Using an Intelligent Interviewer to Conduct Cognitive Assessments

Cognitive assessments are conducted in one of two ways; a trained professional will conduct a face-to-face interview, collecting the answers from the interviewee during the interaction; or a computer will present an inventory as a list of questions and multiple-choice answers. While a trained professional can provide an individualized assessment there is also concern that the interviewee can appear different while interacting with different assessors. Although conducting the assessment using a computer provides a consistent and cost-effective way of conducting an inventory this static process affects interpretation because it does not provide a means of clarification either for the user or the computer: the user to clarify ambiguous wording, or the computer to clarify potential conflicting responses from the user. This dissertation researches an Intelligent Interviewer conducting a cognitive assessment, specifically a learning style inventory. The Intelligent Interviewer was presented to the participants in two forms: an Embodied Interviewer, where the user and computer conversed vocally and visually; and a Text Interviewer, where the user and computer conversed by typing and reading text. Both the subjective and objective evaluation metrics were studied to determine the user experience, effectiveness, and efficiency of the Intelligent Interviewer