RT Book
T1 Automated and Participative Decision Support in Computer-aided Credibility Assessment
A1 Jenkins, Jeffery
A2 Lowry, Paul Benjamin
A2 Jensen, Matthew L.
LA English
YR 2009
UL https://krimdok.uni-tuebingen.de/Record/1866324225
AB History has shown that inaccurate assessments of credibility can result in tremendous costs to businesses and society. This study uses Signal Detection Theory (SDT) to improve the accuracy of credibility assessments through combining automated and participatory decision support. Participatory decision support is also proposed to encourage acceptance of the decision aid’s recommendation. A new hybrid decision aid is designed to perform automated linguistic analysis and elicit and analyze perceptual cues (i.e., indirect cues) from an observer. The results suggest that decision aids that collect both linguistic and indirect cues perform better than decision aids that collect only one type of cue. Users of systems that collect linguistic cues experience improved credibility assessment accuracy; yet, users of systems that collect both types of cues or only indirect cues do not experience higher accuracy. However, collecting indirect cues increases the user’s acceptance of decision-aid recommendations