This study sought to evaluate the influence of initial belt torso contact (belt gap) and belt fit provided by belt‐positioning booster seats on various kinematic and kinetic outcomes. Frontal crash tests (n=18) were conducted at a peak deceleration of 22.5±1.0 g using the Q‐Series 6‐year‐old (Q6), Q‐Series 10‐year‐old (Q10), and Large Omni Directional Child (LODC10) anthropomorphic test devices (ATDs). Each ATD was evaluated on two highback (HB), three low‐back (LB), and one low‐profile (Low) booster seat, which provided varying initial belt fit and gap conditions. Resultant head, chest, and pelvis accelerations, and HIC15 were similar across boosters. Larger gap boosters produced greater peak lumbar MZ (HB:​ ‐28.3 Nm, LB:​ ‐24.7 Nm) compared to smaller gap boosters (HB:​ ‐13.1 Nm, LB/Low:​ ‐3.5 Nm) for the LODC10 and Q10 on average. Larger gap LB/Low boosters also produced greater axial shoulder rotation (43.3°) compared to smaller gap LB/Low boosters (30.1°). This suggests ATDs on larger gap boosters experienced greater torso rotation and lumbar moment due to lack of initial contact between the belt and inferior torso. While no shoulder belt slip‐off occurred, larger shoulder rotations may indicate propensity for slip‐off in more severe crashes, oblique manoeuvres, or with variations in initial occupant posture.