The main aim of this study was to investigate a selection of commonly used performance indicators (PIs) that have been reported to be sensitive to distraction and workload. More specifically, the PIs were tested for their ability to differentiate between task modalities (visual, cognitive and haptic) and task difficulty (easy, medium and hard). It was investigated whether possible differences were constant across two traffic situations (with/without lead vehicle) and two driving simulators. The experiment was conducted in the VTI Driving Simulator III, an advanced moving-base simulator, and in the Volvo Car Corporation driving simulator, an advanced fixed-base simulator. Both simulators were equipped with Smart Eye Pro eye tracking systems. A visual, a cognitive and a haptic secondary task were chosen to test the ability of the PIs to distinguish between the tasks’ loading on different modalities. Some of the main results from the study were:

• There were only minor differences between the two simulators for driving behaviour as described by longitudinal PIs. There was no overall offset, and the main difference was that the visual task led to stronger speed reductions in the moving-base simulator, which influenced both the mean speed and the speeding index.
• Regarding lateral PIs, major differences between the two simulators were found, both as a general offset and for those factor combinations that include modality and task difficulty level.
• With the visual or the haptic task active, the drivers positioned themselves further to the left and the variation in lateral position was higher in the fixed-base simulator.
• The number of lane crossings did not differ considerably between the simulators, but the lane departure area was larger on average in the fixed-base simulator, again influenced by modality, with the largest lane departure areas for the visual task, and in the case of the fixed-base simulator for the haptic task as well.
• Most of the eye movement related PIs had a general offset between the simulators. The drivers in the fixed-base simulator accumulated more time with their eyes off the road, especially during the visual and the cognitive tasks, while the drivers in the moving-base simulator cast longer single glances at the display.