[ Home ] [ Up ] [ Project Summary ] [ Project Description ] [ General Plan of Work ] [ Detailed Work Plan ] [ References ]

To study the nature of CIR, the researchers will use a work-centered approach. Generally speaking, work-centered approaches are concerned with designing technologies that support, facilitate, and enhance daily work practices. In the study proposed here, this approach will view CIR holistically, taking into account the total team and organization’s culture, examining the teams’ and team-members’ goals and priorities, how they share information, and their personal and collective procedures and routines. Various frameworks have been used in work-centered approaches. Web models (e.g., Kling and Scacchi, 1982) were an early approach directed at establishing requirements for information systems. Recent frameworks include Jones (1995) and Ford and Wood (1996). We have experience with the conceptual framework for work-centered evaluation and design developed by Annelise Mark Pejtersen, Jens Rasmussen, and colleagues at the Risoe National Laboratory in Denmark (Pejtersen, 1985; Rasmussen, Pejtersen & Goldstein, 1994, Pejtersen et al., 1998).

 

This conceptual framework was developed as a general approach to help information system designers analyze and understand the complex interactions between (a) the activities and organizational relationships and constraints of work domains, and (b) users’ cognitive and social activities and their subjective preferences during task performance. The framework is based on generalizations from field studies which led to the design of support systems for a variety of modern work domains such as process plants, manufacturing, hospitals, and libraries.

The work-centered approach to evaluation and design assumes that information interactions in the workplace are determined by a number of dimensions. The first step towards evaluation is the development of a framework based on examination of dimensions of work-based information interactions, such as the actual work environment; the work domain; the task situation in terms of work domain, in terms of decision making, and in terms of mental strategies; the organization within the team and within the larger unit in terms of division of work and social organization; and the characteristics, resources, and values of individual work-team members.

The framework for system evaluation, which is based on this analysis, answers questions such as these: Does the system support the cooperation and coordination necessary? Does the system support all the tasks that work-team members perform? Does the system support the members in decision making? Are all relevant strategies supported? Does the presentation of information match sensory characteristics? The answers to these questions are based on the data collected during the initial examination, which is primarily descriptive. This serves as the framework for subsequent data collection, and the system evaluation is the framework for data analysis.

The Framework for Work Analysis has been successfully applied to the design and evaluation of structures that support information retrieval. In the past, it guided the development of the first retrieval system for fiction called BookHouse (Pejtersen, 1992). In addition, it was used to analyze the data collected in a recent study of Web searching by high school students (Pejtersen & Fidel, 1998, Fidel et al., 1999) and in a project at the Boeing Company that was aimed at improving the design of that company’s Web-based intranet (Fidel & Efthimiadis,1998). In both studies, the framework proved to be very powerful in helping to uncover the problems that people experience when using the Web to search for information.

Examples from the high school study can illustrate how we intend to apply the Framework to the proposed study of CIR. In the high school study, Fidel was interested in finding out how the Web supports user strategies. The analysis revealed that students often had problems in applying a search strategy and therefore frequently shifted from one strategy to another. Similarly, with regard to the Web’s ability to support cognitive decision tasks, it was found that students had many problems with analyzing their information needs before and during searching. In both cases - needs analysis and search strategizing - a great deal of interaction resulted among the students themselves and with the teacher and librarian. These interactions tended to disrupt the search process because students had to move away from the terminal either to join a colleague at another terminal, or to join the teacher and the librarian in another room.

Following the Pejtersen framework, the next step in this high school study was to consider the development of organizational structures and facilities on the Web that would make it possible for students to communicate with their colleagues, teacher, and librarian in the ways they wanted to. The study recommended that these sorts of improvements should also be directed at helping students to execute their strategies, shift easily from one strategy to another, and to analyze their information needs. It would appear such facilities have never before been considered in studies about the Web in general or in educational settings, but they make the information on the Web much more accessible to these high school students.

While rich and promising, this work-centered framework is designed to examine individual IR, rather than collaborative IR. One of the study objectives, therefore, is to extend the work-centered framework to include CIR.

Two of the present team have assessed and extended frameworks for collaboration in the context of technology support (e.g., Grudin, 1994; Grudin and Poltrock, 1997). The table below (from Poltrock and Grudin, 1999) identifies fundamental social activities engaged in by teams and organizations-communication, information sharing, and coordination-and specific technologies that support these, either in real time or asynchronously. Although a specific activity or technology may span these categories, these distinctions have served as a useful assessment tool and have guided design and implementation. CIR is likely to span communication, information sharing, and coordination, and could present complex challenges.

The first phases of this research involve observations of four work teams (two from Microsoft and two from Boeing) and the analysis of their collaborative information retrieval activities. The researchers on this project from each company will identify and recruit the participating work teams.

Selection of teams for this study will be based on an analysis of a set of information intensive work scenarios and a description of their similarities and differences with respect to characteristics that shape the collaborative work structure and the related information retrieval patterns.

The teams selected for this study will be engaged in knowledge work that requires access and integration of diverse information. Their work will not be principally information retrieval; this will be a supporting activity of their work. Likely candidates for selection are teams engaged in defining or developing new products. These would include teams developing new software products at Microsoft or new aerospace products at Boeing. The selected teams will be in, or about to enter, a phase of their work that involves intense information gathering. Examples include conceptual design or product definition.

The teams will be selected to explore the effects of variables that are known to influence collaboration. One such variable is physical proximity. Product development teams generally include some people who are physically collocated in the same room or at least the same section of a building and have frequent face-to-face meetings. These teams often include other people at remote sites who participate in the team through use of technologies. Therefore, the study will investigate both collocated and distributed situations.

The selected teams will be engaged in a full-time project, but some members of the teams may participate part-time. For example, a quality control expert in an aerospace team may simultaneously be a member of several teams. Therefore, the study will consider both full-time and part-time members.

The study will use a variety of instruments depending on the specific phase of the research activity. The first four phases will be based on observations and interviews in the actual workplace settings where the work-team is functioning. Before going to the field, researchers will conduct a general study of work scenarios from available documents such as annual company reports, organizational charts, work manuals, procedures and instructions, and published work analyses. Members of the research team will accompany work-team members during their workday. The researchers will sit in team meetings taking notes about the activities of the team and the interactions that occur between work-team members. Where possible, they will collect data on the interchanges that occur on e-mail between work-team members. They will observe work-team members as they seek information individually or in pairs or groups, asking them to think aloud. Finally, the researchers will interview each work-team member to complete the work analysis for the work-team and the individual members. All verbal protocols, including think-aloud and interview protocols, will be transcribed and made available to all members of the research team.

To analyze the data collected in the observation of the first work-team the research team will get together for a couple of days to begin the process of system evaluation for this work-team, that is, the evaluation of present retrieval tools and practices. This analysis will continue until the evaluation is completed, and the research questions are answered for the team. After the observation of the second work-team, the researchers will get together again to begin the system evaluation of the second work-team and to integrate it with that of the first. This will continue until an evaluation that pertains to both teams is completed. After the fourth iteration, this system evaluation process will pertain to all four work-teams. The research team will then develop the survey instrument and interview schedule that will be used in phase 7 of the study.

This research is exploratory in nature. The researchers cannot be certain about what they will find when they observe work-teams using the techniques of work analysis. But the aim is to find out if CIR is a part of the functioning of work-teams in organizations and, if so, what the dimensions and manifestations of CIR might be. The development of a general theory of CIR based on these observations and the surveying of work-teams in the U.S. and Denmark will hopefully provide the basis for the conceptual design of technologies that can be used to support and facilitate the processes that have been identified empirically. This may involve the re-design of existing technologies or recommendations for an entirely new vector for the development of technologies that are needed to support the processes of CIR.

In spite of the exploratory nature of the research, it is possible to illustrate the methods and procedures that the researchers intend to use with a scenario. Let us suppose, for example, that during the observation period, an engineer in a work-team indicates that when he wants to consult with someone about information seeking he always turns to one of two work-team members. During an interview, the engineer may explain that he feels comfortable with these two members and he does not like to bother other people. The researcher might also observe that the engineer’s consultations do not always help his information seeking to progress on a fruitful path. This scenario is recorded during the work analysis stage of the investigation when the researcher describes the task situation in term of mental strategies.

The system evaluation stage of the investigation focuses on a different question -- Does the system support all relevant strategies for information retrieval? In the case of this engineer, the answer would be no. His strategy of consulting with other work-team members is not supported because only two members of the team are available to him for consultation.

A member in the second work-team to be observed by the researchers may display a similar behavior, only she may explain that she is consulting only her friends because she does not want to appear to be stupid. This behavior is repeated in other teams and possibly for other reasons. What is clear in the evaluation is that some team members do not receive all the support they need for CIR. In particular, they receive no support for consultation with other team members.

The survey instrument may reveal the extent of this problem, and may also provide additional data about the factors that affect this behavior. At the same time, several solutions can be considered. For example, suggestions for a more supportive environment within a work-team for CIR can be developed, say by weekly short meetings to discuss IR issues. Technological solutions might also be recommended. For example, a database of information strategies that have been used by work-team members could be constructed to make these strategies available to all work-team members. In addition, an electronic communication system could be established for work-team members to pose questions to one another and to suggest answers anonymously or otherwise.