Recent field data studies have shown that force-limiting belt systems reduce the occurrence of thoracic injuries in frontal crashes relative to standard (not force-limiting) belt systems. Laboratory cadaver tests have also shown reductions in trauma, as well as in chest deflection, associated with a force-limiting belt. On the other hand, tests using anthropomorphic test devices (ATDs) have shown trends indicating increased, decreased, or unchanged chest deflection. This paper attempts to resolve previous experimental studies by comparing the anterior-posterior and lateral chest deflections measured by the THOR and Hybrid III (H-III) dummies over a range of experimental conditions. The analysis involves nineteen 48-km/h and 57-km/h sled tests utilizing force-limiting and standard seat belt systems, both with an air bag. Tests on both the driver side and the passenger side are considered. The purpose of the study is to evaluate whether either dummy could differentiate between the two belt types over a range of sled pulse shapes, seating positions, and belt force limits. The comparison shows that both dummies exhibit significantly (p<0.05) greater maximum chest deflection (Cmax) with the standard belt. In the driver-side tests (3 kN force limit), Cmax for THOR and H-III was 51.6% and 64.4% greater while, in the passenger-side tests (4 kN force limit), Cmax was 16.6% and 30.6% greater. Lateral deflections increased similarly, as did resultant chest acceleration. It is concluded that, over the range of conditions considered here, both dummies consistently ranked the restraint systems appropriately. We conclude by hypothesizing several potential explanations for the different findings among studies in the literature.