Usability in digitalen Kooperationsnetzwerken. Nutzertests und Logfile-Analyse als kombinierte Methode

Abstract

Usability is a key factor when developing new applications. The interaction between the users and the application should be efficient, effective and engaging. Furthermore, a good usability includes a high error tolerance and an good learnability. Different methods allow the measurement of usability throughout the development (process). All methods have in common that the different employed steps like planning, conducting and evaluating are rather time-consuming. When end-users are included as subjects, usability tests are employed. Due to the high time-effort, usually ten or less tests are conducted. The thesis tries to solve this point by trying to combine usability tests and logfile analysis. The empirical work is two-folded. First, usability tests within a learning management system (LMS) are logged in the background. These logfiles are assigned to severe usability problems. Second, the paths of the severe usability problems are combined with logfile data from a "real-world" LMS that runs the same application. The "real-world" logfiles contain a period of about 300 days with 133 active users. Prior to the combination, both data sets converted into a similar format. Being a new procedure, the definite similarity value had to be specified by descriptive statistics and visual inspections. The final combination makes it possible to determine the severity of usability problems on the basis of "real-world" usage data. The proposed method offers a more precise overview of the occurrence of the found usability problems, independent of the test situation. This thesis provides additional value to the fields of (Web) Data Mining, Usability and Human-Computer Interaction (HCI). It also offers additional knowledge to the field of software development, quantitative and quantitative research as well as computer-supported cooperative work (CSCW) and learning management systems (LMS).