Comparison Between Occlusion Methods for Measuring Distraction Caused by Smartphone Interaction

The occlusion method is an established method for measuring the visual demand from in-vehicle interfaces. The method is usually employed by means of occlusion glasses which can be automatically made opaque or transparent. However, occlusion can also be achieved by turning on and off the screen of the interface being studied. It can be hypothesized though, that glasses-occlusion requires more visual re-orientation efforts after each occluded interval which may give longer Total Shutter Open Times (TSOTs) than the method where only the interface screen is occluded. Two experiments were conducted, with 10 participants (5 female) in Experiment 1 and 14 participants (7 female) in Experiment 2. In both experiments, participants were seated in a truck mockup and performed tasks on a smartphone. Two occlusion conditions were employed: 1) Occlusion with glasses, and 2) Embedded occlusion (turning the screen of the smartphone on/off). Participants also performed a baseline condition with no occlusion. In Experiment 1, five different tasks were performed in each condition: destination selection, dialing a phone number, dialing a contact, changing radio frequency and setting the alarm clock. For each task, a note showing the data to be entered was posted next to the smartphone, e.g. the number to be dialed. The four tasks used in Experiment 2 were: visual-manual text entry, make changes in a truck driving log, dial a phone number and select item from a map (using pinch/swipe operations). The participants did not have support from any note but had to remember the data to enter for each task. For both experiments, TSOTs and Resumability (R) were submitted to separate ANOVAs to reveal differences between occlusion and tasks. The results of Experiment 1 showed a main effect of occlusion type on TSOT (p < .05), where Glasses resulted in longer TSOTs (M = 8.1s) than Embedded (M = 7.4s). However, the interaction between occlusion and task was also significant (p < .01), and suggested that mainly the phone dialing task caused the difference in TSOTs. An explanation to this could be that tasks requiring more visual-spatial reorientation are more difficult to perform with glasses occlusion. A simpler explanation could be that participants needed to look at the note while dialing – which can be done during the occluded intervals with the embedded method. The results from experiment 2 did not show any statistically significant effect of occlusion on TSOTs which suggests that the effects found in Experiment 1 were caused by the fact that participants had to look at the paper note while dialing. The current experiments indicate that the embedded method gives TSOTs comparable to the glasses-occlusion method. However, one must be aware of the fact that there are practical differences between the methods which can lead to diverging TSOTs. The current experiments show that such divergence can occur when instructions are given visually. The embedded occlusion method for evaluating distraction is nonetheless promising since it is easy to use and can be integrated in app developers’ toolkit.