سیستم های اطلاعاتی مدیریت اضطراری: آیا تصمیم گیرندگان در انتخاب فرمت های صفحه نمایش پشتیبانی می شود؟

Recent information technologies make it possible to include sophisticated three-dimensional display formats in emergency management information systems (EMIS), decision-support systems that facilitate decision making in crisis situations. However, if decision makers are to improve their decisional performance, they must correctly identify appropriate situations for using these formats. We conduct two experiments and find that, as prior research has suggested, decision makers do not choose the most appropriate display format, but their performance improves when given prospective decisional guidance. We discuss implications of these findings for EMIS design, for the training of emergency management professionals, and for future research on display formats and decisional guidance.

To provide adequate decisional support to manage crisis situations, researchers and practitioners in information system and disaster management have urged attention to the development or enhancement of emergency management information systems (EMIS) [5], [19] and [48]. As a decision-support system (DSS), an EMIS helps members of an emergency management team4 — who are usually performing under conditions of high information load and high time stress — gather and synthesize incomplete but often-critical data [5] and [40]. Prior studies have recommended not only training strategies that may improve EMIS users' decisional performance, but have also suggested guidelines for EMIS design and implementation, such as communications, data integration, and tools for interface and visualization [11], [12], [23], [33] and [40]. Most existing EMIS were designed to render geospatial data as two-dimensional (2D) displays [19] that show orientations and relative positions by laying objects in a plane and by using colors or contour lines to represent elevations (e.g., Google Maps; Fig. 1, Panels 1 and 2). As the availability, performance, and cost-effectiveness of information technology constantly increase, attention has increasingly shifted to three-dimensional (3D) visualization, especially after September 11, 2001 [9]. With an extra dimension, 3D displays integrate the presentations of orientations, relative positions, elevations, and shapes of objects in a single view (e.g., Google Earth; Fig. 1, Panel 5). The literature has suggested that such differences in display format and content richness mandate different levels of user skill or experience, cause different levels of mental workload, and thus lead to differences in decision-making performance [35] and [44]. Although EMIS with either 2D or 3D displays could be utilized in crisis situations, trivial differences in decision performance might cause delays, damage, or casualties. For example, if a 911 call center receives a report of a fire in a downtown skyscraper, incident commanders using an EMIS with 3D views can make decisions promptly, whereas those using an EMIS with 2D displays may require additional assistance to find the exact location. Nevertheless, the technological sophistication and content richness that characterize 3D displays do not guarantee better decision making for all tasks; on the contrary, prior studies in human factors and ergonomics suggest that 2D displays are more useful than 3D when the tasks require judgments on orientation or relative positions, but not on elevations, sizes, or shapes [4], [35] and [36]. In the domain of MIS research, cognitive fit theory (CFT) proposes that the correspondence between task requirements and the information presentation format determines problem-solving performance [41] and [42], suggesting that the dimensionality itself is not the sole determinant of decision performance. Both streams of research suggest a hybrid EMIS that allows users to switch between visual displays flexibly. Two immediate questions arise: Can EMIS users accurately select display formats with the appropriate dimensionality (i.e., 2D vs. 3D) when given a choice? If not, can they be trained or supported by decisional guidance (DG) to choose appropriate display formats? We conducted two related experiments to address these questions. First, using previously validated experimental instruments [4], we examined whether novice EMIS users can choose appropriate display formats without guidance. Specifically, research participants were required to select among several 2D and 3D displays to deal with several tasks that each required different information. Based on the results of this experiment, we found that although participants matched display formats to tasks at a rate reliably better than what would be expected by chance, there was still considerable room for improvement in their display selection performance. Thus, in Experiment 2 we explored whether participants made better display format selections after they were exposed to DG about the functionality of the different formats. The combined results from our experiments suggested that novice EMIS users may not automatically choose the best display format for a task, but DG could significantly improve their performance. Indeed, both DG and training are designed to equip DSS users/trainees with the knowledge and capability to make better decisions. From various formats of DG [29] we chose to examine the impacts of a prospective DG, where decisional supports are provided prior to decision making. Such decisional guidance, as implemented in our study, might also be seen as a form of training. Although examining only a single type of DG, we believe our research serves as a motivator for future research on other formats of DG and yields implications for EMIS design, for the training of emergency management professionals, and for more basic research on information design. The rest of this paper is structured as follows. In Section 2, we provide an overview of theoretical foundations and principles that guided our research hypotheses. Section 3 proposes the framework of the research and experiments. We present the experimental methods and results of our two experiments in 4 and 5, respectively. Section 6 concludes the paper with a discussion of our findings and implications for practice and for future research.

Crisis situations are critical events that require maximally effective and efficient decision making [23] and [33]. Among the many tools available, an EMIS rendering geospatial information by 3D visualization techniques is an important innovation that has been in the spotlight since the terrorist attacks of September 11, 2001 [10]. Before then, most EMIS presented the geospatial information of the target objects by 2D maps, which means that information about height and shape could be communicated only by symbols, colors, verbal descriptions, or contour maps. For EMIS users or developers, a fixed 2D display strategy forces users to make elevation- or shape-related decisions by translating symbols into internal spatial representations, a transformation that may be time consuming and error prone. Based on the results of prior work, a fixed 3D display strategy may also fail to satisfy users' need to process information efficiently or effectively because it provides superfluous information that may distort some perceptual judgments [4] and [35]. Therefore, two possible strategies seem reasonable to take advantage of the benefits of both formats — synthesizing 2D and 3D views together [37], or providing geospatial information both in 2D and 3D displays and allowing users to choose. Focusing on the latter approach, our study suggests that these benefits will be realized if novice EMIS users are given guidance regarding the importance of cognitive fit and methods for choosing display formats for different tasks. Such results not only provide EMIS developers with information regarding potential problems associated with adaptable displays, but also show the significance of providing appropriate decisional guidance. 6.1. Limitations As in any experimental work, this study has several limitations. Creating an experimental scenario that is consistent with its real-world analog is important for ecological validity. Nevertheless, replicating all elements of a crisis situation is difficult, if not impossible. Although our study participants were not facing a real crisis, we simulated an emergency situation by asking them to undertake the roles of incident commanders and by emphasizing the importance of quick and accurate responses. To simplify and expedite the experiment, we adopted the paper-prototyping method [34] and provided DG via paper and pencil rather than by actual computer-based or interactive EMIS. Nevertheless, because we intended Experiment 2 to examine the influences of predefined, prospective DG on novice EMIS users' decisional performance, we believe that the contextual differences between those two methods did not seriously weaken our findings. Future researchers who are interested in the dynamics or interactions between DG and users are encouraged to develop or modify an actual EMIS system. In addition, we developed our Experiment 2 questions for each task based on the existing visual displays in Fig. 1. Although we used questions that we felt were practical and that our research participants would relate to, future research could certainly test other tasks. Developing more questions would certainly contribute to establishing content validity as well. This research examines the influences of a prospective DG on EMIS users' decision making. Future should investigate the impacts of alternative formats of DG (e.g., concurrent). To increase the task complexity and remain consistent with prior work [4], Experiment 1 intentionally implemented two plan view examples that differed by rotation (i.e., Panels 1 and 2, Fig. 1), two elevation view examples that differed in layout (i.e., Panels 3 and 4, Fig. 1), and one example for 3D view. Research participants provided informal feedback that revealed possible confusion and unnecessary cognitive burden, especially when they were using the elevation views. Moreover, Experiment 1 was intended to investigate whether novice EMIS users can choose display formats with the best cognitive fit. Therefore, statistical analyses of Experiment 1 grouped five displays into three types, and Experiment 2 investigated the impacts of DG by contrasting plan and 3D views only, dropping elevation view. Such change is a trade-off between the task consistency of two experiments and their research objectives. In other words, Experiment 2 is intended to document the influences of DG on novice EMIS users' decision making, not to determine how they react spontaneously to different display formats. 6.2. Implications for practice Recently many 3D EMIS have been developed to help incident commanders make decisions [9]. Researchers have urged more thought and effort to identify ways to design and develop EMIS, resulting in frameworks, architectures, and design principles that are helpful to the study of EMIS [4], [9], [10], [11], [12], [19] and [48]. Our study indicates that, as prior studies have cautioned [7], [38], [39] and [42], practitioners should use new technologies judiciously, with appropriate guidance regarding when to use them. To enhance flexibility in interactive systems, Hansen [13] cited eight user engineering principles that formulated the ingredients of data-entry interaction. One of these principles suggested that system developers should change displays as little as possible to carry out users' requests, as users' display selection inertia may influence the functionality of the system [27]. Nevertheless, display selection inertia and its associated concept, familiarity bias, have not been addressed in the DSS literature [p. 1, 22]. We found evidence that DG could help novices select from various display formats and could alleviate the familiarity bias, consistent with the findings and suggestions of Marett and Adams [22]. Such finding may help system developers utilize DG to transition their users to a new DSS with modified or improved interfaces or display formats. The results of Experiment 1 suggest that making task-display compatibility decisions is not intuitive and that EMIS users, when given a choice, might fixate on a specific display format (i.e., 2D or 3D). Thus, before implementing or propagating EMIS that provides adaptable or hybrid displays, novice and even experienced incident commanders must be trained and supported to make the most efficient and effective decisions. 6.3. Implications for future research This research is one of the first studies to examine the roles of DG and adaptable 2D/3D displays in crisis and other decisional situations, which opens further research avenues to determine when and how to provide this guidance. As we used a paper-based DG in our study, a first step would be to replicate the results in a computer-based EMIS and determine whether DG also contributes in users' decisional performance in this context. Second, as Silver indicated [29], it is possible to examine various formats of DG that may further improve decisional performance. In this research, we provided DG before novice users made their task-display compatibility decision, similar to traditional orientation or training. Future research should investigate the impact of other types of DG, such as advising EMIS users while they are making decisions, providing concurrent (e.g., pop-up windows) or on-demand DG, or offering other automatic reminders (e.g., “tips of the day”). These alternative forms of DG may lead to different decisional performance or outcomes. Third, in addition to adaptable displays, various methods have been used to develop EMIS or other DSS. For example, some DSS synthesize multiple 2D/3D views into one [37]; whereas others integrate multidimensional data into a single view (e.g., 4D spatial-temporal displays advocated by [26]). How users interact with those innovative systems and whether DG improves users' decisional performance require more explorations. Fourth, this study examines the selections of “fit” decision aids and focuses only on one attribute — the dimensionality of visual-spatial representation. As decision making becomes more complicated, many DSS have been developed to incorporate or integrate more data into one display. Whether DSS users can choose the most appropriate visual displays from various alternatives deserves more investigations. For example, Google Maps users can easily add more layers to the 2D or 3D maps (e.g., traffic, real estate, webcam). Such multivariate displays are increasingly common in safety-critical domains such as air traffic control, computer-aided surgery, and chemical and physical process control. Unnecessarily adding layers, functions, or colors may increase complexity and inadvertently decrease decisional performance. In fact, a growing literature exists on the benefits of display “decluttering” algorithms given inherent limitations in the spatial resolution of human attention [45] and [46]. As a parallel to the present research, the question will be whether decisional guidance can be used by decision makers to determine how to best apply such information presentation options. Fifth, the effect size of this research deserves further investigation. Specifically, this research assumes that failing to choose the display with the best cognitive fit may result in suboptimal performance or other consequences. How severe these negative influences are and whether users can be trained to avoid these problems remain unanswered for all stakeholders. Sixth, incident commanders work under stress. The role of time pressure and whether it mediates or moderates users' decision making requires more empirical investigation. On one hand, time pressure may lead incident commanders to make decisions chaotically and thus show no clear pattern of display choice; on the other hand, the demand for quick decisions in crisis situations may make users less inclined to manipulate display formats, suggesting higher fixation or display selection inertia. In addition, how DG works under time pressure deserves further empirical investigations. Seventh, an important but unanswered question is whether the demographics (e.g., age, gender) or personal characteristics (e.g., openness to new technologies) of EMIS users influence their display format choices. We encourage future research to determine the independent and joint contributions of DG and individual differences such as experience with different formats and differences in visual-spatial processing capacity. Last, from a theory perspective, our study extends the vast amount of research on the IS theory of CFT and shows that it is appropriate to design displays for incident commanders. The use of CFT and DG (from MIS theories) and PCP (from human factors and cognitive psychology) also shows that technology design and effectiveness evaluation can be enhanced by adopting interdisciplinary perspectives. In conclusion, ongoing advances in computational power and information technology are improving the performance, availability, and cost effectiveness of 3D visualization techniques. Whether a greater quantity of EMIS display choices will improve the quality of decision making is an important practical and academic question. Through two experiments we provide evidence that novice EMIS users, without guidance, often fail to choose the display format with the best cognitive fit. However, even simple decisional guidance improves their performance.