Hamilton, C. W.
Associate fellow, Battelle Memorial Inst., Columbus, Ohio
Pages: 11 Published: Jan 1966
The purpose of this paper is to outline statistical decision procedures that appear to be well-suited to the problem of standard qualitative decision procedures. It is pointed out that such procedures have been previously developed to contend with various mechanical and electronic reliability problems. Thus, it is only required to adapt these procedures to the particular feature of the building materials approval problem. Specifically, a Bayesian decision model is proposed. The essential features of the model are that it provides for: (1) the formal inclusion of laboratory test data, field test data, and “subjective” data in a single model and (2) the successive “updating” of the model as new data are made available. On the basis of this, model formulation procedures are recommended for generating laboratory and field test data to provide a sufficient basis for the decision problem. The procedure is illustrated with an example drawn from a research program for FHA on adhesive-bonded structural components.
construction materials, acceptance, life testing, reliability, Bayes theorem, probability, statistical inference, validity
Paper ID: STP44646S