Conceptions of task as a methodological issue: Scandinavians on information seeking and retrieval research (SIG USE)
Katriina Byström, Louise Limberg, Annelise Mark Pejtersen, Nils Pharo, Olof Sundin, Nicholas J. Belkin, Carol Collier Kuhlthau · Proceedings of the American Society for Information Science and Technology · 2004
During the last decade, there has been a call for empirical research that combines the traditionally mutually exclusive research areas of information needs, seeking and use (INSU) and information retrieval (IR) (e.g., Belkin, 1993; Ingwersen, 1996; Vakkari, 2003; Järvelin & Wilson, 2003). The lack of common ground that would provide an appropriate point of departure for both of these research areas has been one of the main reasons for difficulties in merging them. The session will focus on one possible solution, namely, the empirical (work) tasks based research on INSU and IR. These types of studies are often aimed at fostering the fusion of INSU and IR research, as well as providing insights on why, or why not, such integration is desirable. This session is designed to bring together researchers (both invited and drop-ins) to discuss the methodological issues involved in using the concept of task in information seeking and retrieval studies. Instead of emphasising research results, the aim of the session is to shed light on various conceptualisations of tasks and their applications in research. This is approached by exploring how tasks are used in empirical research and discussing the various pros and cons. The invited speakers will address the major gains that they have encountered and/or problems that they have tangled with in their empirical research, the solutions that they have arrived at, and the challenges that they consider to be critical for future research. The speakers' research cover a diversity of task settings, which are all related to real-life contexts. The speakers' research will be raised to a more general level by commentators with long experience in the research field of information seeking and retrieval. The session also aims to present recent Scandinavian research on the task-based information seeking and retrieval front, and to provide a discussion forum for Scandinavian, American and other researchers involved in task-based research. In the context of work, tasks are the most common form for professionals to perform their duties towards their employers. This fact makes task a fruitful unit of analysis for studies on information activities in different work settings. Byström's empirical research in journalistic and municipal administrative work settings has shown that task performers' perceptions of task complexity influence the need to acquire different types of information as well as the use of various kinds of information channels and sources. This research also demonstrates that although individual perceptions of tasks may vary, the span of information activities applied (especially regarding problem formulations and sought solutions) are in most cases directed/limited by shared professional and organizational conceptual structures. Founded on this research, the speaker will discuss some of the advantages and challenges that become evident when tasks are used as units of analysis in empirical research on information activities. One such advantage is to shift the level of analysis from a more general to a more specific in information seeking studies (from whole jobs to tasks), and simultaneously shifting the level of analysis from more specific to more general in information retrieval studies, which means that these research areas are able to meet. Moreover, the task-based approach allows the study of the interplay in the use of various types of information sources: electronic, human, situations etc. in a process towards a common goal. The concept of task is germane to information seeking in contexts of education and learning. The type of task is generally considered to have strong impact on users' information behavior. However, in her research Limberg has shown that learners' (users') various conceptions of the same learning task influence both their information seeking and use and their learning outcome. The implications of Limberg's findings are that when the same task is experienced in different ways by a group of users, their information seeking and use will take different turns, as regards i.a. relevance criteria, assessment of the authority and value of information sources and their ways of using the information retrieved. This variation will also impact on users' ways of using IR systems. The speaker will discuss a research approach intended to identify and describe patterns of variation in users' ways of conceptualizing task related to information seeking and use. The discussion will compare the purpose of the study of variation with the aim of creating a general model and indicate some consequences of this for the design and outcomes of studies. One specific characteristic of such a research approach is that it acknowledges the user's autonomy of understanding the task and combine s this with the researcher's analysis of information seeking and use related to the same task. As argued by Byström, task complexity affects professionals' information needs and their choice of information sources. Likewise, dynamic work tasks where a stream of unanticipated events occur pose a particular challenge for IR research. Studies of information retrieval behavior can usually devise solutions for anticipated events. In dynamic tasks, individuals and collaborating actors must improvise and retrieve information quickly in new circumstances to maintain work goals. For dynamic work, the task analysis could benefit from expanding the unit of analysis to include features related to the task, such as the work content, the information activities during decisions that have to be made, the social organization and the collaboration of professionals during IR. Such features should become separate units of analysis, and their relationship with IR should be formulated explicitly. The speaker will present this viewpoint and demonstrate this argument with empirical studies from different work settings. The dynamics characterising Web information retrieval can valuably be viewed in light of the interplay between work tasks, search tasks, the searcher, and the searcher's surroundings. Pharo has developed a method schema (the Search Situation Transition–SST–method) focusing on the dialectics of the search process, which changes between interaction with potential real information (Situations) and interaction with meta-information (Transitions). The method advocates the use of triangulation in order to get a picture of the search process as detailed and multifaceted as possible. What can be learned from a work task-oriented viewpoint depends on the level of analysis that is used for handling the data. Different levels of analysis may reveal different kinds of interplay between IR and information seeking (IS) elements, e.g. to look at the influence of work task stage on actions performed during IR-interaction. Based on this work the speaker will address the following central issues 1) what is the appropriate level of analysis; 2) what methods can be used to collect data related to the effects of the searcher's work task during IR-interaction; and 3) what are the identifiable work task dimensions. The presentation will use examples from an empirical study to discuss how these issues can be met. Professionals' information behaviour is often related to task solving. However, professionals do not only perform their duties towards their employer, but also towards their profession. In his research on nurses' information behavior Sundin demonstrates how task solving, and its related information behavior, can be analyzed as something that takes place in different communities at the same time, i.e. a work place community and an occupational community respectively. Each community can be seen as steered by different norms, values and expectations on adequate conduct and behavior. Sundin applies theory of professions in his research on professionals' information behavior and thereby he shows how phenomena such as "relevance" of information and information "needs" are defined within different communities at the workplace, often in competition between, or within, different professions. Thus he makes visible how professional information is not only used in rational task solving, but, perhaps more interestingly, also as a symbolic tool. In his presentation Sundin will address these symbolic aspects of information behavior and methodological considerations that follows from such a perspective. One such consideration is that the concept of task should be seen as interfoliated in a broader socio-cultural context, in which it is essential to empirically study task solving at different analytical levels simultaneously.