لخّصلي

The mean distance at which the pedestrian was first recog- nised was analysed using a three-way linear mixed effects model with the factors of Age (with two levels: Young and Old), distracters (with three levels: None, Visual and Audi- tory) and Clothing (with two levels: Vest and Biomotion),

Night-time pedestrian recognition
with subjects as a random factor, and condition order as a continuous covariate. To examine which of the visual func- tion measures best predicted the overall mean recognition distance for each participant, Pearson bivariate correlations and Pearson partial correlations (controlling for driver age) were conducted. All data met the assumptions of normality and homogeneity of variance, and multivariate tests of sig- nificance were employed to rule out problems associated with sphericity.
Results
Table 1 presents the mean visual function data for the younger and older participants. Although the mean visual acuity for the older group was slightly better than 6/6, it was significantly worse than the mean acuity of the younger group. The mean letter contrast sensitivity and motion sen- sitivity were also significantly worse for the older group, but differences in performance on the UFOV were signifi- cantly slower only for the selective attention subtest, but not for visual processing speed or divided attention sub- tests.
Figure 1 shows the mean pedestrian recognition distances as a function of age group and pedestrian clothing. The mean distance at which the drivers first recognised that a pedestrian was present was 190.3 m col- lapsed across all drivers, pedestrian clothing and distracter conditions. Overall, the older participants recognised the pedestrians at significantly shorter distances than did the younger participants F1,30 = 33.4, p < 0.001; with the older participants recognising pedestrians at approximately half the distance of the younger participants (129.1 m vs 251.5 m).
The clothing manipulation had a significant effect on pedestrian recognition distance (F1,30.95 = 105.77,
Table 1. Mean (and standard deviation) scores on tests of visual perfor- mance for the younger and older participants
p < 0.001). When averaged across driver age and distracter conditions, drivers responded to the pedestrian wearing biomotion clothing at a mean distance (249.0 m) that was almost twice as long as that of the standard vest alone (131.7 m). Response distances were also affected by the presence or absence of distracters (F2,29.56 = 5.254, p = 0.011) such that the visual distracter condition resulted in significantly shorter recognition distances than the audio distracter condition (189.7 vs 219.39 m) but the audio and baseline conditions (207.71 m) did not differ significantly. There was no significant age 9 distracter (F2,29.62 = 0.16; p = 0.86) interaction. The interaction between clothing and distracters (F2,29.68 = 0.46; p = 0.64), and the three- way interaction between clothes, distracters and age (F2,29.64 = 0.03; p = 0.97) were also non-significant. There was however a significant two-way interaction between clothing and age F1,30.04 = 6.30, p = 0.02 (see Figure 1). Although for both groups, the biomotion clothing configu- ration was visible at a longer distance than the vest, this effect was stronger for the younger participants.
An analysis of the visual predictors of the mean distance at which pedestrians were first recognised (after collapsing across the clothing and distracter variables), revealed that visual acuity, contrast sensitivity and motion sensitivity were all significant bivariate predictors, as was the UFOV test 3 (selective attention; see Table 2). In each case, drivers with better visual sensitivity (or faster UFOV response times) tended to recognise the pedestrians at longer dis- tances. The relationship between motion sensitivity and overall pedestrian recognition distance for both age groups is shown in Figure 2.
To determine whether the relationship between the visual measures and pedestrian recognition distance were driven purely by ageing, partial correlations were also con- ducted with age as a covariate (Table 2). The partial