Abstract
We proposed and examined a learning organization framework in higher education within an international context. Using a sample of 687 employees in the UK and Vietnam, we tested the relationships between personal mastery, mental models, team learning, shared visions and systems thinking with their antecedents and outcomes. Our findings support the suggested learning organization model. As predicted, these five variables partially mediate the relationship between the antecedents and outcomes. We also found that employees in a collectivist culture were more likely to be committed to the process of becoming learning organizations compared with those from an individualistic culture.
Introduction
Organizational learning is a well-established and growing area within the wider area of management and organizational research (Argyris and Schön, 1996). However, there is an ongoing debate about units/levels of analysis of learning organization. While some argue that organizational learning is the sum of what individuals learn within organizations, others assert that organizational learning is more than the learning of its individual members (Easterby-Smith et al., 2000). We take organizational learning to be the sum of individual and collective learning.
In order to remain viable in an uncertain and changing environment, organizations and individuals alike rely on an ability to learn (Edmondson and Moingeon, 1998; Senge, 1990). Thus, a learning organization (LO) must constantly strive to develop and implement policies and strategies, which encourage and make use of learning at all levels within the organization (following Church, 2002; Hitt, 1995; Hodgkinson, 2000; Rowley, 1998; Sackmann et al., 2009). In so doing, they are likely to create and deliver innovative products and services (Church, 2002; Corbett, 2005). LOs align people’s learning and development continuously to corporate vision, mission and strategy (Harrison, 1993). A LO works to create values, practices and procedures, in which learning and working are synonymous throughout the organization (Rowley, 1998). Learning is a core part of all operations. The LO process challenges employees and communities to use their collective intelligence, ability to learn, and creativity to transform the existing system (Bierema, 1999). It helps people to connect with each other, their work, and their community. It is not a programme, but rather a new process for understanding and learning together. In our study, we adopt Senge’s definition for LOs:
organizations where people continually expand their capacity to create the results they truly desire, where new and expansive patterns of thinking are nurtured, where collective aspiration is set free, and where people are continually learning how to learn together (Senge, 1990: 3).
LOs are expected to build the capability to adapt and confront change-related issues in order to prevail and excel, and to meet the challenges inherent in an unstable environment (Bokeno and Gantt, 2000; Senge et al., 1999). In the knowledge economy, institutions of higher education (HE) constantly need innovation. Business schools with a strong professional learning base are better able to offer valid pedagogy and be more effective in promoting student achievement (Sackney and Walker, 2006). Though Senge (2000) posits that changes in institutions of public education are harder to deliver than in business, Resnick and Hall (1998) argue that LOs bring about sustainable educational reforms. True learning communities encourage educators to acquire the knowledge and skills that they need from many sources, inside and outside the LOs. They openly share their own knowledge and skills with others because they realize that they are all working toward achieving personal and professional goals (Sackney and Walker, 2006).
While the literature provides ample support for the concept of the LO, both theoretical (Sackmann et al., 2009) and managerial (Garvin et al., 2008), there are clear gaps that call for investigation. First, related studies have been conducted mostly in western work environments (Birdthistle, 2008; Blackman and Henderson, 2005; Nyhan et al., 2004). Second, the educational sector, which is growing in scope and importance, is rarely represented in LO studies (White and Weathersby, 2005). Third, most of the studies which focus on LOs have employed qualitative, as opposed to quantitative, research methods (Eijnatten and Putnik, 2004; Harvey and Denton, 1999; McGill et al., 1992; Slater and Narver, 1995). Fourth, while Senge’s contribution has attracted vast scholarly and practical attention, there remains an acute shortage of empirical investigation of his work (Nair, 2001). Specifically, Ellinger and Bostrom (2002) call for further studies to explore the relationships between characteristics of the LO and organizational performance.
To bridge these gaps, we developed a conceptual framework that included a wide set of antecedents and outcomes of Senge’s five disciplines (the components of a LO), comprising personal mastery, mental models, shared vision, team learning, and systems thinking. We tested the framework within an international context of HE. Employing a multi-level analysis (Klein and Kozlowski, 2000), we started at the individual level (personal mastery) through the collective level (team learning, mental models), and up to the organizational level (shared vision, systems thinking).
Our study answers these deficiencies by addressing the issue in the HE sector, and comparing universities from two distinct cultures. We benefited from a unique opportunity to examine and compare the LO model in two different cultures, namely East and West, to recommend theoretical and managerial implications. Further, we employ quantitative research methods to complement earlier investigations which have tended to be based on qualitative methodology.
Theoretical background
The context of higher education in a global environment
Like any other sector, the HE is under increasing pressure to improve its competitiveness. The competition in HE is getting more severe within and across national borders (Marginson, 2007), with indications of an emerging global phenomenon described as ‘brain drain’. This phenomenon can be observed in the movement of highly educated people from developing countries to developed countries (Baruch et al., 2007; Carrington and Detragiache, 1999). However, there is also evidence to suggest that this phenomenon may turn into a ‘brain circulation’ (Carr et al., 2005; Saxenian, 2006; Tung and Lazarova, 2006), particularly within a global open labour market.
Management should cope with fast-paced social, economic, and political transitions that place extensive demands on the system and its employees. The western HE sector is still operating in a lucrative market, but the situation is changing, as many developing countries are catching up and establishing their own high quality HE systems (for example, see Altbach and Selvaratnam, 1989; Marginson, 2007). In such context, LOs bring about sustainable educational reform (Resnick and Hall, 1998), and as a result, a wide range of organizations across the globe adopt the LO framework (Davies, 1998; Franklin et al., 1998; Patterson, 1999; Rowley, 1998; Willcoxson, 2001; Yeo, 2006; Yeo and Marquardt, 2010).
The learning organization and higher education
Senge’s (1990) LO is depicted as pragmatic, normative, and inspirational (Easterby-Smith, 1997; Roper and Pettit, 2002). Easterby-Smith (1997) adds that LOs encourage organizations to go beyond single-loop learning (actions as learning repeating in a routine) to double-loop learning (attempts to change normal practice, i.e. single-loop learning), and even triple-loop learning (learning about learning, about redesigning the current systems for better learning and operations) (Argyris, 1999; Argyris and Schön, 1996). Amidon (2005) argues that Senge designs a set of practices in a robust way that might lead to a knowledge-worker utopia. Senge attempts to build organizations that ‘serve humans rather than enslave them’ (Amidon, 2005: 408).
Senge’s LO is considered inspirational, as it possesses a power to nurture practical creativity. Though there is no specific prototype for developing universities as LOs, Senge’s (1990) model is highly regarded as an essential framework for their construction (Gudz, 2004; Jackson, 2000; Patterson, 1999; White and Weathersby, 2005). Like business, HE is facing challenges of changes and innovation. Many HE institutions have adapted LO models to facilitate progress and advancement in line with economic changes and technological development (Duke, 1992; Patterson, 1999).
Based on the extant literature we posit a model which comprises several antecedents and outcomes of the five disciplines. These are depicted in Figure 1 and explicated in the hypothesis development section.

The LO model.
Research hypotheses
Personal mastery
Personal mastery refers to a personal commitment to continuously clarify and deepen a personal vision, focus energy, develop patience, and endeavour to see reality as objectively as possible (Appelbaum and Goransson, 1997). It is ‘the learning organization’s spiritual foundation’ (Senge, 1990: 7). It goes beyond competence, skills, and spiritual unfolding though it is grounded on those bases (Senge, 1990). As discussed by Senge (1990), personal mastery is influenced by personal values, personal vision, motivation, and individual learning through training and development. First, personal values are rooted in an individual’s own set of values, beliefs, and aspiration (Homer and Kahle, 1988; Kahle, 1983; Senge, 1990). Since employees are believed to bring their values into the work setting (Robertson, 1991), the impact of personal values is of special relevance in educational systems (Blackmore and Castley, 2005). Second, personal vision is the ‘groundwork’ for continually expanding personal mastery (Appelbaum and Goransson, 1997; Nightingale, 1990; Senge, 2006). There is an increased confidence in employees’ personal vision when universities develop as LOs (Smith, 2003; Wheeler, 2002). Third, motivation has long been studied to explain why humans are inspired to do certain things (Deci, 1975; Kanfer and Ackerman, 2000; Maslow, 1970; Rueda and Moll, 1994; Siebold, 1994). An individual with high personal mastery would be self-motivated (Ng, 2004). Fourth, individual learning is the primary learning-enabling element in LOs (Antonacopoulou, 2000b; Dodgson, 1993; Senge, 1990; Watkins and Marsick, 1993). Individual learning can promote personal mastery (Gong et al., 2009). Life-long learning is a part of commitment to personal mastery (Barker et al., 1998; Davies, 1998), and to LOs (Davies, 1998). Academic scholars are highly qualified in terms of formal education and rich in informal learning (Baruch and Hall, 2004; Knight et al., 2006). Fifth, development and training is essential for employees’ personal mastery (Senge et al., 1994). This process plays a significant role in making employees aware of Senge’s LO concepts, including personal mastery (Kiedrowski, 2006). Research also shows the effect of development and training on personal mastery (Antonacopoulou, 2000a; Blackman and Henderson, 2005). Many countries’ HE makes development and training a top priority (Blackmore and Castley, 2005; Dalin, 1998; Maslen, 1992).
Hypothesis 1. Personal values, personal vision, motivation, individual learning, and development and training are positively associated with personal mastery.
Mental models
Mental models refer to deeply-held assumptions or metaphors through which people interpret and understand the world, and take action (O’Connor and McDermott, 1997; Senge, 1990). Mental models are powerful in influencing human behavior (Senge, 1990). Mental models are assumed to be composed of a number of factors. First, organizational commitment is defined as ‘the relative strength of an individual’s identification with and involvement in a particular organization’ (Mowday et al., 1979: 226). Despite Gallie et al.’s (2001) finding that British employees had a relatively weak commitment to their organizations, organizational commitment is considered to be at the heart of mental models (Kofman and Senge, 1993). It is essential for the development and sharing of mental models (Bui and Baruch, 2010b). Second, leaders are responsible for learning and creating a learning environment for people to continually expand their ability to understand complexity, clarify vision, and improve shared mental models (Fullan, 1993; Horner, 1997; Marquardt, 1996; Marsick and Watkins, 2003; Mazutis and Slawinski, 2008; Mintzberg, 1998). For our study, we consider leadership to be relevant to identifying mental models that challenge organizational members with the question: ‘What values do you really want to stand for?’ (Senge et al., 2000: 67). Third, organizational culture is a key factor of organizational management (Pettigrew, 1979). It is about fundamental assumptions that people share about an organization’s values, beliefs, norms, and symbols that give meaning to organizational membership and are considered guides to behaviours (Bloisi et al., 2007; Tyler and Gnyawali, 2009). Organizational culture is highly influenced by the societal culture in which it is embedded (Dimmock and Walker, 2000; Hofstede, 2001). Thus, organizations with a low power distance culture are more likely to succeed in mastering mental models than those in cultures with high power distance (Alavi and McCormick, 2004) because a culture of trust and openness that encourages inquiry and dialogue is needed to challenge assumptions (Gephart et al., 1996; Jones et al., 2005; Senge, 1990; Watkins and Marsick, 1993).
Hypothesis 2. Organizational commitment, leadership, and organizational culture are positively associated with mental models.
Team learning
Team learning is a ‘process of aligning and developing the capacity of a team to create the results its member truly desire’ (Senge, 1990: 236). It is regarded as a fundamental unit of LOs (Hitt, 1995; Senge, 1990). Despite its importance, team learning among academics remains poorly understood (Nissala, 2005: 211). However, a study on team learning have found that some universities have succeeded in creating team learning through the process of becoming LOs (Bender, 1997). The literature shows that team learning is composed of a number of factors. First, team commitment is the essence of team learning (Katzenbach and Smith, 2004; Park et al., 2005). Talented individuals do not necessarily form talented teams if they lack team commitment (Senge, 2006). When people are committed to team learning, they tend to set clear goals for the team and themselves (Kofman and Senge, 1993). Second, leadership inspires innovation and the creation of knowledge within team members (Kozlowski et al., 1996; Lee et al., 2010; Wageman, 2001). The most successful teams have leaders who proactively manage team learning efforts (Choi, 2009; Edmondson, 2003; Edmondson et al., 2004; Marsick and Watkins, 2003; Morgeson et al., 2010). Leaders are expected to create a learning environment in their organization (Mazutis and Slawinski, 2008). Leadership is of significance in education because it motivates educators, students, and parents to learn together (Sackney and Walker, 2006). Third, goal setting is important for evaluating the result of team learning and team performance (Mehta et al., 2009). The more educated staff are, the more participative and effective their goal setting is likely to be (Ivancevich and McMahon, 1977). Fourth, development and training related to team skills is important for successful learning (Bowen, 1998; Druskat and Kayes, 2000), and enhances team learning (Morgeson et al., 2010; Prichard et al., 2006). Development and training is a priority in HE in many countries (Brancato, 2003; Dalin, 1998; Maslen, 1992). Fifth, organizational culture determines the effectiveness of team learning and team working (Elfenbein and O’Reilly III, 2007). A LO’s culture should support and reward learning and innovation; promote inquiry, dialogue, risk-taking, and experimentation; allow mistakes to be shared and viewed as opportunities for learning; and value the well-being of all employees (Gephart et al., 1996: 39). Sixth, individual learning is at the core layer of organizational learning (Antonacopoulou, 2006; Pedler et al., 1991). In other words, organizational learning is the product of individual learning (Antonacopoulou, 2006; Argyris and Schön, 1978; Fiol and Lyles, 1985; Senge, 1990; Watkins and Marsick, 1993). As a result, we hypothesized:
Hypothesis 3. Team commitment, leadership, goal setting, development and training, organizational culture, and individual learning are positively associated with team learning.
Shared vision
Shared vision is a vision to which people throughout an organization are truly committed (Senge, 1990). It is important for bringing people together and fostering a commitment to a shared future (Dufour and Eaker, 1998; Griego et al., 2000). Shared vision should begin with a personal vision to which one is committed (Appelbaum and Goransson, 1997). Employees’ personal vision is a crucial ingredient in the pursuit of LOs (Senge, 1996). Personal vision relies not only on individuals but also requires the support of their organizations (Senge et al., 1994). Second, personal values also contribute a certain degree of commitment to shared visions (Eigeles, 2003; Gudz, 2004; Senge, 2006). Within any society, educators are expected be have high levels of personal values (Haydon, 1997). Third, leadership is vital in ensuring that organizational vision is shared across the organization, as leaders act as designers, stewards, and teachers (Fullan, 1993; Senge, 1990). Leaders often possess extraordinary vision and commitment to high ideals (Fullop and Linstead, 1999; Senge et al., 2000). They constantly look for new information and opportunities that can help them and others to realize their vision (Mintzberg, 1998; Schrage, 1990). In reality, however, many leaders tend to be good designers and teachers, but not good stewards (Gudz, 2004; Tsai and Beverton, 2007). Fourth, motivation is a key factor that leads employees to sharing their vision with the organization (Senge, 1990). Fifth, organizational culture is regarded as a catalyst for creating a shared vision (Reeves and Boreham, 2006). Sharing vision seems to be more effective in organizations that are embedded in a high societal collectivism and future orientation culture (Alavi and McCormick, 2004), i.e. in an open, dynamic and group-oriented cultures.
Hypothesis 4. Personal vision, personal values, leadership, and organizational culture are positively associated with shared vision.
Systems thinking
Any attempt at creating a LO must start from systems thinking (Appelbaum and Goransson, 1997; Senge, 1990). This is the capacity to see deeper patterns lying beneath events and details, in the full setting of interconnecting elements (Hosley et al., 1994; Senge, 1990; Senge et al., 1994). Bui and Baruch (2010b) propose some antecedents for systems thinking. First, individual competence includes systems thinking which actively contributes to personal and professional success (Anderson et al., 2006; Antonacopoulou, 1996; Marquardt, 1996). People from all parts of the organization who are competent and genuinely committed to deep changes in themselves and in their organizations are considered as leaders (Senge, 1996; Spreitzer, 1995). Second, leaders need systems thinking to recognize those people who will be influenced by their decisions (Dixon et al., 2010; Kumar et al., 2005; Yang et al., 2010). Furthermore, leaders in LOs should be ‘philosopher kings, who are willing to inquire into their own competence for kingship (Coopey, 1995). In HE, the way that universities are organized into disciplines may create the false impression that the real world is divided into fragmented parts (Vo et al., 2006). To overcome such possible pitfalls, leaders in HE benefit from the concept of systems thinking. Third, organizational culture can affect systems thinking. Alavi and McCormick (2004) found that organizations high in societal collectivism tend to be successful when working collectively, as their staff tend to be more inclined to effectively take part in teams for systems thinking. Systems thinking was found to be applied successfully in HE (Austin, 2000; Wright, 1999).
Hypothesis 5. Competence, leadership, and organizational culture are positively associated with systems thinking.
Outcomes
Personal mastery is believed to lead to self-efficacy and better individual performance (Bloisi et al., 2007; Senge, 1990). In addition, personal mastery can create a balanced work and home life (Baruch, 2004; Doherty and Manfredi, 2006; Johnson, 2006). When mental models and team learning are developed and learnt throughout the organization, one of the outcomes is a higher level of knowledge sharing and knowledge creation (An and Reigeluth, 2005; Argyris, 1999; Senge, 2006; Watkins and Marsick, 1993). Developing appropriate mental models and team learning generates more knowledge and can consequently lead to improving job performance (Chan et al., 2003; Davenport et al., 1998; Inkpen, 2000; Pedler et al., 1991). Fullan (2004) also stresses the importance of systems thinking in organizational strategic planning. To generate impact, strategy should be backed by passion and vision from the people who create and implement it (Domm, 2001: 46). Without systems thinking, it would be difficult, if not impossible, to develop a strategy that fits the organization. As a result, we hypothesize:
Hypothesis 6. The five disciplines are positively associated with individual performance, knowledge sharing, self-efficacy, work-life balance, and strategic planning.
We combine hypotheses 1 through 6 to test the following hypothesis:
Hypothesis 7. The five disciplines mediate the relationship between the antecedents and outcomes.
Cultural differences
Cultural differences may have a dramatic impact on management (Fadil et al., 2005; Hofstede, 1980, 1984; Schwarz, 1992; Sun et al., 2004). Because our study was conducted within an international context, cultural differences were considered to be a major factor.
Based on Hofstede’s study (1984, 1993, 2001), the UK scored very highly in individualism and masculinity. Individuals realize their potential in what they do, i.e. they often have intrinsic motivation for their work. People put priority on the task (Hofstede, 1993, 2001). In the UK, a low power distance culture, educators encourage independence in students, encouraging them to challenge themselves, each other, and even their teachers (Hofstede, 1984).
In contrast, Vietnam is a collectivist culture (Grinter, 2006; Smith and Pham, 1996). In collectivist cultures, people place greater emphasis on relationships (Hofstede, 1984; Smith and Pham, 1996). This is particularly true in education where students try to avoid arguments or expressing different viewpoints to their teachers’. In addition to these cultural differences, the two countries are also different in infrastructure. One country is a developed nation whilst the other is developing. This may affect the process of becoming learning organizations in the two countries. Therefore we hypothesize:
Hypothesis 8. The components of the learning organization model in the UK are expected to have higher values/scores than in Vietnam.
Methods
Sample and data collection
The two selected organizations were established universities in Vietnam and the UK. Pilot tests were conducted before the main data collection to ensure that the questionnaires would function similarly in two different countries. An initial pilot study was conducted with a sample of 40 participants (25 in the UK and 15 in Vietnam). The pilot study confirmed the clarity and relevance of the questionnaire, and preliminary analysis conducted in both countries to assess internal reliability generated positive outcomes. These and the respondents’ feedback suggested there was no need for significant changes to the final questionnaires, except for some re-wordings. The questionnaires were presented in English and Vietnamese. In order to prevent any methodological problems that may stem from the translation of the questionnaires (Sperber et al., 1994) a combination of three questionnaire translation techniques was employed including back-translation, committee approach, and pre-test procedure (Brislin, 1976; Hofstede, 1980; Sperber et al., 1994).
Stratified sampling was used to ensure representativeness (Wiersma and Jurs, 2005), and to decrease bias in the data, as it eliminated subjectivity in choosing a sample (Fink, 1995). The population was divided into subgroups of academics (i.e. those who were employed for teaching, research, or both) and non-academics. A number of schools/departments were selected at random within the UK university, including the School of Environment and Earth Science, the School of Computing and Technology, the School of Mathematics, the School of Business, the School of Education, the School of Language and Literature, the School of History, and the Registry Office (including its managing board, administrators and assisted staff). After that, questionnaires were sent to academic and non-academic staff of all levels in those schools. The equivalent process was conducted in the Vietnamese university, after identifying schools that matched those sampled in the UK university.
The staff population in each university was approximately 3200 people. A total of 1391 questionnaires were distributed in person, 711 in Vietnam and 680 in the UK based on strata that we determined above. Participation was anonymous and voluntary. Respondents could either return the completed questionnaires via the stamped envelope or make appointments for the researchers to collect them later. As 96 employees in both universities (61 in Vietnam and 35 in the UK) were away on study, maternity, or sabbatical leave; 1285 employees (640 in Vietnam and 645 in the UK) actually received the questionnaires. After systematically screening for missing data, 687 completed questionnaires (341 in Vietnam and 346 in the UK) were used for analyses. This represents an effective response rate of 53.5%, which is above the norm in social sciences (Baruch and Holtom, 2008; Neuman, 2000; Roth and BeVier, 1998).
Most of the respondents were highly educated with 308 (44.8%) PhD holders, 141 (20.5) Masters degree holders, and 146 (21.3%) Bachelor degree holders. A total of 315 (45.9%) respondents had worked in their organizations for five years or less, and 362 (52.7%) respondents had worked there for more than five years. 446 respondents (64.9%) were academics, while 241 respondents (35.1%) were non-academics.
Measures
Participants responded to the items listed below using an ordinal scale ranging from 1 to 7, with 1 represents ‘disagree strongly’ to 7 represents ‘agree strongly’, or ‘not at all important’ to ‘extremely important’. The choice of the measures employed was based on academic criteria, namely the relevance and suitability of items to our literature review and model development. We opted for existing and already validated scales which had been frequently utilized in earlier studies.
Personal values. Five out of nine items were taken from Kahle (1983). They were self-respect, security, warm relationships with others, and sense of accomplishment. The Cronbach alpha was .72.
Competence. A three-item scale of competence was employed from Spreitzer (1995). A sample item is ‘I am confident in my ability to do my job’. The Cronbach alpha was .85.
Development and training. Though development and training has been studied in a number of research studies (Forssen and Haho, 2001; Noe, 2002), no suitable construct of development and training was found. Therefore, a four-item scale was created to measure development and training in organizations. These items comprised ‘This University encourages staff to develop team-working skills’, ‘This University encourages staff to identify skills they need to adapt to changes’, ‘I was mentored when I first took up the job here’, and ‘I receive the training I need to perform my current job effectively’. The Cronbach alpha was .78.
Team commitment. Four items of team commitment were adopted from West (2004). ‘In my work, I lead myself by the goals of my team’ is a sample item. The Cronbach alpha was .85.
Goal setting. Three items of goal setting for teams were adopted from Ivancevich and McMahon (1977). ‘The goals assigned to our team members are equitable and can be accepted’ is a sample item. The Cronbach alpha was .89.
Motivation. A four-item scale from Siebold (1994) was employed to measure motivation. A sample item of motivation is ‘I am very personally involved in my work’. The Cronbach alpha was .76.
Individual learning. This is a four-item measure. Two items were taken from Baruch and Peiperl (2000). One of them was ‘My own learning and development at work are essential to me’. Another item was borrowed from Kanfer and Ackerman (2000). ‘I prefer activities that provide me the opportunity to learn something new’. One more item was created to measure individual learning, ‘I am committed to life-long learning’. The Cronbach alpha was .86.
Personal vision. With no known measure for this construct existing in the literature, we designed a four-item scale to measure personal vision. We developed the items based on the literature. These items were ‘I set up career goals of my own’, ‘I have my personal vision for my career’, ‘Part of my personal vision is to make the University more successful’, and ‘I understand how the work I do helps this University achieve its vision’. The Cronbach alpha was .83.
Organizational commitment. Four out of eight items of affective organizational commitment by Meyer and Allen (1991) were taken to fit the construct of the research, in line with earlier use of these items (Iverson and Buttigieg, 1999; Meyer et al., 2002; Riketta, 2002). A sample item is ‘I enjoy discussing the University with people outside it’. The Cronbach alpha was .85.
Leadership. Six items to measure leadership were taken from Marsick and Watkins’s (2003) The Dimensions of the Learning Organization Questionnaire (DLOQ). The questionnaire has been validated as a research tool (Marsick and Watkins, 2003; Yang, 2003). ‘At this university, leaders/managers continually look for opportunities to learn for their professional development’ was one of the sample items. The Cronbach alpha was .90.
Organizational culture. Employees’ perceptions of organizational culture were measured by asking them to evaluate the culture within their organization on six scales adopted from Baruch and Peiperl (2000). Each scale ranges from 1 (one extreme) to 7 (the opposite extreme). The scales were: stable—dynamic; closed/bureaucratic—open/interactive; reactive—proactive; individual oriented—group oriented; aggressive—accommodating; reserved—friendly (Baruch and Peiperl, 2000). The direction was such that the second extreme is the one considered more positive or fit for contemporary labour markets. The Cronbach alpha was .90.
Personal mastery, team learning, shared vision, and systems thinking. Reed (2001) employed Senge’s model quantitatively in research on realizing a learning organization in a medium-sized company. The size of the universities was relatively similar to a private sector medium-size company. Although the main approach of his research was qualitative, Reed used quantitative research for his pilot test. He designed a set of questions to measure four constructs: personal mastery, shared vision, team learning, and systems thinking. The Cronbach alpha of personal mastery, team learning, shared vision, and systems thinking were .69, .73, .91, and .80 respectively.
Mental models. With no known measure for this construct in the literature, we designed a five-item measure based on a literature review of the definition of mental models. Those five items included, ‘The University treats its staff fairly’, ‘The University provides necessary skills for professional development’, ‘The University provides opportunities for professional development’, ‘The University has an appropriate mechanism to bring staff together to develop the best ways of problem solving’, and ‘When there is a problem in my department, we also seek help to handle the problem from other schools/departments’. The Cronbach alpha was .90.
Self-efficacy. A three-item scale from Tierney and Farmer (2002) was employed to measure self-efficacy. The scale has been successfully used in a number of studies (Miron et al., 2004; Rindova et al., 2005; Shalley and Gilson, 2004). A sample item is, ‘I have confidence in my ability to solve problems creatively’. The Cronbach alpha was .81.
Work-life balance. Four items to measure work-life balance were taken from Hayman’s (2005) work. Sample items are ‘My job makes me happy’, and ‘My personal life makes me happy at work’. The Cronbach alpha was .83.
Knowledge sharing. Four out of five items measuring knowledge sharing were adopted from Bock et al. (2005), which were originally developed by Fishbein and Ajzen (1975). A sample item is ‘My knowledge sharing would help other members at the university solve problems’. The Cronbach alpha was .94.
Strategic planning. Strategic planning was assessed using four items adapted from two studies (Bryan and Joyce, 2007; Dye and Sibony, 2007). The Cronbach alpha was .93.
Performance appraisal. Two different sets of performance appraisals were used, for academic and non-academic staff respectively. The one-item scale for measuring non-academic staff was based on Baruch (1996). Respondents were asked to rank their performance score from number 1 (too early to assess) to number 7 (outstanding/effective). A two-item scale was used for academic staff, one item was for teaching performance, ranking from 1 (unacceptable) to 7 (outstanding), and the other was for research performance based on RAE score (in terms of the 2001 Research Assessment Exercise). Performance was also cross-checked. Prior to the questionnaire delivery, actual individual research performance in terms of publication output was checked and classified into 7 ranks equivalent to the research performance question in the questionnaires. This external evaluation was based on journal quality rankings, and enabled us to avoid common method bias in the outcomes data. The respondents’ questionnaire answers and the actual individual research performance were highly correlated (R2 = .53, p < .05). The actual research publication output, reflecting individual performance, was used for analysis in this research.
Avoiding common method bias
As most of the data rely on self-reporting, common method variance was a potential methodological problem (Campbell and Fiske, 1959; Podsakoff and Organ, 1986). In order to reduce the risk, a number of procedural remedies suggested by Podsakoff et al. (2003) were employed in this study. First, item ambiguity was substantially removed via the pilot tests. Second, measures of the predictors and criteria variables were obtained from different sources, e.g. the cross-check of research performance data prior to the questionnaire delivery. Third, in the invitation letter, participants were informed that their answers were confidential, and that there were no right or wrong answers, in order to tackle possible evaluation apprehension. Finally, the question order was counterbalanced to control ‘priming effects, item-context-induced mood states, or other biases related to the question context or item embeddedness’ (Podsakoff et al., 2003: 888).
Factor analysis
Both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were conducted to validate all the variables employed in this research. The EFA test was used for validating distinctiveness of the newly developed measures. The CFA test was used for variables already validated in the literature.
Exploratory factor analysis
Two groups of variables, which were related conceptually to one another, were validated. The first group consisted of development and training, personal vision and individual learning. The determinant of the R-matrix is .003 (p > .00001), which shows that multicollinearity is not a problem for this piece of data (Field, 2005). The factor loadings of those variables in the rotated component matrix showed no cross loading, suggesting significant orthogonality among the factors (Field, 2005; Tabachnick and Fidell, 2001).
The second group included the five disciplines of LO. The factor loadings of those variables in the rotated component matrix showed some cross loadings. After revising, an item was dropped from each of the personal mastery, team learning, and systems thinking scales. Therefore, there were four items in these three scales and five items in mental models and shared vision scales.
Confirmatory factor analysis
As some variables in this study were related conceptually, a series of CFAs were conducted to verify each variable’s distinctiveness with the employment of AMOS. Another reason for conducting these was to check for common method variance—Harman’s one-factor test, as prescribed by Podsakoff and Organ (1986). We conducted three sets of CFA. The first (CFA1) covered measures relating to constructs at the organizational level. These were: organizational culture, leadership, organizational commitment, and strategic planning. The second (CFA2), covered measures relating to constructs at the team level. These were: team commitment, goal setting, and knowledge sharing. The third (CFA3) covered measures relating to constructs at the individual level: personal values, motivation, competence, work-life balance, and self-efficacy. The results of these three CFAs are presented in Table 1.
Confirmatory factor analysis.
The results of all the CFAs indicated that the data fit the models well, except the chi-square. The use of the chi-square statistic is problematic because of its sensitivity to violations of normality and its relation to sample size (i.e. with a large sample size, very small discrepancies between the expected and actual correlations will produce a significant χ 2 (Cole, 1987). This study has a large sample size (687) that leads to the problem of significant chi-square values. Other than that, all the fit indices were within the recommended range (Byrne, 2001; Hair et al., 2006; Hu and Bentler, 1999; Tabachnick and Fidell, 2001). Therefore, all these constructs were retained for further analysis.
Table 2 presents the means, standard deviations, number of respondents, their correlations, and Cronbach alpha values for all variables of interest.
Descriptive statistics and Pearson correlational matrix.
Note: n = 693; Alpha coefficients are presented on the diagonal. Decimal points are deleted;* p < .05 (2-tailed); ** p < .01 (2-tailed).
Regression analyses
We employed hierarchical regressions to test the hypotheses. Table 3 presents the regression results for testing Hypotheses 1 through 5. The three control variables including qualifications, job roles, and tenure were entered in Step 1 and the respective independent variables were entered in Step 2.
Prediction of the five disciplines.
Note: n = 687; *p < .05; **p < .01.
Table 3 shows a significant increase in R2, significant Fchange and coefficients in the respective regressions. Hypothesis 1 predicted that development and training, personal vision, individual learning, motivation, and personal values would be positively associated with personal mastery. As shown in Table 3, development and training (β = .373, p < .01), personal vision (β = .180, p < .01), motivation (β = .128, p < .05), and personal values (β = .147, p < .05) were positively associated with personal mastery. Individual learning (β = -.009, ns) was not associated with personal mastery. Hypothesis 1 was mainly supported. Hypothesis 2 posited that organizational commitment, organizational culture, and leadership would be positively associated with mental models. Table 3 shows that organizational commitment (β = .165, p < .01), organizational culture (β = .438, p < .01), and leadership (β = .301, p < .01) were all related to mental models. Thus, Hypothesis 2 was fully supported. Hypothesis 3 predicted that development and training, individual learning, team commitment, goal setting, organizational culture, and leadership would be positively associated with team learning. As shown in Table 3, individual learning (β = .121, p < .05), team commitment (β = .154, p < .01), goal setting (β = .085, p < .05), organizational culture (β = .167, p < .01), and leadership (β = .123, p < .01) were positively associated with team learning. Development and training (β = .032, ns) was not associated with team learning. Thus, Hypothesis 3 was mainly supported. Hypothesis 4 predicted that personal vision, motivation, personal values, organizational culture, and leadership would be positively associated with shared vision. Table 3 shows that personal vision (β = .444, p < .01), motivation (β = .120, p < .05), personal values (β = .119, p < .05), and organizational culture (β = .407, p < .01) were positively associated with shared vision. Leadership (β = .028, ns) was not related with shared vision. Thus, Hypothesis 4 was mainly supported. Hypothesis 5 posited that competence, organizational culture, and leadership would be positively associated with systems thinking. Table 3 shows that competence (β = .173, p < .01), organizational culture (β = .422, p < .01), and leadership (β = .148, p < .01) were all positively associated with systems thinking. Hypothesis 5 was fully supported.
Table 4 presents the regression results tested Hypothesis 6. This hypothesis predicted that the five disciplines, including personal mastery, mental models, team learning, shared vision, and systems thinking, would be positively associated with individual performance, knowledge sharing, self-efficacy, work-life balance, and strategic planning. The three control variables were entered in Step 1 and the respective independent variables were entered in Step 2.
Prediction of the outcomes.
Note: n = 687; *p < .05; **p < .01.
As shown in Table 4, the five disciplines were not associated with administration performance. Team learning (β = .130, p < .05) was positively associated with teaching performance. In contrast, mental models (β = -.226, p < .01) was negatively associated with research performance. Mental models (β = -.170, p < .01) was also negatively associated with knowledge sharing. Team learning (β = .164, p < .01), shared vision (β = .405, p < .01), and systems thinking (β = .106, p < .05) were positively associated with knowledge sharing. Team learning (β = .249, p < .01) and shared vision (β = .130, p < .01) were positively associated with self-efficacy. Mental models (β = -.117, p < .05) was again negatively associated with self-efficacy. Personal mastery (β = .166, p < .01), mental models (β = .172, p < .01), team learning (β = .096, p < .05), and shared vision (β = .124, p < .05) were positively associated with work-life balance. Personal mastery (β = .236, p < .01), mental models (β = .462, p < .01), and shared vision (β = .206, p < .01) were positively associated with strategic planning. The results show that the five disciplines are positively associated with strategic planning, knowledge sharing, self-efficacy, and work-life balance, but not with individual performance, except with teaching. Hypothesis 6 was partially supported.
Hypothesis 7 predicted that the five disciplines would mediate the relationship between the antecedents and outcomes. Based on the results of Hypotheses 1 through 6, a list of the antecedents, including organizational culture, leadership, personal vision, personal values, and motivation and a list of the outcomes, including knowledge sharing, strategic planning, and work-life balance were selected to test Hypothesis 7. Mediation was tested using the method recommended by Baron and Kenny (1986). The three control variables were entered in Step 1, the antecedents were entered in Step 2, and the disciplines were entered in Step 3.
Results from the analyses for Hypotheses 1 through 5 satisfied Baron and Kenny’s requirement that the predictors be related to the mediators (see Table 3). The selected antecedents were significantly associated with the five disciplines. The mediators must also be related to the outcomes when using Baron and Kenny’s method. This condition is satisfied via Hypothesis 6. The five disciplines were significantly associated with the selected outcomes (see Table 4). Results of the five mediating relationships are shown in Table 5.
Prediction of the mediating relationships of the five disciplines.
Note: n = 687; *p < .05; **p < .01.
For the first mediating relationship as shown in Table 5, organizational culture (β = .051, ns) and leadership (β = .001, ns) failed to reach significance when team learning and systems thinking were included in the equation. This pattern of results indicated full mediation of team learning and systems thinking on the relationship between the antecedents (i.e. organizational culture, leadership) and the outcome (knowledge sharing).
For the second mediating relationship, personal vision (β = .002, ns) and motivation (β = .069, ns) were not significant when personal mastery and shared vision were included in the equation. This showed that personal mastery and shared vision fully mediated the relationship between the antecedents (personal vision, motivation) and the outcomes (strategic planning).
For the third mediating relationship, personal vision (β = .018) became non-significant and the β values of personal values and motivation reduced from .021 to 0.11 and from .624 to .535 respectively when personal mastery and shared vision were included in the equation. This indicated that personal mastery and shared vision partially mediated the relationship between the antecedents (personal vision, personal values, motivation) and work-life balance.
For the fourth mediating relationship, the β values of organizational culture and leadership remained significant, but decreased from .664 to .497, and from .202 to .101 correspondingly, while mental models and systems thinking were significant in the equation. This indicated that mental models and systems thinking partially mediated the relationship between the antecedents (organizational culture, leadership) and strategic planning.
For the final mediating relationship, the β values of organizational culture and leadership decreased from .378 to .280, and from .160 to .090 correspondingly yet remained significant while team learning and systems thinking were significant in the equation. This indicated that team learning and systems thinking partially mediated the relationship between the antecedents (organizational culture, leadership) and work-life balance.
The results showed that the five disciplines either fully or partially mediated the relationships between the antecedents and the outcomes. Thus, Hypothesis 7 was partially supported.
Comparative analysis
Table 6 presents the independent sample t-test results of the comparison between the two universities. The UK university had lower scores in most of the components, as compared to the Vietnamese university. The only exception to this was individual performance, which included administration performance (mean difference = -.372, p = .049), and research performance (mean difference = -.594, p = .003). These results do not lend support to Hypothesis 8, in which we posited that UK values would be higher than Vietnam values.
Comparison between the two universities.
Discussion and conclusions
The current study provides a conceptual framework and an exploratory analysis for Senge’s (1990) LO model within the two organizations (universities), one in the UK and the other in Vietnam. An intriguing finding of our study was that most of the components of the LO model in the Vietnam university had higher score evaluations compared with the UK university. We believe that the source of these differences is based on culture. As discussed in the literature review, the British are generally more individualistic, while the Vietnamese tend to be collectivistic. In Vietnamese organizations, people try to encourage harmony among members and protect the image of their organizations/communities (Thêm, 1999; Vượng, 2001), whereas the university in the UK is located in an individualistic culture, in which people tend to have strong individual autonomy (Hofstede, 1993, 2001), especially in academia (Dearlove, 2002; Woodfield and Kennie, 2008). This behavior may lead to the findings in this study. Another possible explanation can be based on the results of an annual global research called ‘Planet Happy Index’, Vietnam ranks number 5 while the UK ranks 74 out of 143 countries on the list (Abdallah et al., 2009). This indicator suggests that Vietnamese people are more likely to feel satisfied than their British counterparts.
Second, all the suggested antecedents were significantly positively associated with mental models and systems thinking. Specifically, organizational commitment, an open and dynamic organizational culture, and positive leadership are positively associated with mental models; and competence, leadership, and organizational culture are positively associated with systems thinking. These findings confirm conceptual models that have been offered in the literature (Bui and Baruch, 2010a, 2010b; Kofman and Senge, 1993).
Third, personal vision, motivation, and personal values were positively associated with personal mastery. Team commitment, goal setting, organizational culture, and leadership were positively associated with team learning. Personal vision, personal values, motivation, and organizational culture were positively associated with shared vision.
Fourth, development and training was positively associated with personal mastery, but not with team learning, while individual learning was positively associated with team learning, but not with personal mastery. These findings contrast with Prichard et al.’s (2006) argument that development and training enhance team learning, at least in HE. There are a number of reasons for this, including stressful, top-down evaluated, and/or insufficiently taught skill and knowledge training programmes (Elkjaer, 2001). As Dodgson (1993), Senge (1990) and Watkins and Marsick (1993) suggest, individual learning is the primary learning enabling learning organizations.
Fifth, leadership was not positively associated with shared vision. This finding suggests that shared vision in HE must not only be created from the top down, but it should come also from the bottom up and lateral directions. This fits with the findings of Gudz (2004) and Tsai and Beverton (2007), who argue that leaders do not tend to be good stewards in supporting shared vision.
Sixth, team learning, shared vision, systems thinking, and personal mastery tend to have positive impacts on self-efficacy, work-life balance, knowledge sharing, and strategic planning (Bui and Baruch, 2010b). Team learning was positively associated with teaching performance, which is in line with research in HE (Pil and Leana, 2009). This finding seems to be in line with the current increasingly demanding requirements for high quality teaching in HE (Antonacopoulou, 2010).
Seventh, and surprisingly, mental models were negatively associated with research performance, knowledge sharing, and self-efficacy. Given the fact that knowledge sharing has been under-researched, especially at the micro level (Foss et al., 2010), our research provides an insight into this area. It highlights a matter in HE policy, which is particularly relevant to countries that are aiming to improve research performance and knowledge sharing, and which use national evaluation projects for this purpose. For example, the UK Research Exercise Assessment (REA) discourages collaboration in research and knowledge sharing among colleagues within the same institution, by employing a policy whereby co-authors can claim credit for the same paper as long as they represent different institutions. This means that academics may seek information and knowledge outside but may not be willing to seek or share from inside their institution. This may partly explain why mental models are not often shared within universities, and why changing mental models is a real challenge in the education sector (Bell and Harrison, 1998; O’Neil, 1995). There must be a revolution in HE to change the current mental models. This finding also implies that there is a long journey ahead for HE institutions, as they strive to become LOs.
Finally, on one hand, teaching and research performance were obviously the strengths of the UK university, as compared to its counterpart in Vietnam. That explains why the UK is one of the largest providers of HE to Vietnam. On the other hand, the employees in the Vietnam university tend to have a better work-life balance than their British counterparts. This is not unexpected, given previous study that finds that collectivist employees have a better work-life balance than their individualist counterparts (Powell et al., 2009).
Theoretical implications
This study has shed light on the empirical exploration possibilities of Senge’s LO theory. We anticipate that this study will draw more academic attention to the critical theory of LO. Below are some key theoretical implications for this research area. First, our study has contributed to the development of LO theory by providing a strong validation to Senge’s framework of LO, with sets of antecedents and outcomes in a HE context. This study has met the challenge presented by jeopardized wide-scale LO development (Smith and Tosey, 1999) and answered the shortage of empirical investigations of LO interventions (Nair, 2001). We identified three antecedents for mental models and systems thinking; four for personal mastery and shared vision; and five for team learning. These findings once more illustrate the critical role of human resource management in the process of organizational learning (Lopez et al., 2006). Second, systems thinking has a positive impact on the process of knowledge sharing, while personal mastery and mental models appear to improve the process of strategic planning of organizations and the pursuance of work-life balance. Team learning and shared vision can promote organizational knowledge sharing, as well as an individual’s self-efficacy and a better work-life balance. Shared vision also advances organizational strategic planning, while team learning can help academics to teach better. Third, mental models tend to have a negative impact on research performance, self-efficacy, and knowledge sharing in HE. This finding challenges previous studies (see An and Reigeluth, 2005; Davenport et al., 1998; Inkpen, 2000). Therefore, this framework needs testing in other sectors in order to confirm or re-examine the role of mental models in LOs. Fourth, this research has contributed to the under-researched area of knowledge sharing by examining the micro level of knowledge sharing amongst individuals within organizations (Foss et al., 2010). The findings show that team learning, shared vision, and systems thinking have positive impacts on knowledge sharing, whilst mental models have negative impacts on it. Fifth, a dynamic, interactive, proactive, group-oriented, and accommodating culture is an important factor affecting the process of becoming a LO. This shows a fitness of organizational culture with the contemporary labour markets (Baruch and Peiperl, 2000). In contrast, a stable, bureaucratic, reactive, individual oriented, and aggressive culture is more likely to hinder this process. Finally, employees in the Vietnam university being embedded in a collectivist culture appear to be more committed to their organizations than their counterparts in the UK, in an individualist culture. The societal cultures do reflect in organizational culture (House et al., 2004).
Managerial implications
Apart from the theoretical implications mentioned above, this research has a number of relevant findings for HE managers. First, HE recruiters should be aware that employees who come from collectivistic cultures often have a more positive attitude towards their organizations, as well as toward the process of creating LOs in comparison to those who come from individualistic cultures. Conversely, those coming from individualistic cultures tend to perform better, in teaching and research in our study. This is one of the benefits of diversity. The findings related to leadership differences between the two cultures in this study also confirms the importance of top management teams with international diversification (Tihanyi et al., 2000). Second, team learning is not likely to gain as much benefit from development and training as individual learning. This contradicts the major tenet of literature about team learning and development and training (Morgeson et al., 2010; Prichard et al., 2006). However, development and training tend to promote personal mastery. Therefore, HE managers should revisit the aims and targets of development and training if they are committed to becoming LOs. Third, as mentioned clearly in Senge (1990), organizational vision does not only come from the top; it must come from individuals at all levels across the organization. The role of employees in forming a shared vision should be as important as the leader’s role in that task. Creating an open, accommodating, and dynamic organizational culture is likely to promote a shared vision for the organization. Finally, becoming a LO tends to benefit the employees of academic institutions in terms of improving their wellbeing through a better balance between work and life. This can be considered as an additional bonus for institutions striving to become LOs.
Limitations and future research
Like any study, there are limitations to our findings. First, team performance can be one of the outcomes of team learning (Chan et al., 2003; Senge, 1990). However, team performance is not merely an outcome of the internal functioning of teams (Ancona, 1990; Joshi and Roh, 2009). Employees in HE institutions may belong to an internal team(s), an external team(s), both internal and external teams, mixed teams, or to no team at all. This may partially explain why team performance was not a significant outcome of the LO model. Second, organizational performance can be another outcome of systems thinking (Fullan, 2004; Senge, 1990; Wright, 1999). As our study was based on only two universities, it is difficult to generalize organizational performance as one of the outcomes of the LO model. In order to measure organizational performance, more cases are needed. Third, though a large dataset was drawn from large and well-established institutions in the UK and Vietnam, covering only two institutions limits the generalizability of the findings beyond this specific research context. Thus, the findings have to be interpreted with caution. To improve the generalizability of this study, further studies should be conducted in more countries and across more industry sectors. These limitations can be addressed in future research. The newly developed LO model we offer can be applied and tested in other universities and in other sectors. This may lend further support to the validity of our model.
Footnotes
Acknowledgements
We would like to thank Professor Eugene Sadler-Smith, Mark Ridolfo, and the two anonymous reviewers for their helpful feedback and comments.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
