A USER-BASED COGNITIVE APPROACH TO MODELING HIGHLY DYNAMIC INFORMATION PROBLEM SITUATIONS
University of North Texas
Information seeking and use situations for today's users present innumerable choices and constant, rapid change. This paper proposes a cognitive approach to modeling users' perceptions in situations in which information content, information sources, and users themselves are constantly moving in time and space. The temporal-spatial model involves users' orientations not only toward external reality (the world) and internal reality (themselves), but also toward their power to exert control in accessing information as they travel through their environments. Their perceptions affect the way they describe their situations and evaluate and use information and information sources. Empirical support for the model is provided by the results of a study of weather information users employed in construction, electric power utilities, and aviation. Implications for development and application of the model are discussed.
Information seekers live in a real world of time and space, a world in which they must make quick decisions based on rapidly changing information from a variety of sources. The challenge is particularly evident in critical situations, such as air traffic control, that require immediate access to specific kinds of spatial-temporal information. In attempting to understand user behavior in such situations, researchers must sort out complex, shifting relationships among a wide variety of factors. The factors involve characteristics of users' situations (problem constraints, information sources), information (content, spatial-temporal attributes), and users themselves (knowledge base, situational role).
The purpose of this study was to begin the process of identifying the most important temporal-spatial factors from a user viewpoint, by examining users' descriptions of their problem situations and their evaluations of information and information sources. The users were people employed in weather-related occupations who sought information about weather conditions from many sources. Their responses suggested a preliminary temporal-spatial model of factors involving users' perceptions of external reality, internal reality, and control. The model holds implications for further development, along with possible applications in researching users and behaviors in a number of occupational fields involving highly dynamic situations.
The approach of this study was cognitive and situational, based on users' perceptions of temporal and spatial aspects of their information problem situations. Generally, cognitive approaches assume that users rely on perceptions of both external factors (e.g., situations, systems) and internal factors (e.g., values, goals) that affect them as individuals. In information science, several writers have viewed user perceptions as manifested in relevance judgment criteria, where relevance is defined in a holistic sense as concerning a dynamic relationship between information and the user's information problem situation (see Harter, 1992; Park, 1993; Schamber, 1994).
A particularly useful holistic approach was offered by Taylor (1986) in his value added processes model of information systems. The model involves three components of the information use environment: the user, who has an information problem that establishes the criteria of choice; the interface or negotiating space, in which values added by the system assist the user in making choices; and the system, where specific processes add specific values. Taylor presented a progression of values, from user to system, that are increasingly concrete and applicable to system design. For example, one user criterion is ease of use, for which an interface value added is formatting, for which a system value-added process is highlighting important terms. Taylor's approach is inherently dynamic. Among other things, he viewed the user as constantly monitoring and evaluating a variety of information sources. He was perhaps the first writer in the information science literature to suggest a dynamic criterion for interactivity and to describe reliability as a value developed over time. Although Taylor's criteria were not empirically tested, they have since been validated by the results of several recent studies in which criteria were derived directly from users (see Schamber, 1994).
Researchers in other disciplines provided further background for the temporal spatial focus of this study. A recent example is work in cognitive psychology by Tversky, Franklin, Taylor and Bryant (1994), who described a series of experiments in which they studied mental models of space based on individuals' verbal descriptions of spatial environments. Among other things, they tested distinctions between route (landmark-orientated) and survey (aerial-orientated) conceptualizations of space and tested the relative difficulty of dealing with scenarios in which either observer or environment was moving.
Researchers in communication and management provided approaches to understanding source choice behavior. For example, the media richness model proposed by Trevino, Lengel, and Daft (1987) explains communication media choice behavior in terms of four criteria related to the potential of a medium to resolve task ambiguity: speed of feedback, variety of communication channels employed, personalness of source, and richness of language used. In the theory of adaptive structuration by Poole and DeSanctis (1990), individuals' choices of technological features are linked to the dynamics of group decision-making in a series of stages of interaction. Where multiple information sources are important in decision making, models of channel selection, as discussed by O'Reilly (1980) and Swanson (1987), were helpful.
Not surprisingly, most research on spatial information has been conducted in fields such as meteorology, geology, environmental science, aerospace, architecture, and computer science. Only recently have information scientists shown a substantial interest in the unique characteristics and uses of spatial information. Geographic data such as elevation and latitude, for instance, not only define content and orient the user, but also allow analyses of data from multiple sources overlaid in visual displays. Temporal data allow monitoring and analyses of variations over time using animated displays.
Generally, the trend in information science has been toward holistic cognitive approaches to understanding information behavior, with acknowledgement of the dynamics involved but little empirical research on actual roles and effects of time and space on various behaviors. Clearly the number and complexity of potential factors that can be considered in research on spatial information is enormous. This study was intended to begin the process of reducing the list to manageable proportions by modeling users' perceptions of important spatial-temporal factors in their own dynamic situations.
METHODOLOGY AND RESULTS
The data were collected through open-ended time-line interviews of 30 individuals employed in three weather-related fields: 10 each in construction, electric power utilities, and aviation. Respondents were each asked to describe the events in one situation requiring a decision that depended on information about the weather. For critical events in their situations, they were asked to discuss all questions asked and answers received, then to evaluate the quality of weather information sources consulted and how those sources presented information. The audiotaped interviews were transcribed and the texts content-analyzed in order to identify and define respondents' perceptions. Results consisted of descriptive data concerning respondents, situations, information, sources, presentation formats, and evaluation criteria (see Schamber 1991a; 1991b). The perceptions of respondents in the model below are derived from further analyses focused only on temporal-spatial aspects of these data.
THE TEMPORAL-SPATIAL MODEL
The preliminary model, shown in column 1 of Table 1, was suggested by the tendency of respondents in the dynamic weather-related occupations to frame most of their responses in terms of temporal and spatial constraints of their information problem situations. The constraints seemed to affect how they perceived, evaluated, and chose weather information sources as well as information content.
|Cognitive Orientation||Occupational Field|
|Situational Factor||Construction||Electric Utility||Aviation|
|World Orientation (external)|
|Information Content||Precip. (yes/no), temp. (± 32¤ )||Precip.(kind), temp., lightning, humid., storm movement, flood||Precip. (kind), temp., wind, clouds, barom. pressure at diff. altitudes|
|Time frame||Day to season; future, some past||Day to year; future, some past||Hour to day; future|
|Space frame||Terrestrial; local job site||Terrestrial; region, counties||Terrestrial, aerial; state to continent|
|Self Orientation (internal)|
|Knowledge Base||High School. No meterological educ.||College. Informal meterological educ.||Graduate or professional. Formal meterological educ.|
|Comprehension||Plain language; text format sufficient||Meterological language; text and graphic formats||Meterological language, acronyms; complex text and graphic formats|
|Control Orientation (external/internal)|
|Source Choice||Self, colleague, mass media, public-access systems||Self, colleague, mass media, some specialized systems||Self, colleague, extremely specialized systems|
|User movement||User static, stays in one location||User static or active, may travel in region||User actively moving or directing traffic through horizontal and vertical space|
|User-source interaction (beyond self, colleague)||static or dynamic; noninteractive sufficient||static or dynamic; some interactive specialized systems||highly dynamic and interactive (including human experts in specialized systems)|
The model takes into account respondents' cognitive orientations, or perceptions of internal and external reality, and prominent temporal-spatial factors in their situations. World Orientation refers to external phenomena; in this study, to information about the weather. Self Orientation refers to internal phenomena, including one's knowledge about the external world. Control Orientation serves as a bridge in that it refers to one's internal ability to exert control in the external world, including the ability to control sources in accessing information.
Under each orientation category are situational factor categories (column 1) and situational data for each occupational field (columns 2, 3 and 4). These were drawn from temporal-spatial aspects of respondents' problem descriptions and defined in part by their evaluation criteria. Differences among occupational groups were evident in nearly every factor category. Most differences appeared to represent a continuum of increasing complexity, from construction to electric power utilities to aviation.
World-Oriented Situational Factors
Respondents' perceptions of the external world pertained to information about physical phenomena -- weather conditions -- within certain constraints of time and space. Their problem situations all required planning decisions. In construction, situations involved the protection of workers and materials during winter construction projects; in electric power utilities, the scheduling of line maintenance and repairs; and in aviation, the scheduling and routing of airplane flights. The weather conditions, all in the Northeast, ranged from violent thunderstorms in the summer to a hurricane in the fall to the coldest winter in a decade.
Requirements for information content ranged from data on only two weather conditions in construction -- precipitation and freezing -- to numerous weather conditions, including conditions by altitude, in aviation. In evaluating information quality, all respondents mentioned the importance of accuracy (proven over time by actual conditions) and specificity of weather data.
Time frames dominated respondents' concerns in these dynamic situations. Decisions affected many people, and they had to be made quickly in response to rapid changes in the weather. For example, a utility employee said that scheduling repair work on power lines can "give you ulcers simply because you've got a time frame of three or four hours to do this work. You've got to shut the electricity off, you're going to have these crews in, you've done all this background work, you've got other departments involved, you've got commercial people off, a lot of loss of revenue, they're losing business -- all of this heavy stuff coming down." The sense of urgency was heightened by the fact that decisions ultimately had life-or-death significance. Safety was constantly mentioned as the primary consideration. As one respondent put it, "We kill people when we make mistakes." Adding further to the stress level, they admitted, was the fact that any decisions based on the weather are uncertain at best.
All three occupational groups looked to the future in weather forecasts, but only respondents in aviation, who dealt with the shortest (hour to day) time frames, did not also require historical data for planning purposes. The importance of time was reflected in all respondents' evaluations of currency in weather information.
Space frames were determined by respondents' geographic areas of responsibility and the location of impending weather conditions. Construction was the most localized geographically, whereas aviation was the least localized, involving aerial as well as terrestrial space. All respondents mentioned the importance of obtaining data on the geographic proximity and speed of approaching weather conditions.
Self-Oriented Situational Factors
Respondents' internal perceptions pertained to themselves; to their existing knowledge -- gained over time -- and to their ability to comprehend information. Within this orientation, they tended to evaluate information sources and presentation formats more than information content.
In Table 1, knowledge base is shown in terms of education levels. General education ranged from high school in construction through graduate or professional studies in aviation. Meteorological education ranged from none in construction to formal in aviation; respondents in electric power utilities had a mixture, often including informal (on-the-job) training in reading weather instruments. Not shown is amount of experience, because nearly all respondents were highly experienced, with an average of 20 years in their fields. Both education and experience helped them not only to read weather instruments and use specialized information systems, but also to serve as their own information sources when directly observing weather conditions. In making evaluations, they often mentioned reliability of the source, based on their own and others' expertise gained over time and/or on their experience with a source's track record for delivering good data.
Another self-perception concerned the ability to comprehend information depending on the way it was presented. As the types of data required became more complex, from construction through aviation, so did language and formats. Clear information was described as having simple terminology, good organization, and well-labeled graphics. Respondents in aviation said standardization of jargon and acronyms was important in minimizing the time it took to convey information. Respondents in general expressed preferences for presentation formats that would help them comprehend information quickly in order to make decisions. One said, "I can look at a weather printout this long, and in a minute or two tell you all I need to know from it. Whereas, if I have to listen to somebody over the phone, I have to listen to it and think about it. It cuts your human data processing time by an extremely large percentage."
Control-Oriented Situational Factors
Neither actual weather conditions nor information about the weather were really controllable by respondents. However, they did exert a great deal of control in positioning themselves in time and space and in choosing and using information sources to access information. In the Control Orientation, they seemed to value sources and presentation formats (external perceptions) that allowed them more power and flexibility (internal perceptions) in making choices.
The weather-related context included seven types of sources. (1) Self was the respondent, usually observing actual weather conditions. (2) Colleague involved personal interactions with others in their fields, often co-workers who were not weather experts. Mass media sources included (3) newspaper, (4) radio, and (5) television. (6) Weather instruments ranged from windsocks to radar. (7) Weather information systems included public-access systems, such as telephone and radio recordings, and dedicated special-purpose systems or services. Specialized systems for power utilities included subscription weather information services and a computerized power load forecasting system. In aviation they included air traffic control, flight briefing services, and computerized systems.
Only the first two sources, self and colleague, were used by all three groups. Otherwise, both the number and sophistication of sources increased from construction to aviation. Public-access weather information systems were consulted only in construction. Mass media sources were consulted in construction and electric power utilities. Specialized systems were consulted in electric power utilities and aviation. All respondents consulted multiple sources during a given problem situation. A source was valued if it delivered information that verified or confirmed information from other sources over time. Verification over space often involved a respondent being at a location directly observing and feeling weather conditions.
The increasing sophistication of sources by occupational field also seemed to correlate with increased physical movement, with respondents in construction traveling the least geographically and those in aviation traveling the most, horizontally and vertically. Respondents constantly consulted different sources as they monitored changing weather conditions and as they themselves changed locations in relation to the weather. During the three most critical events in their situations, they described, on average, six instances during which they sought information from three different types of sources. In their overall situational time-lines, about a third of the respondents included events beyond the workday and workplace, indicating that they checked the weather from the moment they woke up until they went to bed at night. They often evaluated sources in terms of accessibility, referring to the availability and ease of using of sources at any time or place. Additionally, respondents who traveled the most, as in aviation, tended to prefer the most sophisticated graphic formats, including three-dimensional weather data displays.
Finally, the ability to exert control in user-source interactions emerged as a very important consideration in the weather-related situations. The most dynamic and interactive sources were human: self and colleague in all three fields, and experts as part of specialized systems (e.g., flight briefers) in aviation. Otherwise, the use of sources with dynamic and/or interactive capabilities increased from construction through aviation; only construction relied to any extent on newspapers, the one source that could be said to be static. Perceptions of control involved both information content and presentation format. With human sources, obviously, interpersonal communication influenced content. All other sources except weather instruments mediated, in one way or another, information from the National Weather Service. Only in aviation, where systems were extremely specialized, did individuals also contribute information to the National Weather Service through pilots' in-flight reports. Presentation format could be affected several ways. In specialized information systems, control could range from a simple choice of output format, such as disk or hard copy, to some sort of manipulable real-time data display. Respondents in electric power utilities and aviation expressed the desire for electronic systems with dynamic interactive displays including tracking, projection, and zoom capabilities.
CONCLUSIONS AND IMPLICATIONS
The focus of this study supports an increasing interest in temporal-spatial information among information scientists, along with an ongoing interest in improving information systems and services through a better understanding user needs and behaviors. In the weather-related occupations, users' information problem situations were more dynamic and their uses of multiple sources more complex than those of users in the text-based information retrieval settings traditionally studied in information science. It was the pervasiveness of concepts of time and space in users' responses that inspired the development of the preliminary temporal-spatial model presented above. In addition to external (world-oriented) and internal (self-oriented) perceptions, the model involves perceptions of external/internal control that emerged from users who themselves were traveling in space and time and relying on multiple information sources.
Although at present the model can be said to oversimplify or underestimate the variety of factors involved (a detailed data analysis is currently underway), it does provide a framework for development through further user-based cognitive research. Future studies, for instance, might focus on collecting data from inexperienced as well as experienced users in a variety of dynamic situational contexts.
In this study, the apparent continuum, from simple to complex, in the situational factors from one occupational group to the next does suggest tangible parameters for designing task-oriented information systems. The data under World Orientation, for example, hold implications for content and organization of information, whereas the data under Self Orientation suggest the provision of choices geared toward individual differences among users. Perhaps most important, the data under Control Orientation suggest that mechanisms be provided to allow user control of the content and amount of information, choices of presentation formats, and links among multiple sources. The data also suggest that control involves interactivity, or the ability to engage in two-way communication with sources. Several respondents expressed the desire for sources -- human or computerized -- that would allow them to explain their problems in a process akin to query negotiation or the reference interview; to engage in a dialog that would provide feedback as their needs changed.
The process of translating user-defined perceptions into system specifications was suggested by Taylor (1986) in a series of steps. Any field that deals with fast-changing information and time-critical decisions might benefit from the translation of results from studies such as this one. Beyond construction, electric power utilities, and aviation, fields that depend on weather and other kinds of dynamic spatial information include disaster and rescue services, military analysis and command, law enforcement, transportation and shipping, recreational travel, wildlife management, and journalism.
It can be posited, generally, that users who are the most spatially active themselves will require the most sophisticated and interactive information systems. Designers might consider combining appropriate content and display features in portable sources that could be carried by users monitoring information in the field. Finally, considering the sophisticated manipulations of spatial-temporal data required for analysis and interpretation (e.g., visual overlays, animation), any expert intelligence that can be built into such systems would make benefit nonexperts as well as experts.
Harter, S. P. (1992). Psychological relevance and information science. Journal of the Association for Information Science, 43(9), 602-615.
O'Reilly, C. A., III. 1980. Variations in decision makers' use of information sources: The impact of quality and accessibility of information. In R.C. Huseman (Ed.), Proceedings of the 40th Annual Meeting of the Academy of Management (pp. 183 187).
Park, T. K. (1993). The nature of relevance in information retrieval: An empirical study. Library Quarterly, 63(3), 318-351.
Poole, M. S., & DeSanctis, G. (1990). Understanding the use of group decision support systems: The theory of adaptive structuration. In J. Fulk & C. Steinfeld (Eds.), Organization and communication technology (pp. 173- 191). Newbury Park, CA: Sage.
Schamber, L. (1991a). Users' criteria for evaluation in a multimedia environment. Proceedings of the 54th Annual Meeting of the Association for Information Science, 28, 126-133.
Schamber, L. (1991b). Users' criteria for evaluation in multimedia information seeking and use situations. Unpublished doctoral dissertation, Syracuse University, Syracuse, NY.
Schamber, L. (1994). Relevance and information behavior. In M.E. Williams (Ed.), Annual Review of Information Science and Technology, 29, 3-48.
Swanson, E. B. (1987). Information channel disposition and use. Decision Sciences, 18, 131-145.
Taylor, R. S. (1986). Value-added processes in information systems. Norwood, NJ: Ablex.
Trevino, L. K., Lengel, R. H., & Daft, R. L. (1987). Media symbolism, media richness, and media choice in organizations: A symbolic interactionist perspective. Communication Research, 14, 553-574.
Tversky, B., Franklin, N., Taylor, H.A., & Bryant, D.J. (1994). Spatial mental models from descriptions. Journal of the Association for Information Science, 45(9), 656-668.