Measuring effects of user-specific behaviour on selection tasks in HCI

Abstract

Assessing the user behaviour is essential for adaptive multimodal dialogues in human computer interaction (HCI). However, in recent research, the vast range of possibly interacting parameters make it difficult to examine the influence of individual factors such as previous experience on user behaviour. In this paper we apply a new experimental setup, which enables independent examination of a large range of factors, in order to examine the effects of previous experience on the user's choice of modality and performance. The results show that previous experience in terms of an interaction history has a systematic effect and is thus a key factor for the development of adaptive dialogue systems.