Field Study: Influence of Different Specification Formats on the Use Case Point Method

The Use Case Point method (UCP method) allows early, easy estimation of the anticipated effort during a software development project. The basis for such estimation in real industrial projects is commonly a number of rough specifications in different formats and of differing granularity. The success of the UCP method and comparability of the results depend above all on whether and how good use cases can be identified and weighted from the specifications. Within a field study, a total number of more than 200 UCP estimations based on eight different specification formats have been performed. The estimations have been compared quantitatively and qualitatively with regard to the reproducibility of effort estimation and with regard to expert valuations. With the help of statistical methods a mean variance (variation coefficient) between 13 % and 48 % was found depending on the specification format. Thus, a valuation of specification formats for improving estimation accuracy could be derived with the help of variance analysis.