Reduced-scale mock-ups had been developed to assess and regulate smoldering ignition resistance of residential upholstered furniture (RUF). However, there are limited data available on the effect of the mock-up test configuration on smoldering propensity and how it affects the degree of correlation with full-scale performance. In this work, the smoldering propensity for standard mock-ups (SMs, where the foam is in contact with a support frame) and modified mock-ups (MMs, where an air gap is introduced between the foam and the frame), were computationally simulated and compared to experimental results. Experimental data showed an up to a four-fold increase in smoldering mass loss in a MM as compared to a SM for a test time of 35 min with a 51 mm-thick foam. The model results indicated that the buoyant airflow at the bottom of the mock-up was enhanced in the MM, giving rise to a higher foam oxidation rate, a higher peak smoldering temperature and higher mass loss rate as compared to the SM, and; that oxygen supply was dominated by diffusion-driven transport from the boundaries in proximity of the heating source in the SM. Additionally, the effects of foam thickness on smoldering propensity were studied experimentally and numerically with foam thicknesses of 51 mm (2 inch) and 76 mm (3 inch). With an increase in the foam thickness, the smoldering propensity is weakened in the MM but enhanced in the SM. The model was able to predict the ranking of smoldering propensity quantified by the mass loss in experiments: ML(t)SM2 < ML(t)SM3 < ML(t)MM3 < ML(t)MM2, where MLSM2 is the mass loss (ML) with a 51 mm-thick foam in SM, MLSM3 is the ML with a 76 mm-thick foam in SM, MLMM3 is the ML with a 76 mm-thick foam in MM and MLMM2 is the ML with a 51 mm-thick foam in MM. These results indicate that reduced-scale tests based on SM2 tends to underpredict smoldering propensity and that MM2 may offer a near-worst-case scenario, useful to identify the This publication is available free of charge from: https://doi.org/10.6028/NIST.TN.2087 upholstery materials that prevent most smoldering ignitions independent of the construction and geometry of the actual furniture.