An obvious metric to measure the monitoring performance between the different conditions would be to compare how many clicks the users made in average for each condition. Furthermore of interest are the buffer values of the respective buffers at the time of the user's interaction with the simulation (e.g., the input buffer of a certain machine at the time of refilling it). A relatively high average buffer value can e.g. signify that the users do not trust that the respective mode of process monitoring conveys the need for interaction in time, leading the users to switching their attention to the process simulation in regular intervals, and performing interactions just in case. A low average buffer can, on the other hand, signify that the users rely on the respective conditions’ ability to signal interaction needs. On the other hand, if e.g. an input buffer had already been completely depleted at the time of intervention, this may signify that the respective condition has failed to inform the users in time. In many cases, participants used double clicks for their interactions, while a single click would have been sufficient, a fact that was perhaps not communicated clearly enough to the participants. Therefore, if several clicks were performed directly one after another, only the first click was taken into account.
