Previous research has shown that the location of the shoulder belt on the Hybrid III dummy strongly influences the sternum deflection measurement, which could reduce the real-world relevance of crash tests making use of this metric. The objective of the current study was to quantify and adjust for the effect of belt placement on the sternum deflection measurement across a range of vehicle designs. Pre-test belt locations were measured using photographs of US NCAP full width (n=207) and IIHS moderate overlap (n=131) tests. Linear regression was used to estimate the effect of the vertical position of the belt relative to the sternum potentiometer on the measured deflection while controlling for other factors. Statistically significant effects were found for both tests, with a 1 mm increase in the vertical position of the shoulder belt estimated to reduce the peak sternum deflection by 0.13 mm in the NCAP test (SE = 0.013 mm;​ p < 0.001) and by 0.06 mm in the IIHS test (SE = 0.018 mm, p = 0.002). If all tests had belts centered over the sternum potentiometer, the median sternum deflection would be expected to increase 49% (±5%) in the NCAP test and 18% (±6%) in the IIHS test.