Distributed modeling of use case diagrams with a method based on think-pair-square: Results from two controlled experiments

Objective: In this paper, we present the results of two controlled experiments conducted to assess a new method based on think-pair-square in the distributed modeling of use case diagrams. Methods: This new method has been implemented within an integrated environment, which allows distributed synchronous modeling and communication among team members. To study the effect of the participants' familiarity with the method and the integrated environment, the second experiment is a replication conducted with the same participants as the original experiment. The results show a significant difference in favor of face-to-face (i.e., the chosen baseline) for the time to complete modeling tasks, with no significant impact on the quality of the produced models. Results: The results on participants' familiarity indicate a significant effect on the task completion time (i.e., more familiar participants spent less time), with no significant impact on quality. Practice: One of the most interesting practical implications of our study is - in case the time difference is not an issue, but moving people might be a problem, the new method and environment could represent a viable alternative to face-to-face. Another significant result is that also people not perfectly trained on our method and environment may benefit from their use: the training phase could be shortened or skipped. In addition, face-to-face is less prone to consolidate participants' working style and to develop a shared working habit of participants. Implications: This work is in the direction of the media-effect theories applied to require- ments engineering. The results indicate that the participants in the experiments significantly spent less time when modeling use case diagrams using face-to-face. Conversely, no significant difference was observed on the quality of the artifacts produced by the participants in the these tasks.