|
Article Excerpt
INTRODUCTION
Older drivers are overrepresented in fatal traffic accidents on a per-mile basis (Evans, 1988; Hakamies-Blomqvist, 1993; Massie, Campbell, & Williams, 1995; McGwin & Brown, 1999; Preusser, Williams, Ferguson, Ulmer, & Weinstein, 1998; Stamatiadis & Deacon, 1998), most likely because of their fragility (Evans, 1988; Hauer, 1988). After age 75, the risk of intersection accident involvement for older drivers increases dramatically for most intersection maneuvers (Preusser et al., 1998; Staplin & Lyles, 1991). About one half of all driver fatalities for those 80 years of age and older are at intersections, compared with 23% for drivers younger than 50 years (Insurance Institute for Highway Safety, 2000). Typical citations by older drivers, once they are involved in an intersection accident, are failure to yield right of way and violation of traffic controls (Caird & Hancock, 2002). Failures of perception (Caird & Hancock, 2002; Schiff, Oldak, & Shah, 1992; Staplin, 1995), attention (Owsley, 2004), memory (Delorme & Martin-Lamellet, 1998; Guerrier, Manivannan, & Nair, 1999), cognition (Drakopoulos & Lyles, 1997), and action (Caird, Horrey, & Edwards, 2001; Hakamies-Blomqvist, 1994; Lerner, 1994) are frequently used to explain why older drivers are involved in accidents. Research that seeks to understand and predict why intersection accidents occur anticipates the unprecedented demographic shift that will swell the ranks of older drivers in the future (Caird & Hancock, 2002; Hakamies-Blomqvist & Henriksson, 2000; Owsley, 2004).
The current research examines the contribution of attentional failures at intersections. Attentional failures may result from the improper division of attention (Caird & Chugh, 1997; Ponds, Brouwer, & van Wolffelaar, 1988), visual search difficulties (McDowd & Shaw, 2000; Scialfa, Kline, & Lyman, 1987; Scialfa, Thomas, & Joffe, 1994), and/or inappropriate selective attention (Ball & Owsley, 1991; Ball, Owsley, Sloane, Roenker, & Bruni, 1993; Owsley et al., 1998; Owsley, Ball, Sloane, Roenker, & Bruni, 1991 ; Parasuraman & Nestor, 1991). As such, failures of attention may result in drivers failing to detect a potential conflict with another object or detecting the conflict too late to respond appropriately (Caird & Hancock, 2002; Cairney & Catchpole, 1996; Rumar, 1990; Treat, 1980). Knowledge of the nature of visual attention may contribute to the understanding of attentional failures and, consequently, the understanding of driver errors at intersections.
In the current context, the inability of drivers to effectively detect changes in a rapidly changing and dynamic environment, such as a busy intersection, may represent an important attentional failure. Recent studies on change blindness have shed light on the understanding of visual attention. Change blindness is defined as the inability to detect changes made to an object or a scene during a saccade, flicker, blink, or movie cut (O'Regan, Rensink, & Clark, 1999).
In general, change blindness has implications for understanding how humans construct, link, and store visual representations. The long-held view is that people store detailed and coherent picture-like representations of the world from one view to the next. However, recent research into change blindness suggests this may not be the case (e.g., Mack & Rock, 1998; O'Regan et al., 1999; Rensink, O'Regan, & Clark, 1997, 2000; Simons & Levin, 1997). For example, Rensink (2000, 2002) suggested that focused visual attention provides spatiotemporal coherence for the stable representation of a single object or spatial location at a time. As such, accurate visual representations may exist only so long as attention is focused on the region or object in question. When attention is focused in one location, changes occurring in other parts of the visual scene may go unnoticed by observers, simply because there is no detailed representation of the changing location at that particular moment, If focused attention on hazardous objects is required to construct a coherent representation of a traffic scene, it follows that intersections that have increased complexity, traffic flow, and visual clutter will also have a higher incidence of missed changes (e.g., the appearance of a pedestrian from behind an initially occluding object) because drivers will fail to maintain a complete and accurate representation of each aspect of a visual scene.
Change blindness is especially pronounced when brief blank fields are placed between alternating displays of an original and modified scene, which is called the flicker technique (O'Regan et al., 1999; Rensink et al., 2000). In the standard or generic application of this technique, an image (A) and a modified image (A') are presented for a short duration (typically 250 ms) separated by a blank field of 80 ms (i.e., the interstimulus interval, or ISI). The images are alternated repeatedly until the observer detects the changing element or a certain time has elapsed. The blank screen separating the two images simulates a saccade and is used to mask the appearance of new objects in the scene--changes that are readily detected when no such mask is present (Rensink et al., 1997). Importantly, these masks are effective even when they only partially occlude the scene (e.g., "mudsplats"; O'Regan et al., 1999).
Research on change detection has further shown that older adults miss more scene changes than do younger adults, suggesting age-related deficits in the ability to maintain a stable visual representation (Pringle, Irwin, Kramer, & Atchley, 2001). Furthermore, others have shown that changes made to objects of central interest are detected more readily than changes to objects of marginal interest (Pringle, 2000). Richard et al. (2002) extended these results to the driving domain, showing faster and more accurate detection of driving-related changes than of unrelated scene changes. These lines of research, however, adopt the typical flicker technique in which observers are instructed to look for changes. Although this approach effectively demonstrates the effects of change blindness, the task itself (i.e., detecting changes to scenes) is not representative of real-world tasks. Furthermore, it is not clear from past research whether age-related differences might be reduced when the task draws upon previous experience. Specifically, it is not known whether the poor detection performance of older adults might be reduced if they can draw on their driving experience (i.e., the practiced and appropriate allocation of visual attention). The current instantiation of the flicker technique does not afford such use of experience to guide task completion. One goal of the current research is to modify the flicker technique such that the observers' tasks are more representative of driving and that the (implicit) detection of changing features will have an impact on this task performance.
Present Study
In the current study, we modified the flicker method so that it could be used to test drivers' attentional capabilities at intersections. Typically, observers are asked to look for changes in two alternating images. Furthermore, observers are rarely under any time pressure in which to make a decision about whether a change is present. Drivers in a dynamic traffic environment rarely have more than a few seconds to observe a given scene or context. In contrast to the traditional approach, the modified flicker method (MFM) provides observers with a specific goal, rather than simply to monitor for changes, and also imposes some time constraints on observers such that their goal-oriented decision must take place rapidly. In the current study, drivers were asked to decide whether it was safe to complete a certain maneuver (i.e., making either a left or right turn, or continuing straight ahead) at each intersection. Although the MFM does not require participants to search actively for a changing feature, it is assumed that the correct detection of a safety-critical object will impact their decision of whether or not to proceed through the intersection, Imposing a directional goal (i.e., of travel) guides drivers' attention more efficiently and allows them to use prior experience to search for relevant information (Theeuwes, 1996; Yantis, 1998).
In the present study, the MFM was used to determine the effects of time constraints on the performance of younger and older drivers' decision making at intersections. Drivers examined a series of intersections for either 5 or 8 s in order to assess the safety of the intended path of travel. It was expected that a shorter viewing time would negatively impact decision accuracy. To the extent that older adults could draw on experience, there would be smaller age-related decrements in performance. However, if older adults were unable to draw upon related experience, these decrements would remain, suggesting that the impact of less stable visual representations sufficiently offset any benefit of experience.
METHOD
Participants
Sixty-two older and younger drivers were recruited from the following age groups: young (18-25 years, M = 22), middle-aged (26-64 years, M = 39), young-old (65-73 years, M= 69), and old-old (74+ years, M = 78). There were 8 men and 8 women in the first three age groups and 8 men and 6...
|
