Evaluating a Configurator Application for Modeling Data Collection Instruments: An Experimental Study

Software applications are a crucial part of our daily life and become more and more indispensable. Over the years, software itself has gone through a rapid development. Therefore, not only the complexity has increased, but also the quality requirements of customers. One important quality aspect that have to be fulfilled in order to guarantee acceptance of the software application is its usability. However, ensuring the latter is by no means easy. Thus, developers are obliged to test their software product during the development phase in order to gather respective information. Furthermore, testing allows reacting immediately and adapting the software accordingly if needed. Experiments are common methods for testing and, therefore, for obtaining important data as scientific foundation for further analysis.
The thesis at hand evaluates a newly developed configurator application for modeling data collection instruments regarding its usability. Therefore, a controlled experiment is set up in order to investigate how much effort end-users need to properly handle the application. In particular, the overall understanding of the (modeling) concept with respect to the complexity of the configurator is evaluated. In this context, the software is used in order to model specific data collection instruments. More precisely, participants in the experiment are asked to process tasks of different complexity using the configurator application. Thereby, the focus is on the total time needed to solve the tasks as well as the number of errors in the resulting questionnaire models. Considering the results obtained assumptions regarding the intuitiveness of the (modeling) concept can be made.