This study is not without some limitations. First, organizational change is viewed here as very gradual and the data collected did not show any radical changes. On one hand it could be the time period spent at TechProject and six months was maybe not enough, in this case, to see such changes. On the other hand, it might not be related to the length of the time period but to the time period itself. Social representations are best studied when „new‟ issues emerge and the paths of different groups cross (Bauer & Gaskell, 1999). While that was the case at TechProject, a major organizational restructuring also occurred toward the end of the data collection phase. Thus, staying more than six months at TechProject would have introduced a significant level of noise in the findings.
Second, cultural values and symbols affect the evolution of a social representation (Markova, 2000) because social representations create a link between the past and the future. As a member of a particular group, one needs to conform to the social exchange typical to that group and to the groups‟ values, beliefs, symbols, and norms (“the past”). Thus, an ethnography-based research would have provided more insights into the evolution process.
Third, it is important to note that there are always “power relations that underlie the construction of knowledge” (Gal & Berente, 2008, p.149). Any group has power relations and it is impossible to deny their existence especially in the context of a governmental agency. A political interpretation (as well as other theoretical interpretations of the data) could have been possible, but I chose an interpretation grounded on the data, based on the groups‟ social representations.
Finally this study, like most of the studies on knowledge sharing across boundaries, assumes that knowledge integration is beneficial and that the advantages outweigh the costs. Nevertheless, when differences in knowledge at the boundaries do not converge, this assumption does not hold (Howard-Grenville & Carlile, 2006). To control for this potential risk, the teams examined had to already use a knowledge-based system for sharing knowledge within or across projects. However, no attempt was made to objectively assess knowledge convergence at the boundaries. In other words, this research assumed that, if an organization invested time and physical resources into a knowledge-based system, knowledge integration was beneficial.
8.6.Future Research
In addition to simply addressing the research limitations mentioned above, this first study on social representations of knowledge-sharing practices opens the door to several other research opportunities.
The results raised intriguing questions about the influence of social representations on organizational change. Thus, an important research opportunity, albeit rather challenging, is to focus on this relationship and on the motors generating the change (Van de Ven & Poole, 1995). Opposing forces in each EI group created social representations of the knowledge- based systems that were more or less similar across the groups. This variation engendered the adoption of four main patterns characterizing knowledge sharing within and across the groups. Based on these findings, the change process resembled Van de Ven and Poole‟s evolutionary model where some of the patterns are selected and others are eliminated through „natural selection‟. Through anchoring and objectification, the selected patterns shape the
form of the new social representation. Do social representations really influence organizational change following an evolutionary model? If so, how does the „natural selection‟ occur? Do routines and improvisations really mediate the relationship between work practices and social representations? What are the conditions influencing the balance toward the forces promoting change or toward the forces promoting stability?
Furthermore, what is the relationship between the evolutionary change process and the opposing forces of the social representations? Do they support or constrain each other? A similar research avenue could also aim at refining Van de Ven and Poole‟s framework by examining how the two motors (evolutionary and dialectic) work together: as intertwined motors, in parallel, or consecutively.
Another research opportunity is to consider organizational culture, as an important dimension of the social context, exerting a powerful influence on the individuals‟ knowledge needs and their subsequent knowledge-sharing practices (Alavi et al., 2005). Previous research has already pointed out how an organization‟s cultural values influence its knowledge sharing practices (Alavi et al., 2005; Knapp & Yu, 1999; Davenport et al., 1998) as well as how they impact the way the knowledge-based systems are used (Alavi et al., 2005). Based on the TechProject case-study, I found that knowledge sharing across groups was ineffective and that the groups‟ social representations varied significantly. What is the role of the organizational culture when an organization aims to achieve a common social representation? Is the cultural transition the essential predecessor of a social representation harmonized across the organization? Is this cultural transition the key element to achieve an effective and efficient knowledge sharing process?
Finally, another research opportunity is to use the social representation theory to validate Nonaka and Takeuchi‟s (1995) concepts of socialization and externalization. Gourlay (2006) has recently remarked that Nonaka and Takeuchi‟s groundbreaking model does not have enough evidence supporting the socialization and the externalization of technical tacit knowledge. With respect to the former, Nonaka and Takeuchi seem to suggest that “when someone learns under the guidance of an expert that some indescribable kind of knowledge is „transferred‟ by an unknown [emphasis added] process” (Gourlay, 2006, p.1420). As for externalization, Nonaka and Takeuchi based their reasoning on figures of speech suggesting “it is a linguistic process through which, mysteriously [emphasis added], tacit knowledge becomes explicit” (Gourlay, 2006, p.1420). While I agree in general with Gourlay‟s remarks I do not believe the two processes are unknown and mysterious. On the contrary, I think social representation theory is a very appropriate theory for demystifying this black box.
In conclusion, my dissertation examined IT-based knowledge sharing practices and gave an explanation of the social representations formed by each project team and how they affected the knowledge sharing practices within and across groups. Based on my findings, I also offered suggestions for management practices and future research opportunities by recognizing the critical role of social representations in introducing change and accepting the unfamiliar event. It is my hope that this research provided a different way to consider the socio-technical influences in IT change management, in both theoretical and practical terms.