An experiment was conducted to assess the distraction potential of secondary tasks using the Alliance of Automotive Manufacturers’ Principle 2.1B verification procedure implemented using specifications obtained from an Alliance member company. Secondary tasks included radio tuning, destination entry, 10-digit dialling, dialling via contact selection, and text messaging were performed using the original equipment systems of a 2010 Toyota Prius V and an iPhone 3GS smart phone. Sixty-three participants 35 to 54 years old drove a low-fidelity, PC-based simulator while performing the secondary tasks. The driving scenario required participants to maintain a 150-ft following distance behind a lead vehicle that was travelling at a constant speed of 50 mph. Alliance driving performance metrics included lane exceedance frequency and the standard deviation (SD) of car-following headway. Text messaging was associated with the highest levels of driving performance degradation on both Alliance Principle 2.1B metrics, followed by destination entry. Radio tuning had the lowest levels of driving performance degradation. The two phone dialling tasks (contact selection and 10-digit number dialling) had approximately equivalent amounts of driving performance degradation. Both Alliance Principle 2.1B metrics were correlated with task duration. Use of a duration-adjusted metric revealed that text messaging was significantly more demanding than other secondary tasks and that other differences between tasks were due primarily to differences in task duration. Results using smaller sample sizes were consistent with the expected loss of statistical power inherent in the use of small samples. Differences were also apparent between replications of samples of the same sizes. Based on the results of the current study, neither 20 nor 30 participants is sufficient to obtain consistent test outcome results. Two sets of analyses compared present study outcomes with those from a previous study, which used a Dynamic Following and Detection (DFD) protocol that provided both Alliance Principle 2.1B and other vehicle control and visual target detection metrics. Results from the two Principle 2.1B verification procedure implementations were not consistent, despite the fact that both implementations were consistent with Alliance specifications. Differences in driving (car following) task demands appear to have contributed to this finding. The second comparison with previous study results used different metrics and their respective decision criteria. Decisions concerning the acceptability of specific number/text entry tasks based on Alliance Principle 2.1B metrics were not consistent with those made using DFD metrics. The two protocols provided consistent results on 7 of 10 planned comparisons. Different conclusions derive from three factors: (1) driving behaviours represented by the metrics, including target detection in the DFD protocol, (2) metric construction, reflecting the differential treatment of task duration, and (3) decision criteria. Alliance (duration-influenced) and DFD (duration-adjusted) metrics provide complementary information concerning distraction effects, a combination of which provides a better estimate of the total exposure to crash risk associated with secondary task performance than either metric type alone. (Author/publisher)
Samenvatting