The general practice in performance and production efficiency measurement has been to ignore additional products of most transformation processes that can be classified as "undesirable outputs" — which are a subset of the output set. Without the inclusion of these factors, the efficiency evaluation becomes a purely technical measure of the system alone, and does not account for the interaction of the system with the surrounding environment and the impact of policy decisions on the system. In addition, there are also technological dependencies arising due to the relationships between the desirable and the undesirable outputs. One of the analytical tools normally used in efficiency evaluation is Data Envelopment Analysis, DEA. In the course of addressing these problems, a decision-maker encounters multiple and contradictory objectives with respect to the output set. This motivates the exploration of new arenas of measurement of efficiency to facilitate policy decisions and address technological relationships. This research presents five modifications of the traditional DEA technique to give a more realistic and comprehensive score of production efficiency considering both, desirable and undesirable outputs. The models address the following problems: (i) technological dependency between desirable and undesirable outputs; (ii) decision-maker's preferences over inputs, desirable outputs and undesirable output performance and finally (iii) conflicting production objectives through a formulation that uses Goal Programming in conjunction with DEA, a concept known as GoDEA. / Master of Science