Mental Models Informing Interface Design: Evaluation of a Method Applied in the Redesign of a Multiserver Managing Tool
Master's thesis of Amelie Piepenbring
The investigation of mental models of complex technological systems goes back to the beginning of the development of computers. Mental models are representations of objects or processes in the consciousness of a living being. They consist of the relevant components of a system and an idea of how they interact. How do users understand such complex technological systems and what can be derived from the mental models of the users for the design of such systems?
At the heart of this work is a software that allows the management of multiple virtual machines based on hardware virtualization technology. This software is to be redesigned. For this purpose the mental models of the users shall be considered. Possible changes can be: change of processes and menu representation within the system, integration of new functions and renaming of the components. A variety of methods are available for the collection of mental models, such as Card Sorting and structuring techniques. However, these methods are mostly designed to answer specific questions and are not suitable for use in a requirement analysis.
This work is methodologically based on previous work, which aimed at studying the mental models of users holistically. The method consists of a detailed interview with the subject about their understanding of the multiserver managing tool and a subsequent structuring technique in which the relevant components for the subjects are to be put into relation to each other. Thereby we want to address concrete issues such as system structure and wording while at the same time keeping the method open enough to learn about additional problems and needs that subjects may have. It is to be investigated whether the proposed method can fulfill these requirements. The analysis is carried out first by means of a visual diagnosis, followed by a network analysis.
The method makes it possible to assess the relevance of system components and the appropriate wording of those components. A grouping of the system functions and components can also be derived. In this respect, the method is a more open alternative to Card Sorting. It also allows to capture other functions that users of the system want to be integrated into the system. However, the method is not designed for the specific design of the processes of a system, but can only inform developers and designers when designing prototypes for more detailed testing. Advantages and disadvantages of the method, as well as possibilities for standardization and simplification of the evaluation are discussed.