KI-04-2002 mit web-Hinweisen.p65

ed to car tographic point-like symbols. Nonetheless, the qualitative locative information of the landmarks is maintained, for example their position at a corner of an intersection. The benefits of maps are apparent from their ubiquitous existence in our culture. Recently, route maps have become widely available through the Web and on-board navigation systems (e.g. Agrawala & Stolte, 2001). Despite the ubiquitous status of route maps, optimal visual representation methods are still a matter of research. For example, route maps integrated in on-board navigation systems present routes dynamically with a moving dot that traverses through the map to simulate an imagined navigator. In contrast, internet maps present information statically with lines representing the route. Differences in presentation modes are caused by technical constraints of each medium, rather than by consider ing cognitive efficacy. When determining which of these two presentation modes – dynamic or static – would provide better instruction for users, dynamic presentation seems to be more attractive and effective at first glance. Animated graphics have shown to be more effective than their static counterparts in some studies (e.g. Kieras, 1992; Nathan et al., 1992). However, Tversky and colleagues (2001) argue that advantages in animated conditions of these studies are due to other factors, such as interactivity or inclusion of information that are not present in static conditions. Furthermore, other studies fail to demonstrate advantages for animations altogether (e.g. Morrison, 2000; Hegarty et al., 1999). Why do animations fail to show any benefits over static diagrams? One explanation is that although motion in animations is perceived continuously, people comprehend and remember it in discrete steps (Zacks et al., 2001). The discretizations of events occur systematically and the break points of the events (i.e. points where events are segmented) are better remembered than other points within an event. Similarly, when people recall route information, they decompose the route into a set of discrete path segments, consisting of only minimal information such as turns and landmarks at the turns, in congruence with effective wayfinding aids ( Jackson, 1998). Turning points are better remembered than non-turning points, as most of the landmarks and streets along the route are ignored and omitted (Tversky & Lee, 1998; Denis, 1997), or parts of the route are chunked to more complex route segments (Klippel et al., 2002). Acquisition of Landmark Knowledge from Static and Dynamic Presentation of Route Maps This contribution reports on ongoing collaborative research a t the Universit y of Stanford, D epartment of Psychology, and the Universit y of Hamburg, Depar tment for Informatics. Extending the research on the effects of static vs. dynamic route presen tation on per ception and memory (Klipp el et al., 2002), we e xamined differen t route presentation methods that are commonly used t oday (e.g. internet maps, GPS maps, etc.) and their effects on the subsequen t route memory. Participants learned a route from a map of a fictitious town. The route was presented to them either as a solid line (i.e. static), a moving dot (dynamic), or a dot super imposed on a line (mixed). In a subsequent recall task, participants in all three conditions remember ed equivalent number of landmarks, but participants in the dynamic condition rec alled less per tinent landmarks than those in the static condition, who in turn r ecalled less than participants in the mixed condition. The results highlight the importance of landmarks at decision p oints and hint at differences in mental processing of dynamically and static ally presen ted spatial information. Figure 1: Graphical representation of the direction “Shell gas station is nor th of McDonald’s” N Paul U. Lee, Heike Tappe, Alexander Klippel Auszug aus: Künstliche Intelligenz, Heft 4/02, www .kuenstliche-intelligenz.de ISSN 0933-1875, arendtap Ver lag, Bremen fon +49 421 34889-30 fax: +49 421 34889-31 33 KI P r o j e k te 4/02 Based on these observations, we predict that dynamically presented route information would hinder effective segmentation since a continuous presentation would bias the user to attend to features along a path equal to those at turns. In the following experiment, we present route information to participants dynamically or statically and examine how the presentation mode affects their memory of landmarks. Landmarks are important features in organizing the memory for spatial information and they are vital to route directions (e.g. Denis, 1997; K lein, 1982). Landmarks are distinguished features of the environment clearly standing out of a multitude of spatial information perceivable by the senses. They fulfill var ious functions in organizing route directions, for instance demarcating decision points (e.g. “ turn left at the restaurant”), confirming correct progression along the route (e.g. “continue past the school”), or providing a global bearing (e.g. “the sea is on your left and the mountains are on your right”). We predict that a dynamic map with a moving dot would create equal memory trace for landmarks at the turns and landmarks along the route. In contrast, we predict that par ticipants would remember the landmarks at the turns better when they learn from a static map. By studying how var ious presentation methods affect the underlying route memory, this study can suggest how applications should be designed to best utilize cognitive structures. Since landmarks at turns are critical components of route information, we will focus on participants’ recall of these landmarks. 2 Dynamic vs. Static Presentation of Maps