We introduce a novel measure of performance of polymer composites based on physical parameters whose behavior depends on levels of dopant concentration used during their preparation. The performance measure is based on a joint analysis of measurements of the physical parameters exhibiting non‐trivial correlations that vary across different levels of dopant concentrations. In contrast to traditional multivariate analysis, we treat data from parameter measurements as being obtained from functional parameters, and develop the performance measure based on the joint average function of parameters that encodes the correlation structure. An optimal level of dopant concentration is then ascertained with respect to the performance measure. While the proposed measure is general enough to be applicable to any chosen physical parameters, we demonstrate its utility in the context of assessing performance using microstructural and nonlinear optical parameters. Computing of the measure and optimal dopant concentration are carried out using Monte Carlo sampling, which further facilitates uncertainty quantification.