The methodology is based on ordering information within previously ill-structured problem environments using two matrix representations similar to those applied to program planning. The first matrix representation defines component interrelationships and captures the objectives of the decisionmaker regarding the performance measurement problem. The second representation summarizes the properties of alternative performance measure instruments in a manner that emphasizes their differences and reduces the task of comparisons across the alternative set. These steps allow for the application of multiattribute decision technology over decomposed ranges of the problem, yielding a simplified assessment of decisionmaker preferences. The study interprets different forms of component and activity dependence in a decision theory context, and preference functions are suggested for the various independence conditions. 1 figure, 1 table, mathematical equations, and 15 references.