IS Assessment – Evaluating what should be Evaluated: A new Formal Framework to Study Project Representations.
July 13, 2012
Posted by on
Yannick MEILLER and Sylvain BUREAU
Scholarship on the ex-ante evaluation of projects is prolific and offers a whole range of tools to assess projects on financial, technical, strategic and organizational levels. All these analyses start from the assumption that everyone knows and shares knowledge on what is to be assessed. Nevertheless, the system under evaluation is rarely just data, but rather a construction of a representation of a tangible physical reality which can appear under different forms, and the chosen representation may strongly influence the evaluation itself. On the contrary, in the field of automatic decision-making, i.e. Artificial Intelligence, the impact of the representation of issues and the structure of such representations in decision algorithms is widely discussed. Building on this structuring approach, we specifically focus on the initial construction stage of a representation and on its impact on evaluation. In order to grasp these representations, to compare one with another and to study this stage of the evaluation process, we propose a conceptual framework – the space of representations, structured by abstractions and scopes. We propose formal definitions of these concepts. To illustrate and also test our approach, we examine three empirical cases of investment projects related to RFID (Radio Frequency Identification) applications in hospitals. This field study shows that small changes in representation can impact strongly the evaluation results. It also shows how our framework helps us to structure our analysis of the potential representations. Lastly, we discuss how this tool opens several avenues for future research.
Full Paper in PDF Document