Reflective learning: a new leadership development framework driving engineering innovation

ABSTRACT

Innovation is an organizational learning process that demands a reflexive perspective to take on uncertainties and question deeply held assumptions, propelling leadership and organizational structures forward. Discussions on the relationships between innovation and reflexivity predominantly focus on individuals rather than considering it a shared capability within an organization. This study elucidates the impact of reflexive-learning-based leadership development on promoting collective reflexivity, navigating its implications for organizational leadership and structures in innovation. A reflexive-learning-based leadership development program is a set of collective actions comprising four fundamental steps: acknowledging problems, reassessing assumptions, thinking of alternatives, and developing new perspectives. This study demonstrates its effectiveness in developing collective reflexivity within an engineering organization by repurposing the After-Action Review (AAR) as reflexive learning training. The findings underscore reflexive learning, making leaders exhibit the traits of ambiguity tolerance and interdisciplinary knowledge that favor innovation. This research contributes to an academic framework, opening avenues for future investigations into organizational development from a reflexive perspective and providing a proven managerial practice to improvise an effective solution to overcome those uncertainties in real-world innovations.

KEYWORDS:

• Leadership Development
• organizational learning
• innovation
• reflexivity
• organizational structure

Introduction

Innovation is inherently ambivalent due to a lack of information for decision-making, necessitating a learning process to navigate these uncertainties (March, 1991; Mumford et al., 2002), requiring a learning process to cope with them (Argote et al., 1989; Cheng & Van de Ven, 1996; Lukoschek et al., 2018; March, 1991). Reflexivity, a capability to question what is taken for granted, is fundamental to this process (Cunliffe, 2016; De Dreu, 2007; Schippers et al., 2007, 2017; Suddaby et al., 2016), making some see opportunity where others see danger or failure (Suddaby et al., 2016).

While scholars have predominantly conceptualized reflexivity as an aspect of individual cognitive development (Carmeli et al., 2014; Cunliffe, 2016; Farson & Keyes, 2002; MacCurtain et al., 2010; Schippers et al., 2007, 2017), its role has gained prominence in understanding engineering organizations, where holds innovation a more pronounced role compared to other industries (Carmeli et al., 2014; Farson & Keyes, 2002; Hirst et al., 2004; MacCurtain et al., 2010; Schippers et al., 2007, 2017). However, these discussions primarily focus on observing individuals’ reflexive traits and their relationships with innovation rather than considering reflexivity as a shared capability within an organization. This study expands beyond individual self-reflexivity, delving into how an engineering organization can initiate collective reflexivity, characterized as its shared capability, turning into collective actions to challenge unquestioned assumptions in social interactions and change organizational innovation structures.

Reflexivity is not an isolated concept but rather intricately intertwined with social contexts and deeply embedded within institutions (Archer, 2007; Bourdieu, 1990; Bourdieu & Wacquant, 1992). It plays a crucial role in creating opportunities for social transformation (Archer, 2007, pp. 25–61; Garfinkel, 1967; Giddens, 1984). This study conceptualizes a comprehensive framework to address these inquiries: How can an engineering organization foster collective reflexivity in innovation? How does collective reflexivity impact organizational structures and leadership for innovation?

Discussions on engineering leadership and organizational development have also gained significant traction (Athreya & Kalkhoff, 2010; Bayless & Robe, 2010; Crumpton-Young et al., 2010; Daniels, 2009; Farr & Brazil, 2010). However, the investigation of the relationship between collective reflexivity and leadership development remained tangled. This study initiated a reflexive-learning-based leadership development program as an intervention to an engineering organization and measured the changes in organizational structures and leadership. This approach contributes to a broader exploration of collective reflexivity and its correlations with various aspects.

Literature review

Collective reflexivity

Individuals are not passive recipients of social norms and rules (Garfinkel, 1967). Instead, individuals situated in the social interactions that actors learn how others view us, test our perspectives on others, check us with their impressions of us, and actively construct and interpret social reality (Goffman, 1959, pp. 248–250; Mezirow, 1991, p. 185). An individual’s reflexivity allows individuals to project and accomplish their actions (Archer, 2007; Caetano, 2015), making transformation impossible (Archer, 2007; Bourdieu, 1990; Garfinkel, 1967; Giddens, 1984). On the contrary, collective reflexivity, a shared capability to project actions and accomplish them within an organization, involves developing new perspectives that tell a story to legitimize collective actions (Aldrich & Fiol, 1994; Lounsbury & Glynn, 2001; Lounsbury et al., 2019).

A reflexive-learning-based leadership development program driving collective reflexivity comprises four steps: acknowledging problems, reassessing assumptions, considering alternatives, and developing new perspectives (Argyris, 1990; Argyris & Schön, 1974; Mezirow, 1991; Sitkin, 1992). The first step is a situational analysis and framing problems (Farson & Keyes, 2002; Heifetz & Linsky, 2017, Schippers et al., 2008, 2017), which involves scanning ‘what’s going on’ in social contexts (Heifetz & Linsky, 2017). Leaders learn to discern the gap between the goal and the actual in reviewing their leadership at this step.

However, leaders often exhibit a tendency to overlook evidence that challenges their entrenched assumptions, thereby perpetuating distortions in their understanding (Janis, 1982, pp. 174–176). This behavior tends to solidify biases, affirming pre-existing beliefs (Dutton & Jackson, 1987). Reassessing assumptions, positively linked to a leader’s openness to change (Argyris & Schön, 1974), is the second step of reflexive learning that helps counter the biases and beliefs that leaders espoused (Argyris & Schön, 1974, pp. 76–79). This step necessitates the regular reevaluation of critical assumptions by both individuals and organizations (Argyris & Schön, 1974, pp. 71–80). To illustrate, leaders can frequently ask, ‘What if I am wrong’, to challenge themselves.

Thinking of alternative perspectives (Mezirow, 1991, p. 161) is the third step, enabling individuals and organizations to reframe problems and solutions using contrasting narratives. For instance, leaders can underscore the unbearable pains and urgent need for change (Davis, 2002; Smith, 1980). This approach can unleash the strengths to change and subvert inappropriate dominance (Davis, 2002; Ewick & Silbey, 1995). These contrasting narratives need not promise specific outcomes but present the potential for reshaping the future (Davis, 2002; Ewick & Silbey, 1995).

Finally, developing new perspectives involves turning alternatives into actionable recommendations and executable conclusions (Baird et al., 1999; Garvin, 2003; Mahal, 2018) and transforming an organization’s old meaning schemes into new ones (Meyer & Höllerer, 2010; Mezirow, 1991). Empirical studies found that an open, safe, and supportive environment is essential for eliciting new understandings (Garvin, 2003), leading engineers to better performance (Hirst et al., 2004). This study embeds these four steps of reflexive learning into a leadership development program, initiating collective reflexivity for collective actions within an engineering organization.

Innovation leadership

Explaining the emergence and functionality of collective reflexivity within a leadership development program necessitates defining the measurement of leadership. This study characterizes innovation leadership as the learning ability to mobilize followers for innovative activities. First, leadership is the ability to examine possibilities, search for information and anomalies, seek out a conceptual model, understand the problem situation (Mumford et al., 2002), identify performance gaps, orient followers’ attitudes to innovations, and make decisions under uncertainty (Duncan, 1976). Second, innovation is a learning process (Argote et al., 1989; Cheng & Van de Ven, 1996; Lukoschek et al., 2018; March, 1991) to cope with various uncertainties appearing at initiation and implementation stages, which necessitate different organizational structures, strategies, and leadership (Bledow et al., 2009; Chandy & Tellis, 2000; Cheng & Van de Ven, 1996; Christensen, 2016; Dougherty & Hardy, 1996; Drucker, 1993; Duncan, 1976; Furr & Dyer, 2014; March, 1991; O’Reilly & Tushman, 1997, 2004, 2016). In essence, from a reflexive perspective, innovation leadership demonstrates the reflexive capability to project and accomplish their actions within an organization.

Innovation uncertainties

Innovation is fraught with uncertainty, stemming from insufficient information for decision-making (March, 1991; Mumford et al., 2002). Some of those uncertainties pertain to technical works, but others are adaptive (Heifetz, 1994; Heifetz et al., 2009). Technical works usually refer to those jobs with clear definitions and solutions to address imminent problems, whereas adaptive ones revolve around smoldering issues with unclear definitions and solutions to overcome (Heifetz & Linsky, 2017; Heifetz et al., 2009).

At the initial innovation stage, businesses primarily explore adaptive works like market identification and prototyping, alongside technical uncertainties like product delivery. As businesses mature, the innovators contend with uncertainties for sustainability, such as scalability and growth slowdown (Christensen, 1992). At the sustaining stage, those works are predominantly technical, such as integrating innovations with existing businesses and transforming growth strategies to overcome business slowdown. Various stages require adaptable organizational structures (Christensen, 1992; Duncan, 1976; O’Reilly & Tushman, 2016). To effectively measure innovation leadership, aligning different types of innovation leadership with their responses to those uncertainties is essential. Figure 1 indicates the matrix of innovation uncertainties intertwined with two categories of works and various business stages.

Figure 1. Types of innovation uncertainties.

Ambiguity tolerance

Innovation leadership encapsulates diverse individual propensities in ambiguity, domain knowledge, learning styles, and values at different stages of innovation. Ambiguity denotes situations with complexity, novelty, or contradiction (Budner, 1962). The capacity to tolerate ambiguity is a crucial metric for measuring leadership response to uncertainties in innovation (Altinay et al., 2012; Argote et al., 1989; Budner, 1962; Furnham & Marks, 2013; Schere, 1982). Empirical studies consistently showed that ambiguity tolerance was positively associated with risk-taking in innovation (Altinay et al., 2012; Furnham & Marks, 2013) and promulgated conceptual creativity and professional solutions in uncertain settings (Argote et al., 1989).

A high ambiguity tolerance corresponds to exploring disjunctive knowledge and objectives (Budner, 1962). In the face of uncertainties, exploratory leaders illustrated a greater ambiguity tolerance than sustaining managers in novel and complex situations (Schere, 1982). Conversely, sustaining leaders are likely to display a conservative stance as ambiguity poses a threat (Argote et al., 1989; Budner, 1962), challenging their indulgence for certainty (Hedden & Gabrieli, 2006).

Knowledge domain

The knowledge domain is also essential in measuring leadership traits dealing with innovation uncertainties (Bledow et al., 2009; Dyer et al., 2019; Heng et al., 1994; Hirst et al., 2004; Schoemaker et al., 2018). Exploratory businesses necessitate interdisciplinary knowledge to cope with markets, prototyping, and product delivery (Dyer et al., 2019; Heng et al., 1994; Schoemaker et al., 2018). Conversely, sustaining businesses requires leaders well-versed in specific domain knowledge for planning, budgeting, organizing, and staffing (Kotter, 1999, 2012).

Research findings highlight that exploratory leaders tend to engage in extensive learning compared to their experienced counterparts, inclined toward sustaining roles (Hirst et al., 2004). Bledow et al. (2009) further observed that exploratory leaders are keen to explore unfamiliar knowledge domains; on the contrary, sustaining leaders are prone to the familiar knowledge domain. Particularly in roles such as R&D leadership, individuals must continuously acquire new knowledge from their recent experiences and ongoing engagements, employing this knowledge effectively while managing teams involved in innovative projects (Hirst et al., 2004).

Learning styles

Leadership is significantly connected to an executive’s learning style (Armstrong & Mahmud, 2017; Lukoschek et al., 2018), which includes concrete-experience comprehension and abstract-experiment apprehension in learning style (Kolb, 2015, p. 51), has a substantial relationship with leadership and collective reflexivity (Day et al., 2014; Hirst et al., 2004). Skilled intuition, a form of concrete-experience comprehension, proves its efficiency in decision-making by scanning an individual’s experience to find a representation responding to the situation, but it is likely to lead to biases (Kahneman & Klein, 2009). Armstrong and Mahmud (2017) found that senior leadership often relied on intuition or tacit knowledge to solve technical problems and tended to be the learning style of accommodation. Sustaining leadership usually obtains visible outputs from learning concrete experience, observation, training, feedback, following standards, and listening to instructions (O’Reilly & Tushman, 2016, p. 15). In contrast, exploratory leadership is comfortable pursuing new ideas or uncertain outcomes from abstract conceptualization, exploration, and experimentation (Christensen, 2016; Dyer et al., 2019; Kolb, 2015; McCall, 1998).

Values

Several studies have highlighted the significance of value in differentiating innovation leadership traits (Eisenbeiss & Boerner, 2010; Garcıa-Morales et al., 2008; Khan & Khan, 2019; Lei et al., 2021; Sosik, 1997; Zheng et al., 2017). Transformational leadership theory induces a dichotomy between transactional and transformational leadership akin to D. M. McGregor’s (1957), distinguishing reward-oriented from self-fulfillment-centric management. Transformational leaders advocate Theory Y and norm change, behaving as role models, stimulating followers to question assumptions, coaching followers to pursue self-fulfillment, and making followers feel more able to influence their work (Pieterse et al., 2010). On the contrary, transactional leaders embrace Theory X and current norms, rewarding followers for performance and correcting deviances (Bass & Riggio, 2014, pp. 5–12; Burns, 2010 & P. 425–426;, p. 36; Heifetz & Linsky, 2017).

Several studies suggested that transformational leadership positively correlated with innovation performance (Garcıa-Morales et al., 2008; Khan & Khan, 2019; Lei et al., 2021; Zheng et al., 2017) and could generate more original solutions (Sosik, 1997). However, Eisenbeiss and Boerner (2010) discovered that transformational leadership does not consistently relate to innovation but postulated a U-shape model after facilitating a survey to investigate 52 R&D teams. Therefore, transformational values are advantageous in exploratory or late-stage businesses, while transactional values are beneficial for sustaining enterprises. Table 1 summarizes the traits of exploratory and sustaining leadership in ambiguity tolerance, knowledge domain, learning style, and value.

Table 1. The innovation leadership.

Types of the Innovator’s Leadership	Exploratory Leadership	Sustaining Leadership
Ambiguity Tolerance	High tolerance in novel and complex ambiguity Mistake tolerant	Low tolerance in novel and complex ambiguity Mistake intolerant
Knowledge Domain	Multiple-domain, disjunctive, or divergent knowledge	Single domain, familiar, convergent knowledge
Learning Style	Abstract experiment-oriented comprehension	Concreteexperience-oriented apprehension with standards or rules
Values	Inspiration-driven or advocate value toward change (transformational)	Interest-driven value (transactional)

Hypothesis 1 stems from the distinction of innovation leadership based on individual traits. It assumes that a reflexive-learning-based leadership development program can promote ambiguity tolerance, interdisciplinary knowledge, abstract-experiment-oriented learning styles, and transformational values toward change.

Hypothesis 1 (H1):

A reflexive-learning-based leadership development program (L) enhances exploratory innovation leadership (O).

Organizational structures for innovations

Supportive environments

Organizational structures denote the explicit or implicit rules that regulate organizational innovations. This study examines how reflexive-learning-based leadership development influences supporting environment, organizational competency of reflection, and root cause identification corresponding to innovation. Building a supportive environment is positively associated with innovation performance (Elenkov & Manev, 2005; Sarros et al., 2008; Si & Wei, 2012). According to a survey investigating 291 R&D staff and 58 workgroups in Taiwan, a leadership team building an environment open to employees’ voices and ideas correlates to innovation significantly (Chen & Hou, 2016). Amabile and Kramer (2011) also found a significant association between employee performance and the manager’s support.

Freedom and autonomy have been regarded as the cornerstone of a supportive environment for innovation (Kunda, 1995; Pelz, 1963). However, Amabile et al. (1990) developed an instrument to test nearly 12,500 participants and found an insignificant association between freedom and innovation performance. Instead, a supportive environment is critical to innovation. Innovation thrives in mistake-tolerant (Farson & Keyes, 2002; Vomberg et al., 2020), risk-taking (Altinay et al., 2012; Furnham & Marks, 2013; Howell & Higgins, 1990; Mulder, 1987), open to ideas environment (Chen & Hou, 2016; Furnham & Marks, 2013). A German study highlighted that moderate failure tolerance, represented by an inverted U-shaped effect, positively influences performance without inducing more failures or increasing their severity (Vomberg et al., 2020).

Furthermore, individuals tend to manage risks (Kahneman & Tversky, 1983). Yet, exploratory businesses exhibit a higher risk-taking attitude toward change compared to others (Howell & Higgins, 1990; Mulder, 1987), a trait that positively correlate to ambiguity tolerance (Altinay et al., 2012; Furnham & Marks, 2013). Conversely, sustaining businesses are risk-averse as managers have options not to undertake a risk (Christensen, 1992) due to immediate financial returns (McGrath & MacMillan, 2009).

Finally, exploratory businesses are more likely norm-oriented than sustaining businesses with reward-driven structures. I. McGregor and Little (1998) found that reward-driven goal efficacy and norm-oriented goal integrity were positively related to one’s sense of meaning at work. Amitai Etzioni (1964) also argued that modern organizations tended to manage employees by employing symbolic (normative) tools to regulate employees within the organization rather than material, such as incentives and rewards (p. 59).

Organizational competency of reflection

This study views the organizational competency of reflection as the collective actions of acknowledging problems to avoid the biases attached to creative ideas or defensive routines. Reflection involves acknowledging problems in cognition and action, serving as the precondition of reflexive thinking for alternatives and developing new perspectives (Allen et al., 2019; Ripamonti et al., 2016). Organizational learning scholars revealed that leadership often avoids acknowledging problems to prevent individuals or organizations from experiencing embarrassment or threat, unless adopting reflection in cognition and action (Argyris, 1990, pp. 25–44, 1993, p. 15; Argyris & Schön, 1974, pp. 72–73). Studies have also indicated that the most effective way for leaders to overcome organizational incompetence is by publicly testing their assumptions (Argyris & Schön, 1974, p. 76–79). In essence, reflective learning is to identify ‘what is going on here?’ (Heifetz & Linsky, 2017, p. 52).

Root cause identification

After acknowledging problems, root cause identification aims to learn the gap between the intended goal and what actually happened in action (Garvin, 2003; Revans, 2016). Exploratory businesses rely on organizational settings to identify this gap (Garvin, 2003; Revans, 2016) regarding market opportunities and prototyping and cope with the transition in product delivery using a root cause analysis, in contrast to sustaining businesses that circumvent the delivery problems in scaleup and integration and diffuse the transformation in growth strategy (Christensen, 2016; Furr & Dyer, 2014). For example, ‘5 whys’ is a technique commonly used for problem identification to ascertain root causes. This method involves iterative questioning, akin to peeling away layers of an onion, probing deeper into the issue by asking ‘why’ five times.

Organizational structures are only effective for innovation if a reflective learning process exists. The second hypothesis assumes that a reflexive-learning-based leadership development program can catalyze organizational structure transformation toward innovation, enhancing supportive environments, the organizational competency of reflection, and root cause identification, as shown in Table 2.

Table 2. Organizational structures for innovations.

	Exploratory Businesses	Sustaining Businesses
Supportive Environments	Mistake tolerance Risk-taking Open to creative ideas	Mistake prevention Risk aversion Open to improvement
Organizational Competency of Reflection	Acknowledging the problems attached to ideas	Acknowledging the problems to avoid groupthink and defensive routines
Root cause identification	To identify difference between ideas and discovery (product-market-fit)	To identify the problem the gap between performance goal and the actual

Hypothesis 2 (H2):

The reflexive-learning-based leadership development program (L) enhances organizational structures (S).

Hypothesis 3 reflects that various business stages necessitate different organizational structures that require distinct leadership (Bledow et al., 2009; Chandy & Tellis, 2000; Cheng & Van de Ven, 1996; Christensen, 2016; Dougherty & Hardy, 1996; Drucker, 1993; Duncan, 1976; Furr & Dyer, 2014; March, 1991; O’Reilly & Tushman, 1997, 2004, 2016), postulating that innovation leadership is positively associated with organizational structures for innovation:

Hypothesis 3 (H3):

Enhancing exploratory leadership positively correlates to organizational structures for innovations.

Research design

Observing how collective reflexivity emerges within an engineering organization and its impact on leadership and organizational structures for innovation, this study facilitated a reflexive-learning-based leadership development program, conducting a 3-month field experiment with scientists and engineers widely distributed in the United States, China, India, Germany, and Slovakia in early 2022. This experiment employed 325 participants, including senior engineers, project leaders, department heads, and the general managers of engineering centers. Additionally, 67 change agents were randomly selected from various departments and project teams.

All change agents underwent the training to facilitate After-Action-Review (AAR), a form of collective reflexivity initially employed by the U.S. Army, leading the sessions examining team or department leadership. All change agents were organized into 11 diverse groups through different job functions and titles. Their training involved leading AAR meetings using a template provided in Table 3, encouraging open, safe, and supportive discussions, and eliciting participant feedback. Instead of giving participants instructions, change agents are trained to lead the discussions with full participation. Additionally, each group was required to generate meeting minutes or key takeaways for subsequent analysis.

Table 3. AAR meeting template.

Team Name:

Facilitator:

Participants:

Date: mm/dd/yyyy

Objective: Discover the gaps in leadership and organizational structures for innovations

Questions: 30 minutes (Invite every member to talk in an open, safe, and supportive environment rather than give them a lecture or instruction) What did we set out to do? 5 minutes (Invite participants to recall the discovery or delivery capabilities your project requires?) What did actually happen? 5 minutes (Invite participants to construct the theme. Ex. What discovery or delivery capabilities does your team have? What worked well and what did not?) Why did it happen? 10 minutes (Invite participants to re-examine the root cause. For example: Are you comfortable taking a risk in innovation? Do your team tolerate your failure? Do your leaders encourage you to try it again? If not, Why?) What are we going to do next time? 10 minutes (Invite participants to offer actionable recommendations. Ex. How to improve their leadership capabilities or build a more risk-taking and open-to-voice environment?) Tips: Invite participants to question the assumptions of recommendations. For example: Why do you believe it works? Lead participants to think of alternatives. For example: ‘What if’ your assumptions are wrong? Encourage participants to adopt a different perspective. For example: Can you tell us your second thought? What was it to block your sight before?

Despite deviations from the provided template during the initial exercises, the change agents demonstrated improvements in conducting more structured meetings as they progressed. Their objective was to champion the intervention to two dependent variables (O, S) by fostering collective reflexivity through regular AAR meetings in a safe, supportive, open-communication environment, as suggested by action researchers (Garvin, 2003; Revans, 2016). AAR provokes collective reflexivity by asking four questions (Garvin, 2003, p. 107): What did we set out to do? What did actually happen? Why did it happen? What are we going to do next time? These questions are consistent with reflexive learning process: acknowledging problems, reassessing assumptions, considering alternatives, and developing new perspectives. The first two questions aim to acknowledge the situation and reassess the gap between the goal and the actual. The third question was designed to question participants’ assumptions by identifying root causes. The last question helped participants think of alternatives and develop new perspectives.

Data collection

The collected data process involved participant surveys and change agents’ feedback. First, a scale with 48 questions, as shown in Table 3, was crafted and pre-tested using cognitive interviews with 4 people who have worked in the product design team and 40 graduate students with at least 7-year work experience in the tech industry or tech education, at two universities in New York City, coming from the U.S., India, China, Middle East, and Europe. In the official survey, all participants were required to respond to every question using a 7-point scale, ranging from ‘strongly disagree’ to ‘strongly agree’, to ensure data completeness and prevent missing entries.

This scale was sent to all 325 participants, 305 and 303 responses were received, respectively, before and after the leadership development program. However, only 287 responders completed both surveys and fully attended this program. The responses from the first survey formed the baseline for comparison with the second survey. Besides, 22 documents received from change agents contributed to the analysis, informing of what happened during ARR sessions.

Measures

Innovation leadership

Innovation leadership, outlined in Table 3, is a composite measure (O), representing two opposite leadership paradigms – exploratory and sustaining businesses – by aggregating 32 questions by four subfactors: ambiguity tolerance (O1), knowledge domain (O2), learning style (O3), and values (O4), based on Table 1. A higher score indicates a greater inclination toward exploratory leadership, characterized by high ambiguity tolerance, interdisciplinary knowledge, abstract-thinking-&-experiment-oriented learning style, and transformational values. In contrast, a lower the score tends to the sustaining leadership with ambiguity intolerance, single domain knowledge, concrete-thinking-&-experience-oriented learning style, and transactional values. This scale is a derivative of the Multifactor Leadership Questionnaire (Avolio & Bass, 2004), Budner’s Ambiguity Intolerance (Budner, 1962), and Kolb’s Learning Style Inventory Questionnaire (Kolb, 2013).

Organizational structures for innovations

Organizational structures delineate the explicit or implicit rules regulating organizational innovations, measured by three independent factors: participants’ perspective on an organization’s supportive environments (S1), organizational competency of reflection (S2), and root cause identification (S3), illustrated in Table 4. A higher score indicates a greater inclination of organizational structures favoring innovation. This measure includes 16 questions based on Table 2, derived from the Learning Organization Survey (Garvin et al., 2008) and Team Reflexivity Scale (Schippers et al., 2007).

Table 4. The scale of organizational structures and leadership.

Variables & Factors	Subfactors	Questions	Scores + (1 to 7) - (7 to 1)
Organizational Structures- Supportive Environments (S1)	Symbolic control	I can achieve my professional self in my team.	+
	Mistake tolerance	Our team tolerates my mistake and encourages me to retry.	+
	Risk taking	I feel free to take a risk to test new ideas in our team.	+
	Open to creative ideas or voices	Our leader welcomes new ideas and opportunities.	+
	Material control	Our performance goal is well connected to bonuses or rewards.	+
	Risk adverse	Our team must have everything finished ‘just right’ when completing a work assignment.	+
	Mistake preventing	Our team pays careful attention to details at work to ensure that nothing is overlooked.	+
	Open to improvement	Our leader is comfortable talking about problems, disagreements, and improvement.	+
Organizational Structures- Organizational Competence of Reflection (S2)	Prevention of the attachment to ideas and the lack of methodology for exploration	Our team regularly asks questions that challenge my ideas.	+
	The attachment to ideas and the lack of methodology for exploration	I am confident in our product and market direction and never doubt.	-
	Groupthink and defensive routines	I talk to our team in the way they desire.	+
	Groupthink and defensive routines	Our team is very productive but often receives unfair criticism or unreasonable requests.	-
Organizational Structures- Root Cause Identification (S3)	Identify the difference between ideas and discovery	Our team often directly observes how people interact with products and services.	+
	Identify the problem between ideas and discovery	In team meetings, we frequently discuss the gap between product design and customer feedback.	+
	Identify the gap between performance goals and the actual	Our team consistently monitors the gap between performance goals and the actual.	+
	Identify the gap between performance goals and the actual	In team meetings, we can productively review the root cause of the performance gap.	+
Innovation Leadership- Ambiguity Tolerance Variable (O1)	Complexity	In the long run, it is possible to get more done by tackling small, simple problems rather than large and complicated ones.	-
	Complexity	A good job is one where what is to be done and how it is to be done are always clear.	-
	Novelty	A person who leads an even, regular life in which few surprises or unexpected happenings arise really has a lot to be grateful for.	-
	Complexity	The sooner we all acquire similar values and ideals, the better.	-
	Complexity	It is more fun to tackle a complicated problem than to solve a simple one	+
	Complexity	People who insist upon a yes or no answer just don’t know how complicated things really are.	+
	Complexity	Often the most interesting and stimulating people are those who don’t mind being different and original.	+
	Complexity	Teachers or supervisors who hand out vague assignments give one a chance to show initiative and originality.	+
Innovation Leadership- Knowledge Domain Variable (O2)	Single Domain vs. Multiple Demains	Depth is more important than breadth in building my professional self.	-
	Improvement vs. Fundamental Change	In the face of competition, improvement is more likely to prevail than a fundamental change.	-
	Assimilator vs. Accommodator	My default focus is the delivery of results rather than finding principles or rules.	-
	Convergent vs. Divergent	A good leader should put rationality ahead of belief or value.	-
	Single Domain vs. Multiple Demains	I often find solutions to problems by drawing on solutions or ideas developed in other industries or disciplines.	+
	Improvement vs. Fundamental Change	I like to do what nobody else does.	+
	Assimilator vs. Accommodator	I rely more on reasoning than my experience and domain knowledge in decision-making.	+
	Convergent vs. Divergent	Differentiation is more important than performance improvement.	+
Innovation Leadership- Learning Style Variable (O3)	Experience vs. Experiment	I’m never afraid of trying something new in my daily work.	-
	Concrete vs. Abstract	Case studies are more valuable in learning than principles.	-
	Observation vs. Comprehension	Question and answer are more interesting than listening to a lecture.	-
	Instruction vs. Exploration	Following processes and rules is more important than creative thinking in new product design.	-
	Experience vs. Experiment	Learning from the past is more effective than trial and error.	-
	Concrete vs. Abstract	I find it easy to express my philosophy and values.	+
	Apprehension vs. Comprehension	I am good at making up new languages or secret codes in my childhood.	+
	Instruction vs. Exploration	I like to figure out things on my own without consulting anyone.	+
Innovation Leadership- Values (O4)	Interest vs. Value toward change	I often offer my assistance for returns.	-
	Orient norms vs. Disorient	I seldom re-examine my assumptions to view whether things are appropriate.	-
	Get the job done vs. Higher goal	Getting the job done is more important than thinking big.	-
	Material control vs. Symbolic control	I believe that crafting a vision is more effective than carrot (reward) and stick (punishment) in leadership.	-
	Interest vs. Value toward change	I think about tomorrow more than today.	+
	Orient norms vs. Disorient	A good leader is one who makes you wonder about your way of looking at things.	+
	Get the job done vs. Higher goal	I concentrate on discovering new possibilities more than solving problems to meet the organizational requirement.	+
	Material control vs. Symbolic control	Self-fulfillment is more efficient in team building than what one can expect to receive when performance goals are achieved	+

Confidence and reliability

With a sample set of 286 valid responders, representing a 99% confidence level at a 2% margin of error in a 315 population, this study ensures robust confidence in statistics. The Cronbach’s alpha coefficients of innovation leadership and organizational structures were 0.54 and 0.83, respectively, in the first survey and 0.70 and 0.85 in the second. These numbers demonstrate acceptable values about the reliability of the scales used in two surveys before and after the leadership development program. Besides, to prevent potential endogeneity in the research model, this study took cautions to eliminate the executives’ involvement and paused other training programs during the research. Still, the possibility that unobserved variables moderating the observed relationship between the independent and dependent variables may exist.

Findings

Effectiveness

To examine the effectiveness of the leadership development program (L) on the changes in leadership (O) and organizational structures (S) in innovation, a t-test validates the significance of the changes, as illustrated in Table 5. Its p-values (<0.05) support both Hypotheses 1 and 2. This reflexive-learning-based leadership development program (L) enhanced exploratory innovation leadership (O). Notably, the increase in ambiguity tolerance and interdisciplinary knowledge were significant, while the changes in learning styles and values were insignificant. This program (L) also enhanced organizational structures for innovations (S) even though the changes in subfactors, supportive environments (S1), organizational competency of reflection (S2), and root cause identification were insignificant.

Table 5. T-Test results.

	O – Innovation Leadership		O1 - Ambiguity Tolerance		O2 - Knowledge Domain		O3 - Learning Styles		O4 – Values
	Before	After	Before	After	Before	After	Before	After	Before	After
Mean	134.08	136.42	34.64	35.91	32.92	33.62	28.67	28.99	37.86	37.90
Variance	170.26	244.64	32.23	44.35	21.36	28.77	20.22	21.96	27.49	29.71
Observations	287	287	287	287	287	287	287	287	287	287
df	286		286		286		286		286
t Stat	−3.331201		−3.791986		−2.380441		−1.117566		−0.128448
P(T≤t) one-tail	0.000489		0.000091		0.008974		0.132345		0.448943
t Critical one-tail	1.650199		1.650199		1.650199		1.650199		1.650199
P(T≤t) two-tail	0.000978		0.000182		0.017947		0.264690		0.897885
t Critical two-tail	1.968293		1.968293		1.968293		1.968293		1.968293
	S – Organizational Structures		S1 - Supportive Environments		S2 – Org. Competency of Reflection		S3-Root Cause Identification
	Before	After	Before	After	Before	After	Before	After
Mean	81.24	82.45	43.28	43.82	17.56	17.85	20.39	20.78
Variance	131.51	133.43	45.34	49.28	7.36	7.39	20.34	18.01
Observations	287	287	287	287	287	287	287	287
df	286		286		286		286
t Stat	−2.064163		−1.443417		−1.722420		−1.635518
P(T≤t) one-tail	0.019952		0.074999		0.043037		0.051520
t Critical one-tail	1.650199		1.650199		1.650199		1.650199
P(T≤t) two-tail	0.039904		0.149997		0.086075		0.103040
t Critical two-tail	1.968293		1.968293		1.968293		1.968293

Correlations

The Pearson’s correlation coefficients outlined in Table 6 delineate the associations between leadership (O) and organizational structures for innovation (S). The data affirms Hypothesis 3, indicating a slight connection between exploratory innovation leadership (O) and organizational structures (S), both pre-and-post-moderation of the leadership development program (L). Moreover, the study initially found insignificant relationships between leadership traits – such as ambiguity tolerance (O1), interdisciplinary knowledge (O2), abstract-experiment-oriented learning style (O3), and transformational value (O4) – and two facets of organizational structures, namely supportive environment (S1) and root cause identification (S3). However, following the program (L), these correlations exhibited a positive trend, indicating an emerging connection between these leadership traits and the organizational structural factors.

Table 6. Correlational changes moderated by the leadership development program.

		O	O1	O2	O3	O4	S	S1	S2	S3
Before
	O – Innovation Leadership	1.000
	O1 - Ambiguity Tolerance	0.741	1.000
	O2 - Knowledge Domain	0.722	0.390	1.000
	O3 - Learning Style Variable	0.536	0.265	0.222	1.000
	O4 - Innovator’s Values	0.590	0.191	0.302	−0.005	1.000
	S – Organizational Structures	0.015	0.059	−0.037	−0.023	0.027	1.000
	S1 - Supportive Environments	−0.005	0.068	−0.076	−0.026	0.004	0.912	1.000
	S2 - Org Competency of Reflection	0.139	0.141	0.050	0.061	0.098	0.542	0.291	1.000
	S3 - Root Cause Identification	−0.037	−0.036	−0.011	−0.056	0.004	0.856	0.650	0.342	1.000
After
	O- Innovation Leadership	1.000
	O1 - Ambiguity Tolerance	0.815	1.000
	O2 - Knowledge Domain	0.758	0.515	1.000
	O3 - Learning Style Variable	0.527	0.237	0.251	1.000
	O4 - The Innovator’s Values	0.675	0.406	0.347	0.117	1.000
	S – Organizational Structures	0.134	0.143	0.118	0.011	0.086	1.000
	S1-Supportive Environments	0.015	0.068	0.039	−0.055	0.035	0.923	1.000
	S2 - Org Competency of Reflection	0.365	0.336	0.269	0.182	0.217	0.497	0.229	1.000
	S3 - Root Cause Identification	0.069	0.061	0.084	0.006	0.038	0.877	0.711	0.332	1.000

The leadership development program (L) failed to bring a significant change in the relationship between leadership (O) and organizational structures for innovation (S), as evidenced by Fisher’s z-transformation results indicating statistical insignificance (p=0.07), detailed in Table 7. However, the program (L) has notably strengthened three other correlations. Primarily, correlations between leadership traits (O1, O2, O3, O4) and all facets of organizational structures (S1, S2, S3) have experienced a notable tightening, particularly in the organizational competency of reflection. Secondly, there has been an enhancement in the correlation of the knowledge domain (O2) with organizational structures (S). Lastly, the correlations between organizational competency of reflection (S2) and leadership traits like ambiguity tolerance (O1), interdisciplinary knowledge (O2), and abstract-experiment-oriented learning style (O3) have exhibited an increase.

Table 7. Fisher’s Z-Transformation test results.

p-values	O	O1	O2	O3	O4	S	S1	S2	S3
O- Innovation Leadership
O1 - Ambiguity Tolerance
O2 - Knowledge Domain
O3 - Learning Style Variable
O4 - The Innovator’s Values
S – Organizational Structures	0.074369	0.156008	0.031021	0.341270	0.238852
S1-Supportive Environments	0.405165	0.500000	0.081357	0.363470	0.354574
S2 - Org Competency of Reflection	0.001717	0.006140	0.003374	0.069303	0.070916
S3 - Root Cause Identification	0.100978	0.121403	0.126519	0.227834	0.341270

Reflexive process

The understanding of how collective reflexivity emerged from reflexive learning was enriched through feedback provided by change agents. A frequency analysis with 22 documents collected from their AAR meetings employed a coding system based on the consensus between the researcher and three trained coders. An initial coding aligned with the survey variables recorded all salient texts that the change agents perceived in importance using various buckets. Table 8 summarizes a frequency analysis counted by their occurrences. 24 out of the total 28 counts are linked to organizational structures, indicating the most prominent improvement falling on organizational competency of reflection (S2), followed by root cause identification (S3) and supportive environment (S1). Only 4 counts are linked to innovation leadership, delineating the significance of ambiguity tolerance (O1), followed by domain knowledge (O2) and value (O4).

Table 8. Salient texts from change agents’ feedback.

Feedback	Innovation Leadership (4)	Organizational Structures (24)
	Ambiguity Tolerance (2) Uncertainty of problem (1) Uncertainty of solution (1) Knowledge Domain (1) Problem-solving skills (1) Learning Style (0) Values (1) Innovative thinking (1)	Supportive Environments (5) Openness to voices (4) Full participation (1) Organizational competency of Reflection (11) To challenge assumptions (4) To develop alternative thinking (6) To acknowledge problems or situations (1) Root cause identification (8) To identify root cause using 5 Whys (4) To identify the gap between the goaland the actual frequently (4)

The frequency analysis echoes the survey data, indicating that the program notably reinforces organizational structures for innovation (S). Noteworthy aspects within the competency of reflection involved reflexive steps such as ‘acknowledging problems or situations’, ‘reassessing assumptions’, and ‘developing alternative thinking’. In discussions related to root cause identification, the technique of ‘5 whys’, commonly used for problem identification to ascertain root causes, was mentioned several times.

Discussions

The synthesis of the findings serves as a springboard for generating theoretical insights into the impact of collective reflexivity on leadership and the organizational structures for innovation. Figure 2 presents a novel theoretical framework involving leadership, organizational structures, innovation, and intervention to expound upon the intricate workings of reflexivity within innovation. At the heart of this framework lies the pivotal role of leaders. They function as reflexive agents who are adept at not only self-learning but also assimilating external perspectives, validating their viewpoints against others, and aligning self-perceptions with the impressions projected by others during social interactions (Goffman, 1959, pp. 248–250; Mezirow, 1991, p. 185).

Figure 2. The theoretical framework of reflexivity in innovation.

Intricately linked to reflexive leadership is the practice of self-awareness, which involves recognizing personal biases and presuppositions (Antonacopoulou, 2004). The statistical analysis revealed a notable increase in the average innovation leadership scores following the leadership development program. This trend signifies a significant transformation among participants, showcasing a shift from upholding sustaining leadership traits to embracing an exploratory stance. Significantly, this transformation mostly appeared in the leaders with ambiguity tolerance and interdisciplinary knowledge.

Furthermore, leadership varies by the traits of ambiguity tolerance, domain knowledge, learning style, and values. The exploratory leadership exhibited a positive correlation with various dimensions of organizational structures. Notably, the robust correlation was observed in terms of the organization’s reflective competency. This correlation indicates that leaders inclined towards exploratory tendencies tend to exhibit higher levels of reflexivity, in contrast to their counterparts favoring sustaining approaches. Consequently, these exploratory leaders seem to gain more from leadership development programs, as they harness reflexivity to amplify their learning outcomes. These findings resonate with prior research, underpinning the idea that leaders adhering to sustaining approaches often gravitate towards conservatism, primarily due to the perceived threat posed by ambiguity (Argote et al., 1989; Budner, 1962). These discussions do not attempt to discount sustaining leaders whose importance remains unchanged in supporting existing businesses. However, this conservative inclination frequently leads them to rely on their intuitive faculties or past experiences when addressing challenges (Armstrong & Mahmud, 2017).

Secondly, a crucial element underpinning the support for collective reflexivity is the feedback system, serving as a vital social structure facilitating the progression of information from data through knowledge acquisition and ultimately to actionable insights (Antonacopoulou, 2004). In this context, a leadership development program like AAR assumes the pivotal role of a feedback system that catalyzes both individual and collective reflexivity. By furnishing a conducive environment, the AAR program offers leaders a platform to engage openly, securely, and supportively with peers, thereby nurturing an avenue for mutual learning. Participants and change agents can acknowledge problems or situations, reassess their assumptions, think of alternatives, and develop new perspectives (Argyris, 1990; Argyris & Schön, 1974; Mezirow, 1991; Sitkin, 1992).

Furthermore, developing new perspectives involves constructing narratives that establish legitimacy and foster identification for collective actions (Aldrich & Fiol, 1994; Lounsbury & Glynn, 2001; Lounsbury et al., 2019). This dynamic process engenders the emergence of a novel collective identity, a phenomenon neatly encapsulated by Walter Fisher (1984) in this conclusion that where reflexivity comes in, identification comes out; individuals construct a new story of their lives and possess the foundation through reflexivity. Telling a story or crafting a narrative is a sensemaking process, aiming to resonate with its audience to construct a new identity and legitimize new ventures (Davis, 2002; Lounsbury & Glynn, 2001; Lounsbury et al., 2019). For example, an organization can reframe problems and solutions to underscore the unbearable pains and urgent need for change for identity building (Davis, 2002; Smith, 1980). The discovery of communal identity contributes to forming a new worldview between the audience and the storyteller, makes one story better than another (Fisher, 1984; Leitch, 1986).

Unlike the communicative reflexivity that Archer (2007, 2010) defined as the internal conversations that require confirmation by others, sustaining structural continuity, collective reflexivity initiates collective actions toward change. Archer (2007) argued that reflexivity is a personal property mediating between (organizational) structure and actor (Caetano, 2015). However, the role of reflexivity in collective actions was barely found in her works. This field study underscores that participating in a reflexive-learning-based leadership development program can cultivate an organization’s reflective competence in the collective. For instance, the routinization of AAR in the workspace is instrumental to the emergence of collective reflexivity.

This study does not attempt to engage the role of reflexivity in the relationship between structure and agency within social systems. Instead, some scholars suggest that routinization supports the perpetuation of organizational structures (Bourdieu, 1990, p. 130; Weber, 2019, pp. 373–389), potentially leading to a stagnation of innovation (Duncan, 1976; Hannan & Freeman, 1984). Nonetheless, others advocate for the positive role of formalization or routinization in enhancing innovative capabilities (Morey et al., 2002; Ohly et al., 2006; Teece et al., 1997) and reducing uncertainties within social systems (North, 1990, pp. 3–10). This study’s data implies that collective reflexivity can bridge the gap between two opposite camps – the impact of formalization and routinization on the organizational structures is contingent on the degree to which reflexive learning involves challenging entrenched assumptions and fostering new perspectives.

Thirdly, developing alternatives or new perspectives within AAR meetings encompasses a framing strategy for collective actions. This strategy crafts a narrative by presenting challenges alongside potential solutions, urgency juxtaposed with necessities, and difficulties framed with transformative changes (Davis, 2002; Smith, 1980). Such strategic framing serves as an essential role in navigating the uncertainties in innovation, such as market identification, product delivery, scale-up, and crafting new growth strategy. At the core of this approach lies two fundamental tasks of collective action framing: firstly, a problematic framing that represents a shared understanding of problems, and secondly, the provision of solutions to address these predicaments (Benford & Snow, 2000; Snow et al., 1986). It’s crucial that both the problem and its corresponding solution are substantiated through a careful analysis of benefits in relation to costs (North, 1990, p. 84). If the ratio of initiation costs (Boudreau et al., 2017; Edmondson, 1999; Hasan & Koning, 2019) to incentives (Becker, 1962; Sandvik et al., 2020) is too high, implementing changes becomes challenging. For example, enterprises often craft customer value propositions by pairing customer pains with relief and problems with solutions (Olsen, 2015; Osterwalder et al., 2015). Adopting this approach can effectively address the uncertainties in innovation.

Conclusion

This study significantly contributes to our understanding of reflexive-learning-based leadership development as a catalyst for fostering collective reflexivity, which is essential in effectively navigating multifaceted uncertainties in innovation. The effectiveness of reflexive learning, exemplified by AAR meetings, lies in its capacity to initiate collective action by framing narrative strategies within problem-solving contexts. By placing collective reflexivity at the forefront of this research, a novel conceptual framework emerges, enriched by the insights drawn from a field study conducted within an engineering organization. The findings offer compelling evidence of reflexive learning’s profound impact on both leadership and organizational transformation in innovation. Despite inherent challenges in representing and generalizing findings from a field study, this research carefully selected an engineering organization as its subject, albeit with limitations due to a constrained sample size.

This field study took place amid the post-Covid-19 landscape, with AAR meetings embracing a hybrid format with in-person and remote participants. Despite this mixed mode, the leadership development program effectively induced shifts in organizational structures and leadership traits, particularly ambiguity tolerance and interdisciplinary knowledge. While some facets, including the abstract-experiment-oriented learning style, transformational value, reflective competency, supportive environments, and root cause identification, did not show statistically significant enhancements during the program, it’s prudent to consider that a different format or an extended duration might yield altered outcomes. The potential of longitudinal research holds promise for a more comprehensive assessment of these outcomes over an extended timeframe.

Furthermore, the study revealed that exploratory leaders demonstrated more reflexivity and absorptive capacity than their sustaining counterparts within the leadership development program. This study echoes the finding suggested by Hirst et al. (2004) that exploratory leaders will learn significantly more than experienced sustaining leaders, encompassing the design of action research endeavors and organizational leadership development programs.

Yet, this study aims to emphasize the role of collective reflexivity when an engineering organization navigates those uncertainties in innovation. Observing the influence of collective reflexivity on real-world innovation is very challenging. This study designs a reflexive-learning-based leadership development program to intervene in innovation leadership (O) and organizational structures (S). The measurement of the change in O and S reflects the effectiveness of this program. Importantly, two types of exploratory and sustaining leadership are ideal categorizations for analytical purposes for evaluating reflexive learning. A mixture or more types of leadership, for example, ambidextrous leadership (O’Reilly & Tushman, 2004, 2008), likely co-exist.

Finally, this study distinguishes itself by exploring the manifestation of collective reflexivity in the context of innovation. Although reflexive learning undoubtedly instigated significant changes in leadership and organizational structures within an engineering organization, it did not directly measure the correlation between these transformations and innovation outcomes, such as patents, performance, or financial results. This approach introduces a novel research framework that encourages a broader exploration of collective reflexivity and its interconnectedness with various other facets for future research.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Jeremy Wei

Jeremy Wei is an instructor at Harvard University and holds the position of an adjunct professor at the University of Science and Technology of China. Additionally, he founded iDox.ai, a prominent AI-powered data privacy management solution, advocating digital civil rights. His research interests predominantly revolve around AI and data privacy, technology adoption, reflexive learning, leadership, and innovation. Wei earned his Ed.D. from New York University.