008-0188

The supply chain learning laboratory (scmLAB) as a method to foster interorganizational change

Desirée Knoppen¹ (), María Jesús Sáenz¹² ()

¹Zaragoza LogisticsCenter, Avda. Gómez Laguna 15, 1ª planta, 50009 Zaragoza, Spain, Tel: +34 976077601, Fax: +34 976077601

²University of Zaragoza12, 50009

POMS 19th Annual Conference

La Jolla, California, U.S.A.

May 9 to May 12, 2008

Spain

Abstract

Competition takes increasingly place at the level of supply chains or networks rather than between individual companies. Therefore, many companies are implementingcollaborative processes with selected partners. Practice, however, shows the difficulty of successfully implementing these interorganizational processes. Consequently, action research, such as the learning laboratory method, is gaining ground in operations management research. Learning laboratories aim to foster change through combining conceptual issues with meaningful behavioural dynamics. This paper presents the theoretical framework and method underlying thesupply chain learning laboratory (scmLAB), currently being set up by the authors. The lab method is expected to impacton relevant individual characteristics, such as attitude towards change, technical and relational knowledge, and trust in the trading partner. The pretest and posttest of the quasi-experimental designis grounded uponidentical survey instruments, filled in before and after the laboratory session by each of the participants. Nonequivalent dependent variables are included, to increase internal validity.

Key words: supply chain relationships, learning lab, action research, experimental design

Introduction

Many companies are implementing collaborative processes with selected partners. Nonetheless, practice shows the difficulty of successfully implementing these interorganizational processes (Boddy et al., 2000) and collaboration is therefore still far from being achieved (Holmberg, 2000). The main reasons identified are: low levels of trust between partners (Johnston et al., 2004); an overreliance on technology (McCarthy & Golicic, 2002); and, the lack of understanding of the collaboration structure and dynamics (Busi & Bititci, 2006). Especially, it is the lack of a shared understanding of both the current and desired boundary spanning processes (Vennix, 1999) that seems important. Thus, in a similar vein as organizations developing a memory (rules, procedures, beliefs, culture) that maintains and accumulates experiences over time (Hedberg, 1981; Levitt & March, 1988), the partnership should also develop some kind of memory: exchange rules and norms, interorganizational procedures, shared beliefs, and some shared culture. In more practical terms, this could be called “learning to collaborate”. The newly-created knowledge is “unique to the collaboration and independent of any single organization’s knowledge” (Holmqvist, 1999: 428).

Daily rush and competency traps, however, often impede partners to (jointly) reflect on their shared processes andlearning therefore does not materialize (Levitt and March, 1988; March et al., 1991).Consequently, action research is gaining ground in operations research (Naslund, 2002) and several intervention methods have been developed that aim to foster collaborative learning (Feller et al., 2005; Knoppen et al., 2007; Smeds et al., 2006).

Modelling and simulation constitute powerful tools as part of these intervention methods and are employed in an interorganizational context by two main research streams: group model building (Groesser, 2006; Vennix, 1999) within systems dynamics (Forrester, 1961); and, learning laboratories (i.e. constructed microcosms of real-life settings in which management teams can learn how to learn together, Senge, 1990) with European examples such as the SimLab© of the Helsinki University of Technology in Finland (Smeds et al., 2006), the Manufacturing Strategy Laboratory of the University of Aalborg in Denmark, and the Supply Chain Management Laboratory (scmLAB) that is currently being set up by the Zaragoza Logistics Center and the authors of the present paper (Knoppen and Sáenz, 2007; Knoppen et al., 2007). In general, these learning labs have dual objectives. Practically, they enhance organizational development and competitiveness of the participating companies. Academically, they explore novel research lines, facilitated by the rich data obtained through the lab sessions. During a lab session it is possible, for instance, to observe how supply chain partners interchange knowledge. Consequently, in-depth insight may be obtained regarding collaborative learning processes.

On the other hand, the learning lab method is not only expected to impact on the relationship as a whole, but also on each individual participant (Keys and Wolfe, 1990). In the latter case the research takes on anexperimentaldesign (Cano and Sáenz, 2003). Literature, however, does not provide a detailed theoretical framework nor empirical evidence on this hypothesized impact of the intervention method on relevant individual characteristics. Therefore, the aim of this paper is to develop this theoretical framework and elaborate the experimental method itself. The structure of the paper is as follows. First, some background on the scmLAB is provided. Then, the different individual characteristics are elaborated that are expected to be impacted by the scmLAB method. In that regard, hypothesis are developed. After that, the experimental method is elaborated. Given the preliminary stage of the project, we cannot report the results yet.

Some background on the scmLAB

The scmLAB of this paper shares with other learning labs (e.g. the SimLab©) that it: gathers different organizational members in a dedicated physical space away from the usual day-to-day distractions; explores relevant areas of organizational life such as project management (Cano and Sáenz, 2003), or production (Haho and Smeds, 1996); combines these conceptual discussions with meaningful interpersonal dynamics (Keys and Wolfe, 1990; Senge, 1990; Senge et al., 2007); and, shifts attention from the instructor to the learner, or from teaching to learning. It is thus more appropriate to talk about a facilitator rather than an instructor. The role of the facilitator is therefore crucial during the sessions. Besides ensuring the course of the session, this person should unfreeze group dynamics (Dickens and Watkins, 1999) and play an expert role (Cassell and Johnson, 2006). Critical characteristics of an effective facilitator are: having an enquiring attitude, integrity, process structuring skills, conflict handling skills and communication skills (Vennix, 1999).

The unique feature of the scmLAB, compared to other learning labs, is that it focuses on supply chain issues, involving different agents in the chain. The agents that have decided to participate in this initiative do so for two main reasons: the scmLAB provides tailor-made models of relevant boundary spanning issues; and, the focus on behavioural issues seems novel and promising. At this stage, the first case (each case consists of two partnering companies) is about to be concluded. The model from the second case is currently being built. And three additional cases have shown a firm interest to participate. More cases will be approached in a future.

Some 10-12 boundary spanners from different hierarchical levels of the involved companies will be gathered for each scmLAB session. During a session, conceptual models, computational models, visual models (i.e. the projection of the previous two models on the wall) and social models (i.e. the facilitated discussion between the involved agents) interplay (Smeds et al., 2006). These models will evolve for subsequent sessions, and are all defined a priori. During the planning stage of a session the types of group structure have to be determined (individual, small-group, plenary) as well as the group task (divergent, integrative, ranking and evaluation) (Andersen and Richardson, cited in Groesser, 2006). This will be laid down in the detailed agenda of the session. The agenda concludes with the establishment of a change agenda, including activities, responsibilities and a time horizon.

Theoretical foundation and hypothesis

The scmLAB combines conceptual discussions with meaningful interpersonal dynamics. A tailor made simulation model facilitates in that regard the conceptual discussion. More specifically, the interaction of participants with the simulation model through the possibility to change parameters and evaluate performance changes of the simulated system facilitates experiential learning to take place (Kolb, 1985) and knowledge to be created and interchanged (Nonaka and Takeuchi, 1995). Simultaneously, participants are expected to learn about relational aspects (i.e. attitudes and behaviors of the partner, and the compatibility with own attitudes and behaviors) through the interpersonal dynamics of the laboratory session. Therefore, the first two hypotheses to test are:

H1: The intervention method increases individual knowledge on technical interorganizational issues.

H2: The intervention method increases individual knowledge on relational interorganizational issues.

Studies on team learning have shown that psychological safety (i.e. the expectation that group members will not respond negatively to divergent viewpoints) is determinant for group learning and performance (Edmondson, 1999). More precisely, van Ginkel (2007) has shown the relevance of psychological safety as a mediating variable between shared cognition and group performance. The latter study modified the degree of psychological safety, which the present study however does not pretend. Therefore, we posit psychological safety as a moderating - rather than mediating - variableof the impact of the scmLAB on the degree of individual learning that takes place during the lab session. Therefore, a third hypothesis is:

H3: Psychological safety moderates between the scmLAB method and the degree of relational and technical learning.

The intended tangible result of the intervention method is the establishment of a change agenda agreed upon by both supply chain partners. It is expected that experimentation with possible changes by the participants increases support or reduces resistance towards future changes in the day-to-day setting; i.e. it improves the attitude (i.e. “the disposition to respond favorably or unfavorably towards and object”, Chaiken and Stangor, 1987: p. 578) towards change. Attitudes are complex phenomena and have been measured as one- (affect), two- (affect and cognition), or three-(affect, cognition, and behavior) dimensional constructs (Chaiken and Stangor, 1987). This study uses the three-dimensional version of Dunham et al. (1989) which has been broadly employed in other studies (e.g. Lau and Woodman, 1995). It is interesting to point out that Elizur and Guttman (1976) develop a similar 3-dimensional structure that only differs in the label (not the definition) of the third dimension (instrumental rather than behavioral). Not all attitudinal dimensions will be equally affected however. Given the nature of this specific intervention method (simulation in a lab setting) it is expected that especially the cognitive dimension will be affected. In other words, the affective and behavioral dimensions seem to need more time to change. Furthermore, attitudes towards change in general have to be distinguished from attitudes towards a specific change (Dunham et al., 1989). Thus, further hypothesis are:

H4: The intervention method improves the attitude towards general change

H5: The intervention method improves the attitude towards changes in the interorganizational sphere

H6: The cognitive dimension of attitude is more affected by the intervention than the affective and behavioral dimensions

Finally, the intervention method is expected to foster trust in the trading partner (i.e. “a state involving confident positive expectations about another’s motives with respect to one-self in situations entailing risk”, Dreu et al., 1998) because of the interpersonal dynamics and extensive communication while spending a whole day together (Serva et al., 2005). The benefits of trust for socioeconomic relationships have been emphasized in social capital theory; i.e. trust promotes “access to privileged and difficult-to-price resources that enhance competitiveness but that are difficult to exchange in arm´s-length ties” (Uzzi, 1997: p. 43). Moreover, a higher level of trust promotes cooperative behaviors such as shared planning and flexibility in coordinating activities (Johnston et al., 2004).Thus, a final hypothesis is:

H7: The intervention method increases trust in the trading partner

The previously introduced hypotheses are visualized in Figure 1.

Figure 1: Theoretical Model


Method

The method of the broader study of which the present experiment forms part is of an organizational development (Beer and Walton, 1987) or action research nature (Naslund, 2002; Lewin, 1951). This paper reports on a part of the project, that elaborates experiments on the individual level of analysis. This will be elaborated below.

Experimental design

The experiment itself follows a quasi-experimental design, because participants are selected given their involvement in the topic rather than randomly assigned. More specifically, it involves a one-group pretest-posttest design using a nonequivalent dependent variable (i.e. a variable that is similar to the dependent variable of the research model but that should not respond to the intervention method, Shadish et al., 2002), to increase internal validity. Another possibility to increase validity would be the use of control groups. This is not feasible, however, given that it is difficult to find similar employees in the remainder of the company then those who are participating in the experiment. The pretest and posttest is constituted by similar survey instruments, filled in before and after the laboratory session by each of the participants. The pretest thus provides the statistical baseline, whereas the posttest permits to calculate the degree of change in the dependent variable after having participated in the simulation laboratory.

Operationalization

The operationalization of constructs into requests for an answer plus a response scale is a vital step when developing a survey (Saris and Gallhofer, 2007). The variables of the study are:

  1. Knowledge on technical issues of the supply chain relationship. Given that the respondents are professionals rather than students, it is difficult to ask questions that measure directly their knowledge. This may be resolved by measuring how they perceive their knowledge on supply chain collaboration.
  2. Knowledge on relational issues.
  3. Attitude towards general change (18 items, taken from Dunham et al., 1989);
  4. Attitude towards change in the supply chain sphere (3 items, adapted from Dunham et al., 1989)
  5. Trust in the trading partner. Although several studies emphasize the multi-facetted nature of trust (e.g. Ireland and Webb, 2007), we have chosen a general and therefore shorter operationalization of the construct, in line with Dreu et al. (1998).
  6. Psychological safety (6 items, from Edmondson,1999)

An additional nonequivalent dependent variable was introduced for the first and fifth variable.For the first trust variable, this implied asking a same question, but in the context of another strategic buyer/supplier. For the fifth technical knowledge variable, this implied asking about the degree of knowledge on another interorganizational topic, in this case joint product/service development. In that regard it was important that the variable referred Appendix 1 shows the indicators that were developed by the authors of this paper, (related to variables 1 and 2), those that were adapted from other studies (related to variables 4 and 5) and the nonequivalent dependent variable.

Data analysis

The results of the experiment will be analyzed employing multiple-group nonparametric statistics. A potential limitation is that the participants of the different groups (each group of participants stems from another buyer-supplier case) are too different statistically to be added up. For example, organizational values may be distinct in each case making employees more or less receptive for experimentation. Nonetheless, all companies will be located in the Center-East region of Spain which increases their comparability. Moreover, participants from the different cases will have very similar functional activities (related to supply chain management), further increasing the comparability of different cases.

Finally, the data from the survey may be triangulated with data obtained during the session, as 2-3 distant observers will be coding the behavior of the participants during the session, following a previously established and tested coding protocol.

Conclusions

This paper draws the theoretical model and experimental design underlying the scmLAB, that is currently being set-up by the authors. As soon as empirical data become available - several cases are being prepared at this stage - we will be able to answer the developed hypothesis of this paper.

This research forms part of a broader project titled “Proyecto Singular Estratégico”, supported by the Spanish Ministry of Education and Science (Reference: PSE-370500-2007-1).

References

Beer, M. and A.E. Walton (1987), "Organizational change and development," Annual review of psychology, 38, 339-67.

Boddy, D., D. MacBeth, and B. Wagner (2000), "Implementing collaboration between organizations: an empirical study of supply chain partnering," Journal of Management Studies, 37 (7), 1003-17.

Busi, M. and U. Bititci (2006), "Collaborative performance management: present gaps and future research," International Journal of Productivity and Performance Management, 55 (1), 7-25.

Cano, J.L. and M.J. Sáenz (2003), "Project management simulation laboratory: experimental learning and knowledge acquisition," Production planning and control, 14 (2), 166-73.

Cassell, C. and P. Johnson (2006), "Action Research: Explaining the diversity " Human Relations, 59 (6), 783-814.

Chaiken, S. and C. Stangor (1987), "Attitudes and attitude change," Annual review of psychology, 38, 575-630.

Dickens, L. and K. Watkins (1999), "Action Research: Rethinking Lewin " Management Learning 30 (2), 127-40.

Dreu, de and Giebels (1998), "Disrupted trust in negotiation," Journal of applied psychology, 83 (3), 408-22.

Dunham, R.B., J.A. Grube, D.G. Gardner, L.L. Cummings, and J.L. Pierce (1989), "The development of an attitude toward change instrument," in Academy of Management. Washington.

Edmondson, A. (1999), "Psychological safety and learning behavior in work teams.," Administrative science quarterly, 44, 350-83.

Elizur, D. and L. Guttman (1976), "The structure of attitudes toward work and technological change within an organization," Administrative science quarterly, 21, 611-22.

Feller, J., A. Hirvensali, and R. Smeds (2005), "Inter-partner process learning in collaborative R&D-a case study from the telecommunications industry " Production planning and control, 16 (4), 388-95.

Forrester, J.W. (1961), Industrial Dynamics. Cambridge, MA: MIT Press.

Groesser, S.N. (2006), "Learning by doing: Towards the development of a group model building framework," in Conference of the systems dynamics society. Nijmegen, the Netherlands.

Haho, P. and R. Smeds (1996), "Benefits from using simulation games in business process development. Experiences from a Finnish manufacturing company.," in 6th IFIP TC5/WG5.7 International conference APMS, O.H. Tamura and S. Fujii (Eds.). Kyoto, Japan.

Hedberg, B. (1981), "How organizations learn and unlearn.," in Handbook of Organizational Design: Adapting Organizations to their Environments, P. C. Nystrom and W. Starbuck, Eds. Vol. Volume 1. Oxford: OxfordUniversity Press.