Shifting the mental model and emerging innovative behavior:
Action research of a quality management system

Stephen D. Tsai
National Sun Yat-Sen University, TWN

Chung-Yu Pan
Tunghai University, TWN

Hong-Quei Chiang
National Sun Yat-Sen University, TWN


The aim of this paper is to explore an effective approach to consider the innovative behavior in incumbent business, both practically and theoretically. We initiated a two-stage action research project to participate in the whole process of planning, implementation, evaluation and reflection in a quality upgrade program of the researched corporation. We find that the quality performance cannot be improved by reinforcing the established quality system. Thereafter, applying the principles of complexity theory, we nurture a context to shift the prevailing mental model and bring in innovative behavior within the organization. From an intentional perspective, we apply the single and double-loop learning models to explain the distinctive implications of the two stage action research process. In addition, from the emergent perspective, we propose a recursive process model of interventions on complex adaptive system to further illuminate how these interventions bring in innovative behavior. We conclude that an organization, which can create a balance between tradition and the new (emerging) complexity paradigm, will exhibit both adaptive and innovative capacity.


To those organizations which are confronted by an ever-changing environment, the introduction of innovative capacity is very important in order to upgrade their existing quality systems. Traditionally, most Total Quality Management (TQM) implementations are based on philosophies laid down by professionals such as Deming, Juran, Ishikawa, Crosby, Feigenbaum, and Harrington (Borges & Guastello, 1998). They share some common principles such as emphasis in training, in control, and in increasing respect for the people involved. Examining the literature on TQM we can find that surveys have shown that 60-70% of TQM programs fail to achieve their stated objectives (Ovretveit, 2002). Sometimes TQM is associated with other kinds of programs, therefore, it becomes more complex to ascertain the enhancement to the organization’s productivity, quality, and profitability resulting from those implementations.

Recently, as well as in the traditional philosophies, there are some literatures on quality improvement systems which focus on chaos and complexity theory to generate innovation (Dooley, et al., 1995; Borges & Guastello, 1998; Plsek, 2000). It is not surprising that the tools of a traditional paradigm will continue to be effective in improving the quality of work-level processes. Nevertheless, we believe that TQM theory and practice can benefit from a further coupling with theory from the complexity paradigm.

The aim of this paper is to explore an effective approach to introduce innovative behavior in incumbent business, both practically and theoretically. We illuminate an action research project which demonstrates how to balance in practice both traditional and complexity paradigms. The authors experienced and participated in the whole process of planning, implementation, evaluation and reflection in a quality-up-grade project. Concurrently, we carried out a two- stage action research project, which provided some insights for shifting the mental model and bringing in innovative behavior.

Literature review

Research conducted on the profit impact of market strategy has led to the conclusion that “in the long run, the most important single factor affecting a business unit’s performance is the quality of its products and services, relative to those of competitors” (Buzzell & Gale, 1987). However, a search for the definition of quality has yielded inconsistent results. Quality has been variously defined as value (Abbott, 1955), conformance to specifications (Gilmore, 1974), conformance to requirements (Crosby, 1979),fitness for use (Juran & Gryna, 1988), loss avoidance, and meeting and/or exceeding customers’ expectations (Parasuraman, et al., 1985). Regardless of the time period or context in which quality is examined, the concept has had multiple and often muddled definitions and has been used to describe a wide variety of phenomena.

Recently, Prahalad & Krishnan (1999) indicate that the concept of quality has evolved from product and service to innovation. Evolving with the subjective preferences of consumers and the strategies of competitors, the meaning of quality is not only diversified, but will also add to or change the implied meaning by its context. In other words, continuous experimentation and the development of new innovative features play an important role in shaping consumers’ ever-shifting perceptions of quality.

A review of the literature discussing the relationship between TQM and innovation suggests that there are conflicting arguments concerning the relationship between TQM and innovation (Prajogo & Sohal, 2001). Arguments that support a positive relationship between TQM and innovation contend that companies embracing TQM in their system and culture will provide a fertile environment for innovation because TQM embodies principles that are congruent with innovation (Dean & Evans, 1994; Roffe, 1999). In contrast to the above position, several scholars reject the positive relationship between TQM and innovation for the reason that it possesses principles and practices that could hinder innovation (Slater & Narver, 1998; Wind & Mahajan, 1997).

The contest of these opposing arguments can be extended to address a lack of appropriate propositions for organizations to excel in both types of performance. The complexity and multiple perspectives historically associated with the concept of quality have made theoretical and practical research advances difficult. Consequently, the challenge is not to formulate one definition or model that attempts to account for all possible variables, but rather, the challenge is to develop models and definitions that are comparable, even cumulative, and which account for the components neglected up to now. From these considerations, this case study of action research not only provides a significant evidence for building innovative capacity in existing TQM organizations, but also enables the progress of theoretical development of quality management.


Action Research

The concept ‘action research’ was proposed by the eminent social scientist Kurt Lewin more than 50 years ago (Lewin, 1946). In his original contribution, however, Lewin did not explicitly define the term itself. Rapoport, (1970) has framed probably the most commonly used definition of such research, when he says that, “Action research aims to contribute both to the practical concerns of people in an immediate problematic situation and to the goals of social science by joint collaboration within a mutually acceptable ethical framework.”

In an organizational context, managers and consultants are increasingly undertaking action research projects in their organizations (Coghlan, 2001; Gummesson, 2000). There is little doubt that we are living in turbulent times characterized by rapid change and challenges to ways of thinking and acting (Stacey, 1996; Weick, 1995). Against this backdrop, undertaking meaningful organizational and managerial research has become more problematic. The accelerated pace of deep and pervasive change calls for research methods that are dynamic and flexible, capturing the rich complexity and ever shifting ground apparent in the organizations in which we live and work. In this context, the potential of action inquiry strategies for inclusive organizational research is being rediscovered (Greenwood & Levin, 1998).

The value of action research lies in its ability to get really close to a business reality where a researcher can go beyond just gleaning insights into broad phenomena and can get to the specific aspects of problem solving, which are often the bases of business decision-making (Lindgren, et al., 2004). Action research, with its process of longitudinal research in context, including cycles of observation, interpretation, action and reflection, allows for understanding, constructing and testing of explanations, as well as modifications and learning. Thus, there is little doubt that action research has the potential of producing important and useful knowledge of people in context.

Data collection and interpretation

The team for the action research project reported herein involved first and second person practice and third person research. One member of the research team is a senior manager of the researched organization who, in charge of the implementation of the project, can be regarded as a first practice person in the action research. The second member was the consultant of this project who, regarded as the second practice person, was obligated to carry out the effective changes of the quality system. In addition, the third member is an expert of qualitative research who is dedicated to the progress of scientific knowledge in this action research, and hence, can be regarded as the third research person in this project.

Rich and multiple data were patiently collected and interpreted throughout the whole process of this action research. Inclusive of the participation and observation, both qualitative and quantitative data were collected. All the actions, observations and reflections were written in a journal. We interviewed 30 selected participants for 1-2 hours each to discuss the practical possible solutions and emerging theoretical issues. Every agreement in the meeting was recorded and reviewed weekly. The research team worked together in decoding data, shaping constructs, and developing theoretical insights grounded in the data and interpretation. Our insights are derived from an iterative process that involved interviewing and observing organizational members, reviewing the literature related to learning and innovative behavior, and interpreting the working journal and transcripts.

To effectively evaluate the consequences of intervention and to measure the quality performance, we also collected the quantitative data of both product and process parameters. In addition, the monthly market share was regarded to be the index of the competitive advantage of new product development.

Thick description of action research

For the past 50 years, TTL has always held the monopoly in the domestic cigarette and liquor industry ever since its establishment. However, since the establishment of WTO in 1987, TTL has brought in new quality control methods on the production system such as a quality control circle, a proposal appraisal system, and, TQM and ISO certification to compete with global suppliers in the domestic market. In the year 2000, liquor brewing and distribution rights have been further opened to domestic private companies which have made their market competition more serious. For developing new products in the high-priced market in order to create new competitive advantage, TTL decided to launch a quality-upgrading project in September 2003 through external consultancy.

We, on the one hand, playing the role as the research team, have facilitated TTL to develop its ability of organizational learning and shift its organizational mental model, which would further nurture the context of enabling emergent innovative behavior. Consequently, the shifting mental model with its evolving value and context is turning TTL into a vital and more competitive organization. On the other hand, both our ability of initiating appropriate intervention and creating specific knowledge has been enhanced by this action project. In particular, insights developed by this project will provide a better understanding of practical project implementation in similar contexts. In the following section, we first illuminate the practical problem solving and reflecting process in each stage of the action research. Typical procedures in each spiral of the action research are presented with pre-understanding, action planning, implementation, evaluation, and reflection.

First spiral of action research

Pre-understanding refers to such things as people’s knowledge, insight and experience before they engage in a research program. The knowledge, insight and experience of the manager-researchers applies not only to theoretical understanding of organizational dynamics, but also to the lived experience of their own organizations (Coghlan, 2001). We, the authors, play multiple roles, as action manager, consultant or researcher. Our pre-understanding affected, and was itself affected by, the project developing process.

At the initial stage, Statistical Quality Control (SQC) was believed to be an objective professional expertise, which could be transferred to the internal personnel by consulting efforts. We definitely assumed that the internal personnel could apply SQC to related routine processes and make a radical quality improvement.

Since September 2003, we took up the program with the active staff in the headquarters and three factories. At first, we collected and analyzed the documents and data of the existing quality system to survey the potential quality problems, and also tried to work out a solution. Then intensive classes were arranged for senior and middle managers to implement the primary recommendations of SQC skills. After the classes were conducted, a meeting with related senior managers was held to discuss and review the existing quality management system. After discussions held at the company headquarters, it was decided to amend the existing top-down management system as following:

However, under the traditional hierarchical legacy, neither the internal personnel nor consumer experienced any substantive improvement derived from the updated quality system. The primary derivative was simply more red tape. Even though the standard of quality management became more rigorous, it still operated under the existing mental model. One of the quality control operators reflected on both sides of the changed and unchanged systems as:

“Recently, there is so much paper work to do everyday. Whenever I deal with a quality problem, I waste a lot of time to record data that are required by senior supervisors, rather than try to generate solid improvement. The point is that I should follow the new rules and procedures that are documented. If I don’t, I will definitely get into a lot of trouble.”

After the implementation of the enhanced quality program, such unchanged self-defensive behavior was a typical reaction.

In March 2004, at the Spring Convention of TTL, it was concluded that the quality of the final products was still the same as those prior to the first stage of managerial action. In other words, the updated quality system did not demonstrate any substantial contribution of quality improvement. The chief executive officer pointed out that the reason for inefficiency in the management was due to the mental model of internal personnel which did not reconstruct to adapt to the new competitive environment. He claimed:

“In order to come up with the new challenge in the market, we need to abandon the habitual bureaucracy of the past monopoly era and head to a performance-oriented enterprise. The most critical thing is to change the mental model of the organization. This significant change needs to be done as soon as possible!”

We, therefore, were required to share the responsibility with the organizational members to find a more effective intervention to create substantive quality improvement. Hence, the effect of the action was reflected upon and reviewed over the first stage of action research. We frequently worked with the organization members to assess whether managerial action had been effective or not. If not, then, where was the problem? And, how could we modify our managerial action to work out the possible effectiveness? These detailed reflections are discussed in the next section.

Reflections of the first stage action

Although the first stage managerial action was well diagnosed, analyzed, and planned, it failed because of the traditional bureaucracy of the organization which caused the internal members to possess more self-defensive mechanisms than innovative ones. That is the key reason why the top-down managerial action could not realize the expected results. Three context features have been listed to explain why the first stage managerial action did not turn out as expected: 1. nonlinear communication, 2. negative feedback, and 3. increasing tension between organization and environment. The further explanations and interpretations of these reflections are given below.

Nonlinear communication: The communication among organizational members is nonlinear and uncontrollable, and it might even be twisted. The message was interpreted differently due to the unique mental model of each individual. For example, the foreman wished to follow the managerial suggestion, and requested more workforce from his supervisor, but the supervisor expressed his grievance:

“There are just enough people to carry out my daily routine; I don’t have any spare manpower that could be transferred to do this experiment. It is not possible to help you out today, maybe some other day. Please don’t push me so much!”

Even though the supervisor believes this experiment might improve the efficiency, he still cannot assist his subordinate because of heavy daily routine and bureaucratic legacy. It shows that different individual priorities play a potential barrier to the linear top down control system.

Negative feedback loop: To handle the increasing diversity and complexity in the updated quality system, there is a saying among most organizational members, “Where there is a decision at the top, there is a counterplot at the bottom.” It means that the top-down control scheme will naturally employ line staff to build up their own formalized solution to tackle all the problems at hand. In such a manner, most organizational members believe that they can fulfill their jobs effectively and efficiently. Nevertheless, this negative feedback of formalization reinforces a conservative self-defense mechanism rather than the positive innovation. The quality control manager who dealt with consumer complaints reflected on this negative feedback feature:

“When a consumer raises a product complaint, it seems to me that to make a proper apology and compensation is more important than to improve the product quality. Besides, I need to carefully record this problem instead of paying more effort to improving an uncertain process problem. I will not leave any verbal flaw. Otherwise, it will leave a cue for headquarters to investigate.”

It shows that the senior officers usually do not cooperate to solve quality problems, but only contribute further questions and stress. Such a negative feedback loop in the real context proves that it is difficult to bringing about substantive quality progress.

Increasing tension between organization and environment: The turbulent elements of the external environment, such as new equipment and material, emerging product and marketing strategies of the competitors, and the evolution of consumer preferences, have made quality issues more complex and diverse than ever. Whilst more quality problems are raised, the internal personnel and their mental models remain the same. As a result, the documentation-orientated problem-solving pattern leaves the real quality problem unresolved. The existing quality system, with negative feedback, is no longer effective and consequently incurs increasing tension between the organization and environment. In other words, the well-modified and planned action scheme could not effectively develop better internal capacity to upgrade the quality system.

Second spiral of action research

In summarizing the first stage action and reflection, we knew that the mental model should be changed prior to substantial upgrading of the quality system. Based on our research experience and pre-understanding in complexity (Tsai, et al., 2003; McKelvey, 1999), we suppose that replacing top-down control with bottom-up inspiration would create a better quality system to generate better products and new competitive advantages.

Acknowledging that the quality management system did not need more revision, we sought to develop a different approach to inspire bottom-up motivation. Therefore, the second stage managerial action was launched, which excluded the headquarters and focused on a unique factory, which was assigned to develop a high priced high quality product. From April 2004, the second stage managerial action took place, which included four main actions: 1. shaping the vision 2. information exchange 3. participant interaction 4. process conditioning.

Shaping the vision: By participating in this project of developing high-priced products, the concept of ‘Imperial Kitchen’ was proposed to promote the belief ‘High Quality brings High price’. Although there was lack of know-how, we put forward that both ‘carefulness’ and ‘conscientiousness’ could make a product more valuable. In late April 2004, a product exhibition was held based on the idea of a “Hi-Tech District” which not only impressed the visitors, but also inspired the organizational staff to look forward to the future. Accompanying ongoing important events, we are continuously shaping the new mission, values, and goals. Rather than doing this through a one-way, top-down approach, we used an iterative process to develop a new vision, which involved all levels of the organization to encourage alignment.

Information exchange: We believed that sufficient information exchange would facilitate the empowerment of organizational members. Therefore, all departments related to the quality system, such as: R&D, Marketing, Material, Engineering and Production Departments, were invited to exchange their ideas and professional knowledge. In addition, the external suppliers were invited to give presentations to build-up a new interactive web between internal personnel and external suppliers. Such an effective exchange of diverse ideas boosted understanding of knowledge, changed the original operating pattern, and shifted the mental model.

For instance, the control chart had been used to monitor and evaluate the performance of process quality for more than 20 years. It was time consuming and could not provide timely and current analysis, and had made itself a habit rather than a useful quality tool. This issue was highlighted while the statistical quality control knowledge was reintroduced; internal personnel recognized the drawback of the existing quality control system and developed an offline computer aided system. Concurrently, external equipment suppliers held a professional seminar. Based on the interactive understanding from both statistical quality expertise and the features of equipment, the first line staff surprisingly found that an excellent online computer aided system had existed for five years, which was the most effective control tool of process quality. They said:

“Through the intensive information exchange from quality consultants, external suppliers, instruction manual and participative discussion, we were enabled to revisit the online function of the Gauss graphic display on the existing equipment. The coveted solution was right in front of us, but without enough sharing of expertise we could not see it before!”

Actually, this redefinition process of an existing facility is very context dependent. With sufficient information exchange, valuable knowledge could be created unexpectedly.

Participation and interaction: The quality variation on the production line was directly observed and discussed with first-line operators to work out a solution. Similar to the previous stage managerial action, the first reaction from the operators was that these improving programs would increase their workload and most of them revealed defensive and questioning attitudes. To handle such situations, we were aware that building trust and coordination would be significant in the initial phase to achieve an effective improvement. No more command and control, but through intensive participation and interaction, a mutual trust was built-up among operators and supervisors. An operator mentioned how he gradually became more open-minded:

“Although it is a good idea to have all material weighed precisely before loading into production, I couldn’t promise to fulfill it at first. In effect, it would increase the workload of my co-workers and offend them. Therefore, when senior managers asked us whether it would be possible, I usually disagreed. But this time the consultant and managers themselves participated to understand our difficulties and tried to assist us on improving operations. Their warm concern has been very touching and the result of the new operation looks good. It would be embarrassing if I didn’t collaborate on this improvement program.

This breakthrough not only rationalized workload, but also tackled staffs’ problems of recognition, attitude and teamwork. Over time, minor improvements accumulated and interacted, and gradually organizational members changed their attitude from passive reaction into actively improving their process quality. Integration with the synergy of external expertise and technologies has given rise to some innovative behaviors that were beyond our prior expectation. For example, the chief manager pointed out in his final report:

“For the loading operation, it took ten years to change from the first generation of monthly checking to the second generation of daily checking. Subsequently, it took about three years to upgrade to the third generation of batch-by-batch checking. However, from the batch-by-batch operation to the fourth generation of computer aided weighing system only took three months. Furthermore, we believe that it will promote to the fifth generation of network operations very soon, because we now have good teamwork characterized by excellent interaction and communication.”

Process conditioning: We reviewed the second stage managerial action by intensive weekly meetings. These meetings did not emphasize the objectives of action, but focused on the sequence of order, condition of progress, and shape-up of teamwork. Through open discussion and dialog, some workable ideas and methods were proposed and exploratory trials or experiments commenced. The feature of this progress conditioning was not to list all the sequences and alternatives before action, but to expand and standardize them until certain methods emerged with preliminary success. Furthermore, continuous problem-finding and staff interaction were required to trace out the variation of each problem. One first-line operator explains such a patching progress in improving the accuracy of the loading operation:

“They commence with small-scale pilot schemes or experiments, which make it easier to find workable solutions. In addition, they participate in the whole improvement process that facilitates timely brainstorming and understanding of each other’s role and dilemma. Finally, we have discovered the computer-aided system which simplifies complicated operations and improves the accuracy of the loading operation. Experiencing a patching and winding path, we together overcame the impossible mission.”

Shifting the mental model and supporting the emergence of innovative behavior interplayed throughout the second stage of managerial action. With minor innovative process improvements like the one mentioned above, the performance of process and the quality of product was improved. Just like small fires which, when fanned by the winds of active interaction, can become great conflagrations and effect, great waves of change. Consequently, from April to December 2004, a new classic product was successfully introduced into the top tier market and obtained new competitive advantage. It shows that TTL has benefited from quality renewal and quality innovation.

Reflections of the second stage action

Since we became partners in the whole research process, from the inside, we have had the privilege of seeing how new local, contextual, and immediately applied objective expertise can lead to the shifting of mental models and the emergence of innovative behavior. Three contextualized attributes, which were created by the second stage of managerial action, are summarized as:

  1. Cross-level communication, which overcame the nonlinear communication problem in the first stage managerial action. Although the internal personnel still possess their diverse mental models, through cross level communication it is easier to develop the mutual trust which exists in an overlapping mental model. Moreover, the first line staff are inspired to improve teamwork to facilitate their knowledge creation for solving uncertain quality problems.
  2. Positive feedback of fine improvement; which means that we make many small changes rather than major disruptions to improve system performance. These small changes, which explore the innovative direction of quality system, eventually move toward a tipping (bifurcation) point. Beyond this unanticipated point, some emergent innovative behaviors have appeared within the factory, and the quality performance has simultaneously been upgrading.
  3. Open and recursive adjustment, which supports small-scale experiments and/or impromptu changes; exposes the staff to new understanding and unexpected ideas (which they often themselves initiate), which shift the mental models of both individuals and the collective. Such adjustment is not only moving towards a predetermined target, but also creatively solves immediate or upcoming problems.

In short, the established bureaucracy of TTL and its existing quality system still provide the stable fundamentals of a business operation. The effects of the second stage managerial action mainly extended the context formed by the first stage managerial action to create ongoing innovative behaviors. These practical innovative behaviors demonstrate the possibility of developing quality innovation based on the attributes of the existing quality system - we don’t have to “throw the baby out with the bath water”. These consequences not only have solved the practical problems of quality upgrading, but have also theoretically bridged the gap between quality management and quality innovation. The next section explains how the practical and transferable knowledge was created through this project.

Conceptualization of action research

A summary of the research project, presented in Table 1, compares two stages with their different managerial actions and their effectiveness. As well as resolving the practical problem of upgrading the quality system, we explored the theoretical implications of how shifting mental models can encourage innovative behaviors. To illuminate the knowledge created in this action research, we explain these implications from the intentional and emergent perspectives respectively, in the following sections.

A definition of ‘mental model’ and ‘innovative behavior’

A mental model is nothing more than the beliefs about an issue as a whole. In other words, they are rules, regulations, habits, managerial frames, assumptions, mind-sets, paradigms, conventional wisdom, customs, institutional memory, and so on. Research has shown that every human being has a mental model, which develops over time primarily through education and experience (Constantinos, 1997). Similarly, organizations develop mental models, manifested in their culture, routines, and unwritten rules of behavior which will also develop over time through education and experience. Mental models can be good because they allow us to process information and make decisions quickly. However, very strong mental models can hinder active thinking and the adoption of new ideas because they act as filters that screen incoming information, i.e., mental models simultaneously both enable and disable our thinking processes.

 Spiral 1 (Sep 2003-Apr 2004)Spiral 2 (May 2004-Nov 2004)
Pre-understandingQuality expertise can be transferred and implemented by well educated staffQuality expertise can be shared and realized in adequate context
Action planningA serial of educational programs and review discussionsTo create a learning organization
Action taking1. Enhance autonomous quality inspection of first-line operator.
2. More control item and higher standard.
3. Promote and implement quality audit system.
1. Shaping the vision
2. Information exchange
3. Participating interaction
4. Process conditioning
EvaluationHardly any substantial contribution of quality improvement.Successfully launch a new classic product in top tier market.
ReflectionAdhere to existing mental model:
1. Non-linear communication
2. Negative feedback loop
3. Increasing tension between organization and environment
Shifting mental model and emergent innovative behavior:
1. Cross level communication
2. Intensive weaving of fine improvement
3. Open and recursive adaptation
Learning styleSingle loop learningDouble loop learning
Managerial PrincipleTraditional or mechanical paradigmBalancing between traditional and complexity paradigm

Table 1 Summary of action research project

Innovative behavior can be defined as the intentional generation, promotion and realization of new ideas within a work role, work group or organization, in order to benefit role performance at the level of the group or the organization (e.g., Kanter, 1988; Scott & Bruce, 1994). In the second stage action research, our efforts focused on lower level employees who were primarily able to contribute to smaller scale innovations in the domain of their own work roles and work groups. A typical process of encouraging innovative behavior at the second stage of action research is shown in Figure 1. We experienced the evolution of questions and mental models within the context constructed by action and interaction. For example, in the process of developing a precise weighing operation, the first line staff presented their concerns iteratively through ‘how’ and ‘why’ questions. On the one hand, by the ‘how’ question they learned to find the solution for their problematic situation. On the other hand, by the ‘why’ question they tried to make sense of their actions and interactions for encouraging themselves to search for alternative solutions if necessary.

Figure 1 The process of bringing out innovative behavior

At the first stage of action research, due to the mental model deeply seated in TTL’s comfortable history, most organizational members would not challenge any additional problem to build a more effective quality system. Therefore, we focused on shifting the organizational mental model in the second stage of action research. We explain the relation between mental model and innovative behavior by two complementary perspectives in the next two sections.

Intentional perspective

From the intentional perspective, we did have to realize the upgraded improvement in this quality project. Argyris & Schön (1978) proposed that single- and double-loop learning models could address the effectiveness of intentional intervention by fostering organizational change systemically. Our efforts in the first stage of action research basically followed the principle of Plan-Do-Study-Act (PDSA) which is the dominant logic of TQM as shown in Figure 2. The ‘Plan’ stage does in fact represent a planned change, and the ‘Study’ and ‘Act’ steps form feedback loops under which system performance is assessed and new methods are standardized into the existing, equilibrium system. Such a single-loop improvement cycle, in spite of its predictability and controllability, has been ineffective in improving organizational skilled incompetence. Researchers have proposed that single-loop learning merely solves the presenting problems (Dooley, et al., 1995; Argyris & Schön, 1978). It does not solve any problems of why they existed in the first place. We further demonstrated that most organizational members maintained a strong mental model in single-loop learning, which made them adhere to the policies even when they were described as being inadequate.

Figure 2 The single loop of the PDSA cycle in the first stage of action research

In the second stage of action research, we routinely questioned the organizational mental models. Questioning does not necessarily mean abandoning. It allows us to think actively about assumptions that we make about the business and the way it behaves. Compared to the first stage of action research, we created a group-learning loop embedded in the PDSA cycle as shown in Figure 3. This group-learning loop made it possible to bring forth new experiences, which subsequently shifted the strong mental model and shaped the context to bring about innovative behaviors. We supposed that only after having new experiences every organization personnel would escape his/her original mental model and commence to kick-start innovative behaviors. We helped the first line staff to select a few solvable problems from which they could invent solutions, and then fit into the double-loop learning process. Over time, small successes breed trust and confidence which increasingly helps the organizational staff to tackle more difficult problems.

However, scholars argue that such double-loop learning is difficult to create and maintain (Argyris & Schön, 1978; Kirton, 1994). They believe that single-loop learning is prevalent in many organizations because they create systems that inhibit double-loop learning. Accordingly, the bureaucratic structure of most organizations exerts pressure on employees to be methodical, prudent, and disciplined, therefore

Figure 3 The double loop of group learning loop embedded in PDSA cycle

reinforcing negative feedback behavior. It implies that double-loop learning may be created not only by our intentional managerial action but also, implicitly, by other unpredictable and uncontrollable processes as we have found in this action research.

Emergent perspective

Although some events followed our process conditioning, most innovative behaviors were essentially unpredictable emergent events that really surprised us. Their novelties did not occur by predetermined procedure, but burst forth by chance. No individual can explain exactly the details of what really happened, but all of us realized that interactive teamwork is always a positive step toward generating innovative behavior.

From the emergent perspective, we could recognize more radically what we had done in the second stage of action research. For example, in our experience, all organization members collectively keep shaping the vision - continually making sense to themselves rather than only the lip service they get from the senior team. When they believe that it is a worthwhile challenge, then they are willing to think and behave differently to approach the new vision. It shows that shaping vision can encourage active questioning of existing mental models, encouraging open minds that can potentially create other new alternatives. In such a recursive process, accompanied with the new knowledge and experience of experimentation, their collective mental model would shift from the original uninnovative one to one that could underlie novel innovative behavior. In addition, the integration of information exchanges, participative interaction, and process conditioning generates the effect of empowerment for organizational members. We found an autonomous learning group, which was definitely deeply influenced by our intentional intervention. We believe that these intentional interventions should be the prerequisite antecedents of unpredictable innovative behavior.

Based on the above interpretations, we can develop an adequate theoretical implication of how we effectively bring out the emergent innovative behavior by our purposeful intervention as shown in Figure 4. Drazin and Sandelands (1992) argue that the autogenesis of complex social organization can be analyzed in terms of three different types of structures:

Figure 4 The recursive process of intervention on complex adaptive system

  1. Observed structure, comprising categories and terms that apply to the perceptions of social interaction as collectives by observers;
  2. Elemental structure, consisting of interactions among individual actors, and;
  3. Deep structure, which consists of rules that generate and govern individual behavior and interactions.

In our research, we consider that the emergent and unpredictable innovative behavior we observed can be seen as observable structure, which can merely be influenced by a specific context; the autonomous learning group can be regard as an elemental structure, which presents the features of self-organization; the organizational mental model can be recognized as one category of deep structure, which can be shifted, but can not be radically changed by intentional interventions. It is the three interactive structures that co-construct an organization as a complex adaptive system that can exhibit both adaptive and innovative capacity.


To effectively build an innovative capacity into an established TQM organization, we have appropriately fulfilled a two-stage action research project in a traditional business. This research approach makes it possible to reach a fundamental understanding of the structure, process and driving forces of upgrading a quality system, rather than a superficial establishment of correlation or overly simplistic cause-effect relationships.

On one hand, from the intentional perspective, our second stage interventions created a group learning capacity which helped the organization to bring forth new experiences, which subsequently shifted the strong ‘linear’ mental model and shaped the context to bring innovative behaviors. This research further provides empirical evidence for the relevance and applicability of double-loop learning theory as developed by Argyris and Schön (1978; see also Argyris, et al., 1985). On the other hand, from the complexity paradigm, it demonstrates that three interactive structures (forming a complex adaptive system) explain how effective managerial actions are the important antecedents of emergent innovative behaviors. In this framework, the autonomous learning group presents the features of self-organization while the organizational mental model can be recognized as one category of deep structure. We conclude that an organization, which can create a balance between both traditional and complexity paradigms, can exhibit both adaptive and innovative capacity - existing (linear) approaches cannot be discarded in the creation of new (nonlinear) approaches.

Similar to the principle of TQM, Gronhaug & Olson (1999) suggest that the action research emphasizes gradual learning and gradual improvement. Nevertheless, in this quality-upgrading project we observed that the resulting synergy of action research is that ‘the whole adds up to more than the sum of its parts’. It shows that longitudinal knowledge creation and learning in cyclic action research is an appropriate methodology to catch knowledge ‘on the move’. Such a process of integrating action and research can reduce the gap between practical and theoretical knowledge. In addition, it enables process theorizing rather than variance theorizing of increasing complex and dynamic business contexts.


Abbott, L. (1955). Quality and competition, New York: Columbia University Press.

Argyris, C. and Schön, D. A. (1978). Organizational learning: A theory of action perspective, Reading, MA: Addison- Wesley.

Argyris, C., Putnam, R. and McLain S. D. (1985). Action science: Concepts, methods, and skills for research and intervention, San Francisco: Jossey-Bass.

Buzzell, R. and Gale, B. (1987). The PIMS principles: Linking strategy to performance, New York: Free Press.

Coghlan, D. (2001). “Insider action research projects: Implications for practicing managers,” Management Learning, 32(1): 49-60.

Constantinos, M. (1997). “Strategic innovation,” Sloan Management Review, 38(3): 9-23.

Dean, J. W. and Evans, J. R. (1994). Total quality management, organization and strategy, West Publishing, St Paul, MN.

Dooley, K. J., Johnson, T. L. and Bush, D. H. (1995). “TQM, chaos and complexity,” Human Systems Management, 14(4): 287-302.

Drazin, R. D. and Sandelands, L. E. (1992). “Autogenesis: A perspective on the process of organizing,” Organization Science, 3: 230-249.

Gilmore, H. L. (1974). “Product conformance cost,” Quality Progress, 7 (5): 16-19.

Greenwood, D. J. and Levin, M. (1998). Introduction to action research, London: Sage.

Gronhaug, K. and Olson, O. (1999). “Action research and knowledge creation: Merits and challenges,” Qualitative Market Research, 2(1): 6-14.

Gummesson, E. (2000). Qualitative methods in management research, London: Sage.

Juran, J. M., and Gryna, F. M., Jr. (eds.), (1988). Juran’s quality control handbook (4th ed.), New York: Mcgraw-Hill.

Kanter, R. (1988). “When a thousand flowers bloom: Structural, collective, and social conditions for innovation in organizations,” in B. M. Staw and L. L. Cummings (eds.), Research in organizational behavior, Volume 10, Greenwich, CT: JAI Press, pp. 169-211.

Kirton, M. J. (ed.). (1994). Adaptors and innovators: Styles of creativity and problem-solving (2nd ed.), New York: Routledge.

Lewin, K. (1946). “Action research and minority problems,” Journal of Social Issues, 2(4): 34-46.

Lindgren, R., Henfridsson, O. and Schultze, U. (2004). “Design principles for competence management systems: A synthesis of an action research study,” MIS Quarterly, 28(3): 435-462.

McKelvey, B. (1999). “Complexity theory in organization science: Seizing the promise or becoming a fad?” Emergence, 1(1): 5-32.

Ovretveit, J. (2002). “Producing useful research about quality improvement,” International Journal of Health Care Quality Assurance, 15(6/7): 294-302.

Prahalad, C. K. and Krishnan, M. S. (1999). “The new meaning of quality in the information age,” Harvard Business Review, 77(5): 109-118.

Prajogo, D. I. and Sohal, A. S. (2001). “The relationship between TQM practices and innovation performance: A literature review and research framework,” Technovation, 21(9): 539-58.

Parasuraman, A., Zeithaml, V. A., and Berry, L. L. (1985). “A conceptual model of service quality and its implications for future research,” Journal of Marketing, 4(4): 41-50.

Plsek, P. E, (2000). “Creative thinking for surprising quality,” Quality Progress, 33(5): 67-73.

Rapoport, R. N. (1970). “Three dilemmas in action Research,” Human Relations , 23(6): 499-513.

Roffe, I. (1999). “Innovation and creativity in organizations: A review of the implications for training and development,” Journal of European Industrial Training, 23(4/5): 224-37. Scott, S. G, and Bruce, R. A. (1994). “Determinants of innovative behavior: A path model of individual innovation in the workplace,” Academy of Management Journal, 37(3): 580-607.

Slater, S. F. and Narver, J. C. (1998). “Customer-led and market-led: Let’s not confuse the two,” Strategic Management Journal, 19(10): 1001-1006.

Stacey, R. (1996). Complexity and creativity in organization, San Francisco: Berrett Koehler.

Tsai, S. D., Chiang, H. Q. and Valentine, S. (2003). “Integrated model for strategic management in dynamic industries: Qualitative research from Taiwan’s passive component industry,” Emergence, 5 (4): 34-56.

Weick, K. E. (1995). Sensemaking in organizations, CA: Sage Thousand Oaks.

Wind, J. and Mahajan, V. (1997). “Issues and opportunities in new product development: An introduction to the special issue,” Journal of Marketing Research, 34(1): 1-12.