Experimental Designs for Estimating the Correlation between Two Destructively Tested Variables

Abstract A specimen can have two or more important strength properties. These could be its bending strength and tensile strength. Assuming joint normality, we study experimental designs for estimating the correlation coefficient. The specimen is proof loaded—say, in bending. If it does not fail, it is loaded in tension to failure. Optimal proof loads are determined. Large sample properties of the maximum likelihood estimator are also given.