found to enhance the lift performance of the NACA 3415 and had little effect on the NACA 23012. The cause of this phenomenon is hypothesized to be a combination of energizing the boundary-layer and the pressure distribution established by the presence of the ridge shape. The boundary-layer-scaled equivalent of that ice shape was observed to reverse this phenomenon. Boundary-layer calculations indicated that the geometrically-scaled ice shape was approximately the same height as the local boundary thickness at angles of attack near stall. The boundary-layer-scaled ice shapes were observed to cause greater penalties than both the 2-D and 3-D geometrically-scaled ice shapes. Boundary-layer-scaled 3-D ice shapes remain to be tested. It should be noted that data regarding the effect of full-scale runback ice accretions are not available at this time. Therefore, it is difficult to judge which scaling method is appropriate. However, it is clear from this work that geometric scaling may not be sufficient for scaling runback-type ice accretions for aerodynamic testing.