|M.Sc Student||Yael Auerbach-Shpak|
|Subject||Adaptive user Interface: The Effect of the Adaptation|
Strategy on User Performance
|Department||Department of Industrial Engineering and Management||Supervisor||Full Professor Erev Ido|
Adaptive user interface is an interface that changes its displays in order to create a convenient workplace according to specific trigger and timing. There are several adaptation strategies: Semi-automatic Adaptive Interface: the system detects automatically the trigger for adaptation and recommends changing the interface. The user controls the adaptation. Automatic Adaptive Interface: The user does not have any control over the system that detects the trigger for the adaptation and changes the interface accordingly.
The aim of this research was to answer three main questions: Which interface, Static or Adaptive can lead the user to better performance? Can a Semi-automatic Adaptive Interface improve user performance more than an Automatic Adaptive Interface, or will the decision making required in the semi adaptive interface increase user workload and therefore decrease the performance? What is the effect of timing of adaptation on user performance?
In order to examine these issues, we used a simulated support center with dynamically changing amounts of service requests waiting to be handled. These changes were the trigger for the adaptation. Participants were assigned to six groups: Two control groups, each with a fixed interface designed for either a low or high workload. The adaptive interfaces changed between those two static interfaces according to the workload. There were four experimental groups: two Automatic Adaptation group and two Semi-automatic Adaptation group. Both those groups were crossed with two adaptation timing conditions: Immediate timing- the adaptation occurs immediately after the occurrence of the trigger without a preliminary warning. Delayed timing the adaptation was delayed relative to the occurrence of the trigger, and a graphic warning suggested the up-coming adaptation.
According to the findings, if there is a known static interface that can assist the user through all tasks, it is better to design one interface. Furtherer more, when the user is given control over the adaptation, it is not utilized in an optimal way. Therefore, it may be better not to allow the user control the changes in the interface.