Hamilton College computer science professor Stuart Hirshfield has received a $458,900 grant from the U.S. Air Force Office of Scientific Research to establish the Hamilton College Next-Generation Usability Laboratory. The grant has funded unique equipment that is being used to model and quantify levels of trust that exists in typical interactions between human users and computer systems and to identify and evaluate patterns and levels of trust, suspicion and frustration with various computer interfaces.
Stuart Hirshfield, the Stephen Harper Kirner Chair of Computer Science, and Research Associate Leanne Hirshfield have begun studying the real-time, quantitative assessment of computer users' mental states to enhance usability testing. They are able to make concurrent cognitive, physiological and behavioral user measurements using a fNIRS, a cutting-edge piece of equipment which measures blood movement in the brain, and other non-invasive equipment. They have begun evaluating and developing adaptive interfaces that react to user workload, moods, and emotions.
Both Hirshfields have expertise critical to this research: in usability testing, user interface design, computer programming, machine learning and cognitive psychology. Leanne Hirshfield completed her doctorate in computer science at Tufts University in the area of Human-Computer Interaction (HCI), and she presented the first published paper on using fNIRS for usability testing.
In recent years, Stuart Hirshfield has focused his teaching and research interests on his roots in HCI. Early in his career, he worked as a research scientist for Xerox when the corporation was developing some of the first computer interfaces. He has been involved in numerous research efforts that deal directly with HCI development and evaluation at the Air Force Research Laboratory in Rome, NY.
Among the non-invasive equipment the Hirshfields are using to measure user states are:
In addition, the team employs usability software that measures more traditional remote metrics such as speed, accuracy, key strokes, mouse movement and screen captures.
Their work in modeling and quantifying the level of trust that exists in typical interactions between human users and computer systems can be applied in many ways, among them:
No entries found