Elsevier

Long Range Planning

Volume 40, Issue 1, February 2007, Pages 30-44
Long Range Planning

Social Capital, Psychological Safety and Learning Behaviours from Failure in Organisations

https://doi.org/10.1016/j.lrp.2006.12.002Get rights and content

This study explores failure-based learning behaviours in organisations. It examines the role of social capital and psychological safety in the development of failure-based learning behaviours. Survey data were collected from 137 members of 33 organisations in a variety of industries in Israel. The results indicate that in organisations where there is strong social capital, the development of psychological safety and failure-based learning behaviours is enabled. It was also found that social capital, through psychological safety, is directly and indirectly associated with failure-based learning behaviours. Follow-up analyses traced the practices organisations use for expanding learning capacities and their outcomes at both the organisational and individual levels. Finally, the implications for learning from failure in organisations are discussed.

Introduction

How much do companies learn from their mistakes? In the introduction to the special issue on organisational failures in Long Range Planning in 2005, Wilkinson and Mellahi, the guest editors, noted that “while most managers know they are at risk of failure, they try to shun the subject, rather than actively seek to guard against it, or face up to it and be prepared to learn from it when it does occur. Managers need to understand two key issues concerning organisational failure. The first relates to understanding the causes and processes of organisational failure so as to help them design effective strategies to avoid or handle failure in the future. The second issue relates to understanding barriers to learning from failure and identifying strategies to overcome them.”1 This paper is concerned with the second issue and examines the importance of social capital and psychological safety for enhancing learning behaviours or processes from failures in organisations.

These failure-based learning behaviours are treated as a unique component of organisational learning. Organisational learning is a complex process through which organisations seek to understand and critique what is taking place.2 Fostering and applying learning behaviours or processes in organisations is critical for survival and viability.3 In recent years, there have been two key developments:

  • (1)

    scholars have increasingly focused on studying intra-organisational learning behaviours, suggesting that learning behaviours vary across work units within the same organisation4; and

  • (2)

    emergent research interest has been devoted to a better understanding of the process of learning from failures.5

However, these two lines of research need further development. Specifically, although investigating learning behaviours at both individual and group levels are valuable, this line of research does not fully capture Cyert and March's argument that individual organisations can learn in ways that are independent of individuals (and by implication, intra-organisational work units or groups).6

Cyert and March, Edmondson and Weick and Sutcliffe indicate that organisations can learn a lot from failures or by encountering problems.7 This led to increasing interest among researchers in learning from failures. Yet “there have been so few studies of learning from failure” that this line of research is still only in an embryonic stage and requires considerable investigation.8 Edmondson recently proposed a comprehensive model of learning behaviours that introduces the antecedents (e.g. leader behaviour, trust) of psychological safety, which itself plays a key role in learning behaviours.9 However, though Edmondson used constructs such as informal dynamics, trust and quality of interpersonal relationships, to the best of my knowledge, no one has explicitly examined the role of social capital, defined as the resources (e.g. knowledge, ideas and opportunities) that flow through a web of internal and external relationships, in enabling the development of psychological safety and failure-based learning behaviours.10 In a context of positive social capital, where people share and exchange resources, they are likely to feel comfortable in expressing themselves, thereby expanding their capacity of learning. This is evident in team learning research that indicates that effective learning is likely to occur when members enjoy high-quality interpersonal relations.11

Yet, how organisations develop, facilitate and expand their capacities of learning as a result of failure remains unanswered. This study attempts to address these issues and contribute to management literature in several ways. I provide a theoretical conceptualisation and operationalisation of this construct. In addition, I present social capital as a key factor in explaining both psychological safety and failure-based learning behaviours. I examine two alternative models:

  • (1)

    a mediation model in which social capital affects failure-based learning behaviours through psychological safety; and

  • (2)

    a moderating model in which social capital interacts with psychological safety to explain failure-based learning behaviours.

Finally, using supplementary in-depth analysis, I investigated the practices organisations use to develop and facilitate these learning processes and expand their capabilities. In so doing, I hope to illuminate how leaders can foster and develop failure-based learning behaviours in their organisations and thus improve organisational processes and outcomes.

Cyert and March proposed that individual organisations learn. Easterby-Smith and Lyles and Vera and Crossan indicated that the central issue in organisational learning research is how (the processes by which) organisations learn.12 Organisational learning is an interdisciplinary field derived from psychology, management science, strategy, production management and cultural anthropology.13 As such, researchers often focus on a particular aspect in defining this construct. In this study, I used the term “learning behaviours” coined by Edmondson to denote learning as a process of change and improvement in organisational actions through better knowledge and understanding.14 This definition is consistent with other views of organisational learning, such as that of Argyris and Schön, who say the term refers to the process of “detection and correction of error”; or of Fiol and Lyles' “… process of improving actions through better knowledge and understanding”; or of Huber who maintains that organisational learning is the processing of information by which the organisation's potential behaviours are changed; or of Vera and Crossan as “the process of change in individual and shared thought and action, which is affected by and embedded in the institutions of the organisation”; and of Argote, Gruenfeld and Naquin as the process by which knowledge is acquired, shared and combined.15 The learning process is iterative in nature and involves “reflective thought” (i.e. critical thinking) in which judgment is suspended, healthy scepticism allowing the rejection of traditional presumptions is maintained and an open mind is exercised.16 These definitions, according to Edmondson and Garvin, entail change and modification of the underlying beliefs, assumptions and actions, with the objective of improving the system (e.g. organisation, group). In this sense, there is some change in the process of learning by understanding and then acting, versus in that of acting, and then interpreting.17

Argyris and Schön claimed that learning in organisations is primarily of a single-loop nature, in that an error is detected and corrected, but the underlying causes are not explored or challenged. This, in spite of accumulating evidence provided by Dewey and by Cyert and March that critical thinking (Dewey), encountering problems (Cyert and March), and preoccupation with failures (Weick and Sutcliffe) are more beneficial than merely detecting and correcting errors or learning from success (Baumard and Starbuck; Edmondson; Sitkin).18 A double-loop learning process occurs when the detection and correction of the error involves the modification of an organisation's underlying norms, policies and objectives (Argyris and Schön).

In this study, I focus on failure-based learning behaviours. These learning behaviours refer to specific processes to help us learn from failures by not merely detecting and correcting errors, but also by challenging and exploring their underlying causes. They are similar to both a double-loop learning process (Argyris and Schön) and second-order problem-solving (Tucker and Edmondson), as both refer to organisations in which members, when facing a problem, not only confront and solve it so the task is completed successfully, but also take action to address the underlying structure, presumptions and causes of that problem.

The literature offers various definitions of social capital. These definitions, according to Adler and Kwon, can be categorised into three groups: external social capital, internal social capital and neutral (external and internal) social capital.19 External social capital usually refers to the direct and indirect relationships an actor or a participant establishes and maintains with other actors or participants in a social network. This could be “friends, colleagues and more general contacts through whom you receive opportunities to use your financial and human capital”.20 Internal or organisational social capital focuses on the relationships among actors (i.e. individuals or groups) within a social collective (organisation, community, nation and so forth). Fukuyama's definition clarifies this view: “the ability of people to work together for common purposes in groups and organisations”.21 A third view is neutral in the sense that theorists define social capital by integrating both the internal (collectivity) and the external (network) elements. To illustrate this view, Nahapiet and Ghoshal define social capital as “the sum of the actual and potential resources embedded within, available through, and derived from the network of relationships possessed by an individual or social unit”.22 In this work, I focus on internal/external social capital as a key enabler of failure-based learning behaviours in organisations.

Research has yet to explore the relationship between social capital and failure-based learning behaviours in organisations. Edmondson's notable study examined the relationship between the quality of interpersonal relationships within a hospital unit and detected error rates. Contrary to the expected outcomes, Edmondson found that in units in which quality relationships were reported, there was a higher rate of detected errors. In these units, it appeared that leaders managed to establish a positive climate (e.g. openness), thereby creating the belief that it was safe to discuss mistakes or errors (i.e. psychological safety). Thus, the number of reported errors in units characterised by good interpersonal relationships was higher than that in units with low-quality interpersonal relationships. This is because members who enjoy good interpersonal relationships share the belief that their unit is a safe environment for interpersonal risk-taking, and that it will not embarrass, reject or punish its members for speaking up. Similarly, Tjosvold et al discovered a positive relationship between learning from mistakes and co-operative goals. In teams in which members work together toward achieving goals (i.e. co-operative goals), there was a high degree of learning from mistakes, compared with teams in which members had lower levels of co-operative goals.

Internal/external social capital is about the web of social relationships work unit members establish among themselves, and between themselves and outsiders (e.g. customers, suppliers).23 Through this web of relationships, they learn, create and acquire useful knowledge. This view draws upon social learning theory, according to which learning is a relational activity involving human interactions.24 Through these interactions among and between actors or participants, better understanding and knowledge are created. Doise and Mugny's experiment, which shows that a conservation task was not solved by two different-aged children because a social relationship had not been formed (the older child did not regard the younger child seriously) proves how important interactions are for learning.25 Drawing on Doise and Mugny's work, Bogenrieder introduced the sociocognitive theory of learning and proposed that a social design is a prerequisite for organisational learning, indicating that it is not merely through cognitive diversity but also through social relationships that learning is enabled and fostered.26 Similarly, Dutton and Dutton and Heaphy argued that the creation of a high-quality connection between two people helps expand learning capacities.27 High-quality connections are vehicles through which useful knowledge is communicated, conveyed and absorbed. Learning processes are made possible through the generative, attentive and flexible interactions people form.

Hence, to draw failure-based learning behaviours, it is essential to bind participants socially. Social capital is an important factor that builds psychological safety and enables the development of failure-based learning behaviours. However, it is expected that psychological safety will have an intervening role in the relationship between social capital and failure-based learning behaviours. As indicated above, the quality of interpersonal relationships is a key factor through which members develop a shared belief that their work unit is a safe environment for interpersonal risk-taking, and that it will not embarrass, reject or punish its members for speaking up. This shared belief of psychological safety is critical for facilitating failure-based learning behaviours. On the basis of this rationale, the following hypotheses are suggested:

Hypothesis 1:

Social capital is positively associated with failure-based learning behaviours (second-order problem-solving behaviours).

Hypothesis 2:

Psychological safety is positively associated with failure-based learning behaviours (second-order problem-solving behaviours).

Hypothesis 3:

Psychological safety mediates the relationship between social capital and failure-based learning behaviours (second-order problem-solving behaviours).

An examination of an alternative model is also noteworthy. This model stipulates that the relationship between social capital and failure-based learning behaviours depends on the degree to which members feel safe to take interpersonal risks, speak up, have honest conversations and discuss errors and mistakes without risking embarrassment, rejection or punishment (Edmondson). Hence, I suggest the following hypothesis:

Hypothesis 4:

There will be significant interaction between social capital and psychological safety in predicting failure-based learning behaviours (second-order problem-solving behaviours), such that the positive relationship between social capital and failure-based learning behaviours will be stronger when psychological safety is high.

I conducted this research among 33 organisations in both the industrial and public sectors. These organisations operate in a variety of industries (a list of the organisations and their industries is in Appendix A). In an attempt to increase the ability of the study to be generalised, I chose to create a diverse sample population. I therefore restricted myself to the organisations to which I had access, through two research assistants who previously knew of these organisations through first-hand acquaintances. All the organisations I contacted agreed to participate. As I was interested in evaluating social capital, psychological safety and failure-based learning behaviours at the organisational level, I sought to include members holding different managerial and non-managerial positions within the organisation to complete a structured survey.

I was interested in a representative population from within each organisation and therefore randomly selected four to seven respondents per organisation and asked them to complete a questionnaire. For each organisation, I was able to obtain surveys from at least one member in a senior executive position and another in a mid-level position (often heading a division or department). Other participants per organisation held low-level managerial or non-managerial roles. The participants were guaranteed confidentiality. The questionnaires were given and collected on site by two research assistants. Of the 170 surveys administered, 137 respondents working in 33 organisations completed the questionnaire, yielding a response rate of 80.58 per cent. Fourteen organisations are from the industrial sector and 19 from the service sector. The average organisational age and size was 30.89 years (s.d. 24.01) and 1,131.55 (s.d. 1,854.26) employees, respectively. Sixty (43.8 per cent) of the individual respondents were women. The average age and tenure of the individual respondents was 33.58 years and 6.15 years respectively.

After collecting data via questionnaires, I sought to enrich our knowledge of the social capital-related practices that organisations use to expand their capacities of learning. I carried out an in-depth analysis in three firms by interviewing their senior executives, middle-level managers and other organisational members who did not hold a managerial position. Overall, we interviewed seven members of Alpha, a knowledge company operating in the software industry; seven members of Gama, an agricultural company; and five members of Delta, a medical equipment company. These open interviews were conducted on site.

In this study, the organisation was considered to be the unit of analysis. Hence, all measures were specified on the organisation level. Respondents were asked to evaluate the social capital, psychological safety and learning behaviours of the organisation as a whole.

Social capital. This measure was assessed by the five-item scale developed by Subramaniam and Youndt.28 This index captures the core ideas of social structure, as presented by Burt, and the more specific ideas on knowledge management, found in the literature.29 This measure assesses the extent to which members in an organisation collaborate to diagnose and solve problems, share information, exchange ideas, interact with customers, suppliers and partners to develop solutions and apply knowledge across the organisation. Sample items are: “Our employees are skilled at collaborating with each other to diagnose and solve problems” and “Our employees interact with customers, suppliers, partners, etc to develop solutions”. Items were all anchored on seven-point scales ranging from 1 = strongly disagree to 7 = strongly agree. The Cronbach's alpha for this measure was 0.82, slightly higher than the reliability (alpha = 0.71) that Subramaniam and Youndt reported in their work.

Psychological safety. To gauge an organisation's shared beliefs, with regard to the extent to which its members feel psychologically safe in taking interpersonal risks, speaking openly and discussing failures, I adapted Edmondson's seven-item team psychological safety scale. To assess psychological safety at the organisation level, I replaced the word “team”, as originally used by Edmondson, with the word “organisation”. In doing so, I preserved the theoretical meaning of the assessed construct. Sample items are: “If you make a mistake in this organisation, it is often held against you (reverse scored item)”, “It is safe to take a risk in this organisation”, and “No one in this organisation would deliberately act in a way that would undermine my efforts”. Items were all anchored on a five-point scale ranging from 1 = strongly disagree to 5 = strongly agree. The Cronbach's alpha for this measure was 0.82, as in Edmondson's study.

Failure-based learning behaviours. To specifically measure failure-based learning behaviours, I drew upon Tucker and Edmondson's study to capture failure-based learning behaviours of a second-order problem-solving type (second-order problem-solving refers to organisations in which members, when faced with a problem, not only solve it so the task is successfully completed, but also work to address the problem's underlying causes). In addition, I drew upon a wide body of literature suggesting that failure-based learning behaviours entail constant questioning and reflecting on the underlying assumptions that inform the actions. To this end, I constructed a seven-item scale. Five items were newly constructed and two were borrowed from Edmondson's study (see Table 1). Sample items are: “When an employee makes a mistake, her co-members in the workplace talk to her not with the intention of blaming her, but rather for the value of learning” and “When employees make a mistake, they inform the relevant manager to enable others to learn from it”.

All items were subject to an exploratory factor analysis. Results of this procedure yielded a one-factor solution, consisting of seven items, with an eigenvalue of 4.34 and explained 62.10 of the variance. The results of this factor analysis and the item description are shown in Table 1. The Cronbach's alpha for this measure was 0.89.

Control variables. Organisational age and size as well as sector were used as control variables. I measured size using a natural logarithmic transformation of an organisation's number of full-time employees. Age was assessed by the number of years since the organisation was established. Sector was divided by two people who were familiar with the organisations into two types: industrial (= 1) and service.

A one-way analysis of variance was performed on the following variables: social capital, psychological safety and failure-based learning behaviours. Results show that there is a greater variability in the ratings between organisations than within organisations (p < .05). I also calculated intra-class correlations (ICCs) for assessing agreement of organisational members. ICC(1) indicates the extent of agreement among ratings from members of the same group. ICC(2) indicates whether groups can be differentiated on the variables of interest. The ICC(1) and ICC(2) values were, respectively, 0.49 and 0.83 for social capital, 0.34 and 0.67 for psychological safety, and 0.47 and 0.82 for failure-based learning behaviours. These values exceed conventional standards for aggregating individual questionnaire responses to team level in field research and, thus, support an aggregation to an organisation level.30

Table 2 presents the means, standard deviations and correlations found among the research variables. The failure-based learning behaviours factor was positively and significantly related to both neutral (internal/external) social capital (r = .71, p < .001) and psychological safety (r = .61, p < .001), providing preliminary support for both Hypotheses 1 and 2. A positive and significant relationship was found between social capital and psychological safety (r = .39, p < .05). None of the demographic variables and sector differences had significant relationships with social capital, psychological safety or failure-based learning behaviours.

Although only Hypothesis 3 was concerned with mediation, each of the three hypotheses was efficiently tested using the procedures for testing mediation, outlined by Baron and Kenny.31 Model 1 in Table 3 presents the results of the first regression, in which failure-based learning behaviours was regressed on social capital and the control variables. Consistent with Hypothesis 1, there was a positive relationship between social capital and failure-based learning behaviours (β = .82, p < .001). The test for Model 2 showed a significant relationship between social capital and psychological safety (β = .74, p < .001). This supports Hypothesis 2. Model 3, which regressed failure-based learning behaviours on both social capital and psychological safety, only partially supports the mediation hypothesis (Hypothesis 3). The effect of social capital remained significant, (p < .01), as did the effect of psychological safety (p < .01). This indicates direct and indirect effects (through psychological safety) of social capital on failure-based learning behaviours. Figure 1 presents the results of the mediation model.

The social capital and psychological safety variables were mean-centred as is suggested for variables that are to be constituents of product terms.32 To test Hypothesis 4, I entered the control variables of organisational age, size and sector environment (see Table 4). Social capital and psychological safety were then entered in the second stage. In the third stage, the product term of social capital × psychological safety was entered. Whereas the main effects (social capital and psychological safety) remained positive and significant, the interactions between them were not significantly related to failure-based learning behaviours, and thus Hypothesis 4 was not supported.

Matters do not always proceed in the desired manner. Actually, managers often face situations in which things drift into the wrong path. The problem is not so much with failures, as they are not entirely avoidable, but rather with the lack of attention, the disregard and even the concealing of mistakes that occur in organisations. While this is the case in many organisations, there are managers who direct their organisation to another path – involvement in understanding failures and learning from them. Corporate governance failures are prominent examples where trust and reputation have been destroyed and serve as critical points for learning processes aiming to improve governance structure and context.33 Thus, two major managerial questions emerge. First, why are some organisations able to design structure and content that enable failure-based learning behaviours and others not. Second, how have those “learning from failures” organisations designed enabling structures and contexts of learning?

Considering that there are many organisational failures, from which organisations can learn and improve their reliability and crisis-preparedness, the study of why and how organisations learn, or fail to learn from failures, is critical for organisational viability. The goal of this study was to illuminate further the importance of learning from failures in organisations. Drawing on a limited yet growing body of research, it is argued that failure-based learning behaviours constitute a special type of organisational learning processes that can facilitate better responses and adaptation to changes. In the words of Weick and Sutcliffe: “an organisation that is ignorant about failure, its location, genesis, and trajectory, is less mindful than it could be” (p101).

In this study, I found that social capital is a key factor in the development of both psychological safety and failure-based learning behaviours. I conceptualised and operationalised social capital in the context of gaining better knowledge and understanding. Social capital refers to a web of quality relationships and interactions among members within a social collectivity (internal social capital) and between members of a collectivity with other participants in a network (external social capital) aimed at acquiring and creating useful knowledge. In organisations where strong (positive) social capital exists, members reported that they felt safe to speak openly and discuss errors and failures without being threatened by the possibility of punishment or embarrassment, and that there was a high rate of failure-based learning behaviours. The results of this study show that the integration of both internal and external social capital is a key mechanism through which psychological safety and learning behaviours are enabled.

An important stream of research led by Edmondson emphasises the need to create a context in which people feel safe to discuss failures. The findings of this study confirm Edmondson's notion and show that when people feel psychologically safe, learning from failures is enabled. In response to the task of “helping Unilever act more intelligently” (i.e. learn faster and leverage what it knows), David H. Smith, head of knowledge development for Unilever, realised that the “solution to our [Unilever] problem wasn't to work harder. We've got to learn to work smarter …”. This notion is true but a bit trite and easier said than done. How can we build context that enables us to learn to work smarter? According to Smith, organisations make the same mistake twice. There are many cases of repeated mistakes in organisations. Why does it happen? Smith says that “fear is the No. 1 reason: fear of being embarrassed, chewed out or worse. Many people and companies are so busy trying to hide boners (from the boss, from stock analysts, from customers and competitors) that they tuck away the learning along with the evidence”.34

Another example of how psychological safety makes is possible to learn from failures is Alpha, a knowledge company competing in Israel's software industry. To foster learning processes, in general, and learning from failures, in particular, the top management team had recently decided to develop a new knowledge management mechanism through which members can find detailed descriptions of past successful transactions as well as failed transactions with customers (e.g. competitors' services were preferred over the firm's services; customer dissatisfaction). Interviews with seven members of these firms reveal a climate of openness that allows individuals to speak up without fear of being embarrassed or risking their job and status. In addition, members indicated that they shared a sense of a family atmosphere, in which they are able to express themselves, discuss things freely and feel that they were being listened to.

Learning behaviours are conceived as a key element in the development of valuable knowledge, which is based on an iterative process of inquiry and reflection. In this sense, reflective thinking is embedded in failure-based learning behaviours because both entail an “active, persistent and careful consideration of any belief or supposed form of knowledge in light of the grounds that support it and the further conclusions to which it tends [that] includes a conscious and voluntary effort to establish belief upon a firm basis of evidence and rationality” (Dewey, p9). The process produced by reflective thinking is an important element of failure-based learning behaviours as it enables organisations to better understand and cope with evolving problems and challenges. Because reflective thinking and learning processes in organisations involve interactions among and between members, strong social capital becomes a valuable asset that organisations would do well to develop and maintain. From a research perspective, however, we should invest substantial efforts in exploring and understanding the social structure and context that facilitate or impede organisational learning behaviours.

If failure-based learning behaviours are beneficial, how can organisations develop and facilitate such learning processes? This supplementary qualitative research identified various practices, defined as organisational routines and manners of doing things, which organisations use to expand their capacities of learning. Table 5 presents two forms of social capital: internal social capital (i.e. relations among individuals or groups within the organisation) and external social capital (i.e. relationships of the organisation with its stakeholders) and indicates practices for expanding capacities of learning, in addition to both organisational and individual outcomes. I discuss these practices and outcomes in regard to each form of social capital.

We identified three key practices organisations use to expand learning through their internal web of relationships. First, they engage in studying organisational failures and near-related failures. At the Gama agricultural firm, an independent consultant helped the managers institutionalise monthly case studies of failures within the firm and those of other firms in the industry. This practice helped the organisation develop high sensibility (i.e. ability to recognise market changes) and reliability (i.e. ability to constantly correct imperfect processes). It also created a climate in which failures were discussed openly and employees felt comfortable bringing up and discussing imperfections. Second, at Alpha (a knowledge company competing in the software industry), I found a flexible work structure that enabled the top management to mobilise about 20 per cent of the employees from one task to another in just one to three days. This practice of mobilising employees, enabled by flexible organisational structure and context, produced a highly responsive firm. This firm was able to meet extraordinary customer requests better than any other in its industry. Customers realised that when they needed speedy and high-quality response, Alpha was the ultimate service provider. This practice also helped the firm produce strong human capital, as employees often tackled a wide variety of tasks. Employees also felt more motivated and satisfied when involved in task variety and new challenges.

A third key practice is weekly meetings of top management teams. Through these meetings, members exchanged ideas about market conditions and trends, organisational context and capabilities and things that went wrong as well as tasks that were not met. The ability to maintain weekly meetings, even in those companies in which top management team members are often abroad, is not a trivial matter: it requires commitment and an appreciation of the importance of these meetings. At Delta, a medical device company, I found that these meetings helped increase awareness to external and internal dynamics and changes, and as a result of the exchange of information and knowledge, new knowledge was created. Through this heightened awareness and knowledge management, the company, which engaged in highly risky technological projects, was able to avoid major breakdowns even in times of significant budget constraints. These meetings also gave members a sense that they were valued and thus enhanced their organisation-based self esteem.

I identified one key practice organisations use to expand learning through their external web of relationships: mentoring and teaching. I discovered that in several firms, senior managers not only mentor their employees but also each other. For example, the CEO of Alpha regularly takes top management team members to meetings with customers and competitors and trains them for their own future interactions with these and other potential stakeholders. At Gama, when senior members visit customers abroad they are often accompanied by members from other departments to strengthen the firm's relationships with and deepen its understanding of its customers. In this particular firm, most customers are based in South America and South-East Asia and such mentoring patterns entailed institutionalised routine. Through this mentoring pattern, the firms built strong human capital and increased knowledge-sharing between people within the organisation. The people involved in this practice of mentoring often reflected on the difficulties and errors that had been observed. This proved very useful in terms of refraining from oversimplifying problems and avoiding repeated mistakes. On the individual level, people gained new knowledge, often vital for their career development. They felt the companies perceived their value and trusted their competence. Ultimately through this practice of investing in mentoring, people felt relatively strong normative attachment (feelings of obligation to remain with an organisation) as well as affective attachment (feelings of emotional commitment to remain with an organisation).

One could question the merit of learning from previous failures, by claiming that they are contextual and idiosyncratic in nature. In response, research suggests that managers often tend to disregard failures owing to their idiosyncratic nature. This prevents the vital process of becoming mindful through preoccupation with failures from taking place. In this respect, one avenue of research is examining the personality, attributes and attitudes of managers and their effect on the organisation's tendency to learn about and understand failures and their actual involvement in the process.

This study primarily aimed to show the extent to which failure-based learning behaviours may be enabled through the development of strong social capital and a shared belief of psychological safety. Though the notion of failure-based learning behaviours has recently received increased attention and is compelling, one should realise that learning can take many forms and learning from success is also obviously important. This study, however, attempted to shed light on the merit of failure-based learning behaviours in organisations. It may well be that a balanced approach would prove fruitful, but before we turn toward this approach, there is a need to explore further the substance and essence of failure-based learning behaviours. Additionally, this study left many questions unanswered. For example, we know little about the dynamics through which failure-based learning behaviours evolve; indeed, social relationships are one important mechanism, but its nature is often left understudied. We also need to examine more systematically the implications of failure-based learning behaviours on organisational outcomes (e.g. to what extent and under which conditions do failure-based learning behaviours enhance organisational performance?).

Although a theoretical rationale and reasoning was developed for the research model, one should avoid making causal inferences from a cross-sectional study. In addition, I attempted to assess failure-based learning behaviours in organisations, and thus used multiple respondents in each organisation. Though this approach is more reliable than a single informant, there is the potential of response bias results from self-report data.35 An evaluation of this problem using Harman's one-factor test revealed that when all items were subjected to factor analysis, a one-dominant factor did not emerge, suggesting that although a potential response bias exists, it is not severe.36 One possible solution is to use independent observers to assess more objectively failure-based learning behaviours.

This study emphasised the importance of learning behaviours or processes from failure in work organisations. I found that social capital is a key enabler of psychological safety and organisational learning from failures. Both survey data and case study analyses support this claim. Through in-depth analysis, I also observed the key practices organisations use to expand their learning capacity. These social capital-related practices proved very useful and resulted in enhanced organisational sensitivity and reliability, high organisational responsiveness to customers' extraordinary needs, a high level of awareness to external and internal cues, strong multiple skilled and highly trained human capital, and both knowledge sharing and creation.

Section snippets

Acknowledgements

I wish to thank students who helped me with data collection and Sharon Erez and Rachel Schlosses for their editorial work. I thank the Editor, Professor Charles Baden-Fuller, and the anonymous reviewers of this journal for their insightful comments and suggestions.

Abraham Carmeli received his Ph.D. from the University of Haifa. He is now a faculty member at the Graduate School of Business Administration at Bar-Ilan University. His current research interests include complementarities of intangible resources, top management teams, organisational prestige and image, crisis management, learning processes and individual behaviours. His research has appeared in journals such as Human Resource Management, The Leadership Quarterly, Managerial and Decision

References (36)

  • A. Wilkinson et al.

    Organizational Failure: Introduction to the Special Issue

    Long Range Planning

    (2005)
  • P. Baumard et al.

    Learning from failures: Why it may not happen

    Long Range Planning

    (2005)
  • A.C. Edmondson

    The local and variegated nature of learning in organizations: A group-level perspective

    Organization Science

    (2002)
    C.M. Fiol et al.

    Organizational learning

    Academy of Management Review

    (1985)
    D.A. Garvin

    Learning in Action. A guide to putting the learning organization to work

    (2000)
  • M. Easterby-Smith et al.

    Introduction: watersheds of organizational learning and knowledge management

    E.W.K. Tsang

    Organizational learning and the learning organization: A dichotomy between descriptive and prescriptive research

    Human Relations

    (1997)
  • Ji-Yub Kim et al.

    Vicarious learning from the failure and near-failure of others: Evidence from the US commercial banking industry

    Academy of Management Journal

    (2006)
  • A.C. Edmondson

    Psychological safety and learning behavior in work teams

    Administrative Science Quarterly

    (1999)
    A.C. Edmondson

    Learning from mistakes is easier said than done: Group and organization influences on the detection and correction of human error

    Journal of Applied Behavioral Science

    (2004)
  • M.D. Cannon et al.

    Confronting failure: Antecedents and consequences of shared beliefs about failure in organizational work groups

    Journal of Organizational Behavior

    (2001)
    M.D. Cannon et al.

    Failing to learn and learning to fail (Intelligently): How great organizations put failure to work to improve and innovate

    Long Range Planning

    (2005)
    A.C. Edmondson

    Learning from failure in health care: Frequent opportunities, pervasive barriers

    Quality and Safety in Health Care

    (2004)
    D. Tjosvold et al.

    Team learning from mistakes: The contribution of cooperative goals and problem-solving

    Journal of Management Studies

    (2004)
    A.L. Tucker et al.

    Why Hospitals Don't Learn from Failures: Organizational and Psychological Dynamics that Inhibit System Change

    California Management Review

    (2003)
  • R.M. Cyert et al.

    A Behavioral Theory of the Firm

    (1963)
  • K.E. Weick et al.

    Managing the Unexpected: Assuring High Performance in an Age of Complexity

    (2001)
  • A.C. Edmondson

    Psychological safety, trust, and learning in organizations: A group-level lens

  • J. Coleman

    Social capital in the creation of human capital

    American Journal of Sociology

    (1998)
    W. Baker

    Achieving success through social capital

    (2000)
  • C.P. Earley et al.

    Multinational teams: A new perspective

    (2002)
  • D. Vera and M. Crossan, Organizational learning and knowledge management: Toward an integrative framework, in...
  • M. Easterby-Smith

    Disciplines of organizational learning: Contributions and critiques

    Human Relations

    (1997)
  • C. Argyris et al.

    Organizational learning: A theory of action perspective

    (1978)
    G.P. Huber Organizational learning

    The contributing processes and the literatures

    Organization Science

    (1991)
    L. Argote et al.

    Group learning in organizations

  • J. Dewey

    How We Think: A restatement of the relation of reflective thinking to the educative process

    (1933/1986)
  • M.M. Crossan et al.

    Organizational Learning: Dimensions for a Theory

    International Journal of Organizational Analysis

    (1995)
  • S.B. Sitkin

    Learning through failure: The strategy of small losses

    A.L. Tucker and A.C. Edmondson...
  • Cited by (0)

    Abraham Carmeli received his Ph.D. from the University of Haifa. He is now a faculty member at the Graduate School of Business Administration at Bar-Ilan University. His current research interests include complementarities of intangible resources, top management teams, organisational prestige and image, crisis management, learning processes and individual behaviours. His research has appeared in journals such as Human Resource Management, The Leadership Quarterly, Managerial and Decision Economics, Organization Studies, Public Administration and Strategic Management Journal, among others. e-mail: [email protected]

    View full text