|M.Sc Student||Iris Ben-Yair|
|Subject||Using Embodied Conversational Agents as Medication|
Instructors for the Elderly:
Does Anthropomorphism Matter?
|Department||Department of Architecture and Town Planning||Supervisor||Professor Nabarro Ron|
|Full Thesis text - in Hebrew|
The present study examines the correlation between the level of visual anthropomorphism of the ECA and the preferences of the elderly user. The dimension of visual anthropomorphism chosen is “humanness” (man, animal or object). The study also examines the correlation between the level of ECA's humanness and the perception of its persona, as well as the willingness to use it as a medicinal guide. All research assumptions were tested on both the elderly (experimental group) as well as the young (control group) subjects.
9 video-clips were produced (3 animated figures * 3 medicinal explanations). The figures were selected during a preliminary experiment. 36 subjects participated in the experiment: 18 elderly and 18 young. Each group split identically in terms of gender. Each subject viewed 3 movies, each featuring one of the three different figures. After viewing the clips, the subject answered an API questionnaire (Baylor and Ryu, 2005) checking satisfaction level and perception of the ECAs' persona, preference questions, and questions regarding demographic details, and habits of computers use and acquiring drug information.
The results of the experiment concurred with the research assumptions. The main research proposition was validated as the elderly significantly preferred the "man" ECA over the "animal" and the "object". This is in contrast to the young subjects, who did not show significant preference for any of the ECAs, thus correlating with earlier studies. When taking into account first, second and third choice, it was found that young subjects also preferred the "man" ECA, although not statistically significant.
The second research proposition was also validated, as the elderly gave higher scores than did the young subjects for all four persona components for each of the three ECAs. Both age groups assigned to the "man" ECA the highest scores for all four persona components, in comparison to the "animal" and "object".
The conclusive research proposition, stating that the elderly will want to use the ECA as a drug information source, was also verified. 72% of both the elderly and the young groups showed high willingness to use the video-clips. Both the elderly and the young subjects preferred to continue to receive instructions from the "man" ECA. In both age groups the "man" ECA elicited the highest level of willingness to accept the ECA as the only source of medicinal information, without the need for any additional explanation from another source.