Abstract
Efforts to measure performance difference attributable to test-to-test variability of a particular dummy, dummy-to-dummy variability of dummies of the same manufacture, and variability from one make of dummy to another are described. Comparisons were conducted under a number of restraint conditions.