Empirical validation of metrics for UML statechart diagrams

It is widely recognised that the quality of Object Oriented Software Systems (OOSS) must be assessed from the early stages of their development. OO Conceptual models are key artifacts produced at these early phases, which cover not only static aspects but also dynamic aspects. Therefore, focusing on quality aspects of conceptual models could contribute to produce better quality OOSS. While quality aspects of structural diagrams, such as class diagrams, have being widely researched, the quality of behavioural diagrams such as statechart diagrams have been neglected. This fact leaded us to define a set of metrics for measuring their structural complexity. In order to gather empirical evidence that the structural complexity of statechart diagrams are related with their understandability we carried out a controlled experiment in a previous work. The aim of this paper is to present a replication of that experiment. The findings obtained in the replication corroborate the results of the first experiment in the sense that at some extent, the number of transitions, the number of states and the number of activities influence statechart diagrams understandability.