Assessing individual performance based on the efficiency of projects

Abstract

Assessing the performance of each of the team members in any project is a difficult task since in most cases the outcomes mask the individual contributions. In this paper a two-step approach is proposed to estimate the relative performance of the members of the organization who have participated in the different projects carried out. Firstly, a DEA model evaluates the efficiency of the projects, whose results are used later on in a regression model to assess the performance of each individual. The proposed approach is validated using 480 instances identifying the factors that may affect the accuracy of the results. The application to a case study involving a real-world dataset of 46 software development projects carried out by a ten-member team is also presented. The results show that the proposed approach estimates, on average, the unknown true efficiency of the individuals rather well and is robust against data noise.