INTRODUCTION

Cognition in the wild (Hutchins, 1995) has been elaborated and researched as naturalistic decision making (Klein, Orassanu, Calderwood, & Zsambok, 1993) and situated cognition (Brown, Collins, & Duguid, 1989; Suchman, 1987). In these areas, contextual factors produce problems that are ill-structured, uncertain, dynamic, shifting; have competing goals, contain action/feedback loops; are coincident with time stress, distributed across multiple players, and constrained by organization norms and goals (Zsambok, 1997). Indeed, as the digital global economy of this information age expands and generates extreme complexity, knowledge workers encounter new missions that challenge how they "put knowledge to use" both collaboratively and individually. For example, in Naval peacekeeping operations, we are studying the impacts of collaborative technologies and knowledge integration tools on the effectiveness of multi-cultural team problem solving (McNeese, 2001). This effort provides a backdrop for the study reported here as workers engage in collaborative and individual work that is interdependent, distributed, and often ill-defined. Teams may be composed of different nationalities with conflictual belief systems, salient gender differences, and at-opposition ethnicity. When teams of this nature must work together for the good of the system but have many inherent turbulent socio-cognitive factors, situated transfer of knowledge may be both risky and difficult. In this sense, these type of situations represent what we call risky cognition.

The purpose of this paper is to report initial findings on socio-cognitive factors that impact the transfer of knowledge from one situation to another similar situation. Teamwork often requires team members to acquire new knowledge as they work together to solve situated problems. This knowledge is often put to test as team members are required to transfer knowledge (as individuals) to similar, related tasks (group-to-individual transfer processes) often under severe time pressures. Some of the more salient socio-cognitive factors that affect transfer of knowledge are establishing a common ground of interaction through perceptual anchoring (Cognition & Technology Group at Vanderbilt (CTGV), 1993), shared mental models, metacognition, and the technological support of work.

The Jasper Research Environment

In order to begin study of group-to-individual transfer for situations engaging risky cognition, we draw upon previous research involving group-to-individual transfer in collocated settings that utilizes the Jasper-Repsaj research paradigm (McNeese, 2000).

The selected Jasper problem (Rescue at Boone's Meadow, CTGV, 1997) represents a research paradigm that is predicated on planning, learning, and making decisions involving complex problem-solving tasks that initially involve ill-defined and emergent operations. The Jasper problems are presented in a perceptually oriented, video-based story format. They involve scenarios, event cognition, and embedded data that enable planning and use of specific knowledge to solve the problem presented in the video. The goal of the Jasper series is to help problem solvers / decision makers transform a set of given data or facts into conceptual tools that will enable teams or individuals to solve complex problems (CGTV, 1997). The storyline-contents revolve around a multi-step search and rescue mission that involves flight and ground transportation elements adaptively employed to rescue life in remote places. Contents require problem solvers to work variations of "Distance = Rate x Time" physics problems, given interdependent demands of the problem context as portrayed by the video they are presented. By establishing subgoals and pursuing various plans that tradeoff spatial, temporal, and practical interdependencies, an optimal answered can be derived. However, various qualities of solution can be obtained in addition to the optimal one (as in real world satisficing). The environment may be utilized for ad hoc, first-of-kind decision making either for group or individual problem solvers. Therein it emulates the type of problem solving that incurs in Naval peacekeeping operations. The paradigm can also include a transfer problem (Repsaj; McNeese, 2000). The Repsaj problem is similar in underlying structure to Jasper, but contains a different surface structure and a different mode of representation (text). This creates the kind of extreme, risky cognition (in putting knowledge to use) that is of interest. In addition to a number of performance-based measures such as time, speed, accuracy of the solution etc., the measures include use of problem space analysis and statement analysis as forms of protocol analysis (CGTV, 1997) to understand how subjects think aloud. Finally, recall measures can be utilized to assess how much information research subjects remember from the Jasper and Repsaj problems three days after the problem-solving session. The Jasper-Repsaj research environment simulates characteristics involved in naturalistic decision making and situated problem solving. One may note that teams or individuals can solve Jasper but Repsaj is only solved by individuals. This affords creation of group-to-individual and individual-to-individual conditions that can effectively assess the value of socio-cognitive factors.

Previous Research Using Jasper-Repsaj Tasks

Many team operations ensue with shared face-to-face interactions and are transacted in a collocated setting. Previous research with the Jasper environment focused on this type of setting. The collocated setting study compared performance of 2-person problem solving dyads with individual problem solvers for the source task (Jasper) and a successive analogical transfer task (Repsaj) on a number of different measures (McNeese, 2000). Results indicated one key to establishing a team cognitive process (given disparate levels of team member knowledge) is the perceptual anchor. Hence, perceptual anchors may be significant in the formulation and development of shared mental models that oversee problem solving.

Perceptual anchoring. Problem solving requires both common ground and unique individual knowledge for optimal solutions. Establishment and articulation of a common ground with perceptual anchors is important for quickly integrating disparate beliefs. Perceptual anchors can create a shared experience of the problem (i.e., joint and mutually agreed upon recognition of affordances, effectivities, and constraints that emerge from a problem context) that can become the basis for distilling a shared mental model (McNeese, 2001). When knowledge is acquired (anchored) and differentiated in a perceptual environment, subsequent use of that knowledge for similar, analogous tasks or domains - without being told or informed to do so - can be increased (CGTV, 1993). For our previous research involving group-to-individual transfer, the perceptual anchor was the video representation that presents the Jasper problem as a real world, naturalistic scenario that (1) embeds all the knowledge to solve the problem in the context of the video (2) allows problem solvers to directly experience salient elements of the problem space. Hence Jasper is anchored perceptually in that subjects may perceptually compare, contrast, pickup, differentiate, and understand the problem in a natural meaningful way. Anchors have been suggested as a foundation for communities of learning (CGTV, 1997). The results of our previous research suggest that the perceptual anchor led to benefits for individuals and teams; and facilitated further articulation and use of metacognitive strategies in the collocated condition. Additionally, the results demonstrated that working with a problem grounded in perceptual attributes resulted in specific transfer outcomes on a novel, yet structurally similar task (Repsaj). Thus, if knowledge acquisition can be represented in a way that highlights perceptual contrasts and comparisons, we may be able to predict knowledge team members are likely to use on subsequent tasks.

Shared mental models. A shared mental model may be highly facilitated by perceptually anchored environments like Jasper, and can create bridges of opportunity that allow access of individual unique knowledge as part of the team cognitive process. Shared mental models are defined as representations of information within each team member’s mind, that are jointly held by several (if not all) members of a team (see Endsley & Jones, 2001). Mental models allow members to reason and predict the thoughts and possible actions of their teammates. However, in distributed cognition situations, the sharing that articulates and constructs mental models for future use may be different than face-to-face situations because of constraints inherent within communication technologies. Shared mental models may also develop differently when team members have to overcome cultural, political, and/or gender differences in team composition.

Initial Assessment of Distributed Cognition

When teamwork is distributed across remote locations and mediated by various communication technologies, the nature of problem solving and in turn group-to-individual transfer may change significantly. When teams are required for globally based, international remote operations it is likely that team members will be composed of differing ethnicity-nationalities, gender, and ages. The combinations of these conditions within the perceptually anchored Jasper problem create what we pose as the basis of our preliminary study.

The purpose of the study was to explore how distributed cognition (3-person teams supported by online chatrooms) affects transfer/recall performance; in contrast to individual cognition, and control conditions. Two types of transfer will be discussed: (a) group-to-individual and (b) individual-to-individual. Additionally, this study explored individual differences (age, gender, and ethnicity) on the subjects' ability to solve problems by transferring knowledge from Jasper to Repsaj.

Expectations are that distributed teams utilizing chatrooms, may be different from the results of our face-to-face team research. Despite having the benefits associated with an anchored problem, the formation of shared mental models may be more difficult at a distance. Group-to-individual transfer performance may decay because metacognitive activities and shared mental models fail to form around the perceptual features of the Jasper problem. The individual condition in contrast is expected to maintain focus on the perceptual events of the problem and therein show positive individual-to-individual transfer. If transfer in the experimental conditions is significantly effective, performance should be greater than the control condition. However, if recall is less, the control condition performance may be greater owing to the reduction of memory load (i.e., not having to deal with information overload from the Jasper task). The individual difference conditions are also an exploratory initiative designed to perform an initial assessment on whether they might influence performance on any of the variables (as we expect they would in multicultural team operations).

METHODS

Participants

The participants were 240 undergraduate students enrolled in courses in a large rural university. Thirty participants were randomly assigned to the control condition. Thirty participants were randomly assigned to the individual condition. Ninety participants were randomly assigned to the distributed cognition condition, and then randomly assigned to groups of three. Note that another 90 participants were randomly assigned to an "enhanced condition" (combination of individual and team process), and were also randomly assigned to triads for the cooperative part of solving the Jasper problem. However analysis of this condition is not reported as part of this paper. A random sample of 50 transcripts (25 Jasper and 25 Repsaj transcripts) was selected for an initial preliminary analysis of the data.

Materials/Tasks

The tasks involved in the study included the Jasper-Repsaj problem set. Additionally, an on-line recall task assessed how much information students remembered from the Jasper and Repsaj problems three days after the problem-solving session (McNeese, 2000). Subjects in all conditions had access to the Jasper Problem via a CD-ROM at their individual computer. This provided opportunity to go back into the perceptual-based problem to seek additional facts and contrast scenes as part of their problem solving activities

Procedure

During the problem-solving session, the Jasper acquisition problem was presented. Participants in the control condition did not view or attempt to solve the source acquisition problem. Participants in the individual condition were presented the Jasper video and solved the problem individually. Participants in the distributed cognition condition solved Jasper together in chatrooms after watching the video. Participants in all conditions wrote down their thoughts while attempting to solve the problem. Next, participants in all conditions individually read and solved the (Repsaj) transfer problem. Finally, three days after the problem solving session participants were sent a recall task via electronic mail. They were required to write what they remembered from the acquisition and transfer tasks and electronically submit their responses to the experimenters.

RESULTS

Preliminary results focus on Repsaj transfer, and recall performance although significant Jasper results are provided for individual-cultural differences. A more detailed explanation of these results may be found in McNeese (2001). Only significant results are reported here.

Separate one-way ANOVAs were used on twenty-five participants’ transcripts for the Jasper problem to first investigate whether individual differences (participant age, gender, or ethnicity) resulted in differences on any of the dependent measures presented in Table 1 for the Jasper problem. These ANOVAs indicated that there were no significant differences on the dependent measures due to age or gender differences. However, the data did show that men had a greater number of correct details mention information (CDM) in their transcripts as opposed to women, and that men attempted more solutions to the Jasper subproblems as opposed to women. An ANOVA on ethnicity found significant differences across participants on CDM, mention, attempts to solve Jasper subproblems, and misinterpretations / confusions / recall errors (see Table 1). It should be noted that for gender and ethnicity the number of participants in these groups is unequal and may have influenced the results.

Separate one-way ANOVAs revealed a significant difference between conditions on the amount of information recalled from Repsaj three days after the research session. Participants in the control condition remembered more information compared to participants in the other two conditions. This may be attributable to control condition subjects having only to remember Repsaj information; thereby their memory load is much less in contrast to other conditions. Additionally, participants in the individual condition remembered more information from Repsaj than participants in the group condition. Also, participants in the individual condition solved a significantly greater number of Repsaj subproblems, compared to participants in the other two conditions, the results revealed a trend for participants in the individual condition to have more correct detailed mentions of information from Repsaj than participants in the other two conditions.

Separate one-way ANOVAs performed on twenty-five participant’s transcripts of the Repsaj problem were used to investigate whether participant age, gender, or ethnicity caused differences on any of the dependent measures presented in Table 1. The results of the ANOVAs for age indicated that there were no significant age differences on the dependent measures. ANOVA results for gender showed that male participants attempted to solve a greater number of Repsaj subproblems compared to female participants. Finally, the ANOVA for ethnicity revealed nearly significant differences on two dependent measures; number of attempts to solve Repsaj subproblems and total number of information recalled from Repsaj three days after the session. As previously noted for gender and ethnicity, the number of participants may have influenced the accuracy of the results.

CONCLUSIONS

Examination of results in lieu of expectations indicates there is initial evidence for positive transfer performance in the individual condition. This reifies the position that individuals get the maximum benefit from the perceptual anchoring and likely focused on perceptual differentiation that could then be accessed on the transfer/recall tasks. These results replicate McNeese (2000) for the strong role perceptual anchoring has in individual cognition and subsequent individual-to-individual transfer. However, under conditions of distributed cognition, group-to-individual transfer does not show positive results for this study. Because this was an exploratory condition, we believe it is likely that even though Jasper was first viewed and experienced perceptually it is likely that when subjects went to the chatroom they did not form shared mental models or engage in productive metacognitive activities. One explanation is that distributed cognition teams produce minimal contrast and comparison of different perceptual features versus collocated teams or individuals. Because perceptual differentiation was likely diminished, the distributed cognition teams probably lacked a strong common ground to integrate their metacognitive processes. These results suggest that perceptual anchors were not well attended to as the distributed team may expend more attention dealing with demands of remote communication. Intelligent interfaces or collaborative technologies may be needed to support teamwork wherein cognition can be integrated with the perceptual features that define the context of the problem. Currently we are exploring agent-based interfaces that function as perceptual support systems for distributed cognition.

Our findings with respect to how individual differences / cultural makeup could impact acquisition and transfer of knowledge are extremely preliminary. They represent some of the first studies on how multinational-cultural differences might effect use of knowledge individually and in teams. The results of this study, though constrained, present challenges that are at the heart of our research agenda. Further studies will attempt to answer these challenges and provide a solid understanding on how individuals use unique knowledge to share understanding in a multicultural, multinational team working in a peacekeeping domain.

REFERENCES

Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition and the culture of learning. Educational Researcher, 18, 32-42.

Cognition and Technology Group at Vanderbilt (CGTV) (1993). The Jasper series: Theoretical foundations and data on problem solving and transfer. In L.A., Penner, G. M. Batsche, H. M. Knoff, & D. L. Nelson (Eds.). The challenges in mathematics and science education: Psychology’s response (pp. 113-152). Washington, DC: American Psychological Association.

Cognition and Technology Group at Vanderbilt (CGTV) (1997). The Jasper project: Lessons in curriculum, instruction, assessment, and professional development. Mahwah, NJ: Lawrence Erlbaum Associates.

Endsley, M. R., & Jones, W. M. (2001). A model of inter- and intrateam situation awareness: Implications for design, training, and measurement. In M. D. McNeese, E. Salas, & M. Endsley (Eds.). New trends in cooperative activities: System dynamics in complex environments (pp. 46-67). Santa Monica, CA: Human Factors and Ergonomics Society Press.

Hutchins, E. (1995). Cognition in the wild. Cambridge, MA: MIT Press.

Klein, G. A., Orassanu, J., Calderwood, R., & C. E. Zsambok (1993). Decision making in action: Models and Methods. Norwood, NJ: Ablex.

McNeese, M. D. (2000). Socio-cognitive factors in the acquisition and transfer of knowledge. Cognition, Work, and Technology, 2, 164-177.

McNeese, M. D. (2001). Situated cognition in distributed multi-national, multi-cultural teams: A research plan and a perspective. IST Working Paper. The Pennsylvania State University, School of Information Sciences and Technology, University Park, PA.

Suchman, L. (1987). Plans and situated actions: The problem of human-machine communication. New York: Cambridge University Press.

Zsambok, C. E. (1997). Naturalistic decision making: Where are we now? In C. E. Zsambok & G. A. Klein (Eds.), Naturalistic decision making (pp. 1-15). Mahway, NJ: Erlbaum.