A series of seventy-two pendulum-type impact tests were performed on six NHTSA Side Impact Dummies (SID) to assess dummy repeatability and reproducibility. A quantity called the Normalized Integral Square Error (NISE) is used to quantify the difference between acceleration responses from repeat tests. Limits for the NISE are developed to define acceptable differences in terms of phase shift, amplitude, and shape. Results indicate that the SID is repeatable in all of the test cases considered and fairly reproducible in 90° lateral impacts although this is not shown conclusively.Before the testing could be performed it was necessary to correct several durability problems with the SID that were identified while early production versions of the dummy were being tested. These modifications are described briefly.