The studies of the safety of structures must incorporate concepts covering three subareas. A perfect part, where St. Venant's principle is invalid; and errors in design and construction, which occur in both parts. This project is concerned with the identification of such errors in the field and their inclusion, in a blind fashion, into laboratory tests of imperfect parts. A group of experts are used to offer opinions concerning the capacity of the parts after simulated seismic loading, which has not been extended to destruction. The experts are asked to suggest remedial measures as well as verbal opinions of the structural competence. Continuing loading to failure will provide a measure of success of such opinions. Where possible Bayesian statistics will be used, but where the axioms of probability theory are not satisfied, reliance is made on fuzzy set theory.