Agents, affectivity and aesthetics in user-interface design

  • Karen Wilson

Student thesis: Doctoral Thesis


Human-Computer Interaction (HCI) research has suggested that there has been a general shift away from more traditional aspects of design, such as usability, and the focus is now more pleasure, or emotion, based. Jordan (2000) states that pleasure based design is associated with more hedonic aspects, such as enjoyment, and the interaction between the user and the product. A number of studies have investigated the relationship between the user and product in relation to agents (e.g. DeAngeli, Lynch & Johnson, 2002). While it has been widely acknowledged that the use of agents enhances the user experience (e.g. Lester, Converse, Stone, Kahler & Barlow, 1997; DeAngeli et al, 2002), these agents have mainly been fully animated, which may influence perceptions as they are considered more ‘human-like’. Conversely, other studies (e.g. Koda & Maes, 1996) have found that the appearance of an agent has little influence on user perceptions of attributes such as intelligence, but only if an interaction has taken place. A series of experiments were conducted in order to investigate user perceptions of agents, or more specifically the influence of aesthetics, context and interaction. Experimental work investigated rating of agents, on a variety of different attributes, in an implied financial context. It was found that there was a general positive regard for female agents, but it was unclear whether stereotypes relating to the context were driving these judgements. Additional work showed that this positive view of female agents was consistent when no context was implied. Further investigation of the role of context indicated that while context (compared to imagined context and no context) had an overall detrimental influence on perceptions of agents, there was a high general regard for both attractive agents and female agents. However, to ensure that this result did not simply arise from the choice of website (financial), additional studies investigated the extent of occupational stereotypes and whether these occupational stereotypes extended to agents. This was found to be the case. However, although the most appropriate agent for a given occupation was one that was gender-congruent with the occupation, there was still a general positive regard for attractive agents. Finally, the influence of interaction was explored, and it was found that the aesthetics, or even presence, of an agent had no effect on user perceptions. This suggests that the quality of interaction may be the most salient aspect of agent perception. These findings are discussed in relation to the literature and future studies are considered.
Date of AwardJun 2007
Original languageEnglish

Cite this