Assessing the Role of Human Factors in Software Effort Estimation: A Quantitative Study

Accurate effort and cost estimation is vital for software project success and even though human factors play a major role in estimation accuracy, they are often overlooked. This study explores the quantitative impact of some of these human factors by designing and conducting an effort estimation-related experiment involving 20 past real-life projects and 40 software industry professionals. Information about software development experience (total as well as project management-related) was obtained from each practitioner who was also asked to attempt the standard Five Factor Model personality test. These practitioners were then assigned to different four-member teams and each team was asked to estimate the software development effort (in person hours) for four different past projects using Planning Poker. The Weighted Team Homogeneity Index (WTHI) metric was used to quantitatively measure the homogeneity of these teams. Once all the estimation sessions had been completed, these teams were then clustered into different groups for determining the stand-alone impact of experience and team homogeneity on estimation accuracy. Estimation accuracy of all teams was measured using the widely-used mean magnitude of relative error (MMRE) metric. Experiment results reveal that, by and large, more experienced and more homogeneous teams come up with more accurate effort estimates. Thus, software project managers should consider both experience and homogeneity while staffing their teams at project inception.