Although Quality of Experience (QoE) plays a major role in the design and development of mobile applications and services, QoE assessment is still challenging, especially in real-life (so called `living lab') contexts. This paper presents results from an exploratory, interdisciplinary study on QoE and more specifically, on the acceptability of varying audiovisual qualities during video watching on a mobile device, in a realistic setting. The subjectively measured acceptability is modeled by means of a decision tree which uses information about the network type, transport protocol, technical video characteristics, and user behavior. The proposed model may contribute to the estimation of users' subjective evaluation of the quality during mobile video watching and to QoE optimization by dynamically altering the parameters that have the largest influence on this subjective evaluation.