This report introduces the concepts of a previously developed methodology which could readily be extended to the field of performance assessment for high-level nuclear waste isolation systems. The methodology incorporates sensitivities previously obtained with the GRESS code into an uncertainty analysis, from which propagated uncertainties in calculated responses may be derived from basic data uncertainties. Following a definition of terms, examples are provided illustrating commonly used conventions for describing the concepts of covariance and sensitivity. Examples of solutions to problems previously encountered in related fields involving uncertainty analysis and use of a generalized linear least-squares adjustment procedure are also presented.