PPIG 2003 - 15th Annual Workshop
Software Effort Estimation: Unstructured Group Discussion as a Method to Reduce Individual Biases
Kjetil Moløkkeni, Magne Jørgensenii
Abstract: The effort of software projects is often estimated, completely or partially, using expert judgement. This estimation process is subject to biases of the expert responsible. Generally, this bias seems to be towards too optimistic estimates regarding the effort needed to complete the project. The degree of bias varies depending on the expert involved, and seems to be connected to both conscious and unconscious decisions. One possible way to reduce this bias towards over-optimism is to combine the judgments of several experts. This paper describes an experiment where experts with different backgrounds combined their estimates through group discussion. Twenty software professionals were asked to provide individual effort estimates of a software development project. Subsequently, they formed five estimation groups, each consisting of four experts. Each of these groups agreed on a project effort estimate through discussion and combination of knowledge. We found that the groups were less optimistic in their estimates than the individual experts. Interestingly, the group discussion-based estimates were closer to the effort used by the actual project than the average individual expert, i.e., the group discussions led to better estimates than a mechanical combination of the individual estimates. The groups’ ability to identify more project activities is among the possible explanations for this reduction of bias.