Gove Nathaniel Allen, Ph.D. Assistant Professor of Information Systems, Brigham Young University   


Statement of Research Interests

Humans have significant strengths and weaknesses with respect to information processing. Some of the weaknesses are particularly evident in everyday life. They forget birthdays and anniversaries, they have great difficulty dealing with more than seven (plus or minus two) units of information at a time, they have difficulty making other than linear projections, virtually all of their memory is subject to decay, they remember things that never happened, and they ignore new information which contradicts previously obtained information. Computer-based information systems have been developed to help deal with many of these weaknesses.

However, humans also have significant strengths with respect to information processing that are particularly problematic for computer-based information systems. Researchers in the fields of cognitive science, neurophysiology, psychology, and others have identified many such competencies. For example, humans are particularly adept at pattern recognition (Clark 1995). This capability is evident in the human capacity to differentiate and interpret facial expressions. Infants only a few months old are able to discern among different classes of facial expressions (Nelson and De Haan 1997). Another human information processing competency is evident in the verbal development of children, who begin life without any language skills and are speaking in nearly complete sentences by age three (Pinker 1991).

Another human information processing competency is seen in our ability to deal with information when it is presented in the format of a story—or as current research calls it, a narrative. Brunner (1990) has identified that information is shared effectively through the use of narrative. Robinsin and Hawpe (1991) further identified that humans use narrative for sensemaking. Orr (1990) demonstrated that narrative is an effective tool for communicating tacit knowledge. Boland and Tenkasi (1995) discuss that narrative and "perspective taking" can be effective at spanning the boundaries between communities of knowing. They further discuss ways of analyzing narrative and identify that narratives can be broken into a sequence of events of varying impact to the overall story. Further Tulving (cited by Stien and Zwass 1995) discusses that human memory is organized into event and skill memory. The work of Shenk (1991) supports the notion that events are fundamental to the way that humans organize information about their environment.

The human information processing competencies associated with narrative structure as well as the evidence that human memory has specific structures for dealing with event information lead me to pursue the area of narrative/event structure as tool for improving the development of information systems. I do not intend to develop new methods for implementing information systems; rather, I intend to use the theories that underlie the narrative competency in the comparison of existing methodologies that make use of the ideas of event-organization and narrative in varying degrees.

Data modeling is a specific area of information system development in which event-organization and narrative are particularly significant. A data model is intended to represent the "semantics" or "natural structure" of "things" in the domain serviced by the information system. An analyst must determine these semantics through interaction with the community of users. Methodologies proposed to accomplish this vary in their emphasis on events and narrative.

The most commonly used methodologies, as summarized in McFadden and Hoffer (1985), organize their data representations around "things," such as Customers, Employees, Orders, and Inventory Items. Although they acknowledge that events change the states of these "things," they often do not represent events explicitly, nor do they emphasize the business activities (narrative) supported by the data model. McCarthy (1979), on the other hand, proposed a methodology in which events are represented explicitly in the data model. "Things" are significant insofar as they are affected by events; agents (such as employees and customers) are significant because the participate in events. The resultant data model reflects the story of how events affect the domain supported by the information system.

Addressing the task of data modeling from an event-organization and narrative perspective should be a fruitful area for research for several reasons. First, in the development process, entity-relationship diagrams are often used as a communication tool between analysts and users to make sure that the analyst has adequately understood and represented the relevant data necessary to manage a business process. Because the entity-relationship diagram is used as a tool of communication, aligning it with narrative through event-organization seems like an effective idea. Second, since a major element of organization memory is contained in the organization’s databases, it makes sense that interfaces between organizational databases and human decision makers which facilitate sensemaking can improve organizational memory. Third, since people use narrative for sensemaking, analysts (whose main task is to make sense of the user’s needs) who use a methodology which parallels narrative should be more effective than those who do not.

Companies spend vast amounts of money on the development of information systems. All too often, the systems developed do not meet companies’ needs. There are theoretical reasons to believe that event-organization and narrative aid human understanding of complex phenomena. Business phenomena for which information systems are developed can be extremely complex. Analysts and users must both understand them. Event-organization and narrative should facilitate this process. If it does, then this type of methodology should result in more effective system development efforts and more effective information system utilization.

ibm