Leaders need to be led: complementary followership in the context of community-driven development program

ABSTRACT

Leadership has often been recognised as a major driver for successful team effectiveness. However, even weak leadership may lead to good team performance, and it is worth studying how weak leadership can be helped and complemented by followership. To investigate the paradoxical mechanism behind leadership – followership practices, we examined (1) multidimensional figures of leadership and followership using the multifactor leadership questionnaire (MLQ) and (2) the impacts of the combinations of leadership and followership on team performances both during and after a community-driven development (CDD) program. To that end, this study examines a rural CDD case implemented by the Korea International Cooperation Agency and the Myanmar Ministry of Agriculture, Livestock, and Irrigation. The analyses present some common patterns of how weak leadership and strong followership can lead to better performance than other combinations of leadership and followership. We conclude with theoretical and practical conditions of “complementary followership”, i.e., the complementary combinations of leadership and followership in group performance.

KEYWORDS:

• Leadership
• followership
• community-driven development
• group performance

Introduction

Leaders have been considered important for team effectiveness as they can help in enhancing a team’s effectiveness by setting goals, assigning responsibilities, supporting and developing resources and management systems, and overcoming major hurdles (Andrews & Boyne, 2010; Yukl, 2013). Much of the literature on leadership addresses the importance of leaders in the context of relationship (Crossman & Crossman, 2011; Graen & Uhl-Brien, 1995) and public service (Crosby & Bryson, 2018; Hameduddin & Engbers, 2022). Still, team effectiveness can be enhanced by not only leaders’ roles but also teams’ collective behaviour in which members assume their own responsibilities and collaborate with one another. Followers, sometimes as part of a leadership shared among members, can also help in reducing conflict and reaching a consensus (Bergman et al., 2012; Crossman & Crossman, 2011). In this sense, recent studies suggest that leadership and followership cannot be considered separately for the sake of team performance (Grego-Planer & Baker, 2022).

In order to explore the dynamics of leadership and followership, it is important to first clearly define these two terms. While leaders and followers may hold formal positions as managers and subordinates in organisations, their roles can be interchangeable. In some cases, formal followers (subordinates) may take on proactive roles as informal (actual) leaders, while formal leaders (managers) may act as informal (actual) followers who support proactive subordinates. Therefore, the terms “leadership” and “followership” may refer to either formal positions or informal (actual) roles in organisations. To ensure clarity in this study, leadership and followership are operationally defined as the relational characteristics expected of or exercised by individuals holding formal positions (managers and subordinates) with corresponding formal authorities, where authority refers to the legitimate power vested in the position derived from the organisation’s formal structure, allowing the individual to make decisions, give orders, and enforce rules.

Many studies on the situationality of leadership assert that the effectiveness of certain leadership styles depends on the characteristics of followers and many other situational contexts (Hersey & Blanchard, 1993; Yukl, 2013). Imagine a simplified situation where leadership and followership can be characterised as either strong or weak. Although the idea that strong leadership may lead to better team performance might be intuitively acceptable, the opposite notion can also be feasible. In other words, even weak leadership can leave space for followers to participate more proactively in team performance and thereby complement the weak leader. In short, it is conceivable that, under certain circumstances, a combination of weak leadership and strong followership can be another positive driver for team performance. With this in mind, this study explores the patterns and drivers of this paradoxical case (i.e., weak leadership and strong followership leading to better team effectiveness).

To look into such paradoxical impacts of the leadership – followership mix, we used a community-driven development (hereafter CDD) case implemented by the Korea International Cooperation Agency (KOICA) and the Myanmar Ministry of Agriculture, Livestock, and Irrigation (MOALI). We surveyed the CDD leaders and randomly selected residents of 100 villages. In doing so, we examined two aspects of CDD practices: (1) multidimensional figures of leadership and followership, using the multifactor leadership questionnaire (MLQ), and (2) the impacts of the combinations of leadership and followership on the performance and sustainability of CDD. In the next section, the existing arguments and theories on leadership – followership as drivers of team effectiveness are revisited; from them, the research questions are then driven and presented.

Theories and research questions

This section provides a review of existing arguments and theories on the variables of interest in this study. Firstly, the literature on the dependent variable, team effectiveness, is revisited. Secondly, the drivers of team effectiveness and various arguments on leadership and followership are also reviewed. This review ultimately leads to the formulation of research questions.

Team effectiveness

A team is a group of individuals who share a sense of membership, interdependent roles, and common goals (Northouse, 2016). However, the exact definition of a team is far from fixed, as the organisational environment surrounding the characteristics of “team” is ever-changing (Wageman et al., 2012). Besides the variations in “team” definitions, “team effectiveness” is another composite of various concepts. For instance, team effectiveness is often a collection of performances (task accomplishment) and development (team maintenance and cohesiveness), as described by Northouse (2016).

As for the focus of this study, “team effectiveness” in terms of community development has a very diverse definition consisting of at least four criteria. The first criterion of community development’s effectiveness is “inputs”, such as financial and technical support to the community (Wong, 2012). Second, the “outputs” of community development are contributions from and communication among the people (Nguyen & Rieger, 2017), in addition to ownership and emotional connection among residents (Cloutier et al., 2019; R. Chase & Woolcock, 2005). Third, outcomes of team efforts in community development include access to basic services, such as water, health, and nutrition (Arcand & Wagner, 2016; Wong, 2012), income generation (Baird et al., 2013; Mansuri & Rao, 2004; Nkonya et al., 2012), social capital (Yalegama et al., 2016); and conflict resolution (Khwaja, 2009). Finally, “longer-term outcomes” of community development imply the community’s capacity for self-sustainable community development (Dasgupta & Beard, 2007; Lund & Saito-Jensen, 2013; Platteau, 2004), such as inclusive governance (Arcand & Wagner, 2016) through distributed information and power (R. S. Chase & Christensen, 2014).

Drivers of team effectiveness

The modifiers of team effectiveness have been studied by many scholars and practitioners. For instance, Larson and LaFasto (1989) argued that team effectiveness is a result of various factors: (1) clear and elevating goals, (2) results-driven structure, (3) competent team members, (4) unified commitment, (5) collaborative climate, (6) standards of excellence, (7) external support and recognition, and (8) principled leadership. Hackman (2012) also suggested a comprehensive list of factors behind team effectiveness, such as compelling purpose, right people, clear norms of conduct, supportive organisational context, and team-focused coaching. In addition to the drivers, team type and situation are considered moderating variables for team performance (C. S. Burke et al., 2006; Hulsheger et al., 2009).

When it comes to community development, team performances are influenced by several factors. First, there are external factors for community development program design: sustainable and timely financial support (Wong, 2012); coordination and collaboration among the supporting stakeholders and organisations (Bowen, 2005; Platteau, 2004; Wong, 2012); and the design of the participation and evaluation mechanism (Grossman & Hanlon, 2014; Khwaja, 2009). Second, the community’s external environment is another factor to consider, including the natural environment (Brinkerhoff et al., 2018; Wong, 2012) and the local economy and education (Baird et al., 2013; Nkonya et al., 2012). Finally, and most importantly in this study, the community’s internal environment is an indispensable condition for community development: governance, participation, capacity development, and leadership (Bassoli, 2010).

Leadership and followership as drivers of team effectiveness

Leadership can be an internal driver for team effectiveness, and leadership can be defined by two roles for monitoring and acting in two environments: (1) internal affairs (such as task and intra-relations) and (2) external relationships (Hackman & Walton, 1986). More specifically, leaders are expected to promote the setting and sharing of vision, inform people of the progress and impacts of the change, encourage innovation and learning, and reward exemplary behaviours (Yukl, 2013). Building a coalition to support and guide organisational change is another role of leaders (Denis et al., 2001).

By focusing on empowering (or participative) leadership, good leaders are believed to involve and empower competent people to formulate and implement change for the various benefits of participation, such as realistic decision-making, compliance, and capacity building (Yukl, 2013). Good followership (followers having initiative, skill, and a sense of ownership) as well is known to be an important condition for successful empowerment (Yukl, 2013).

Situational leadership is often studied as a type of leadership that also considers followership (Northouse, 2016). This type of leadership suggests a nonlinear pattern of desirable leadership styles between directive and supportive leadership according to the competence and commitment of followers across time (Hersey & Blanchard, 1993; K. Blanchard et al., 1993, 2013). Similarly, shared leadership (Day et al., 2004), also referred to as team leadership (Yukl, 2013), which is similar to distributed leadership (Oborn et al., 2013), concerns leadership that considers followership more actively. Shared leadership happens when leadership behaviours and influences are shared and distributed among team members, unlike leadership in a hierarchical structure (Bergman et al., 2012). In practice, teams with shared leadership exhibit more trust and cohesion than teams without shared leadership (Bergman et al., 2012; Northouse, 2016).

In addition to examining the effects of each leadership style on team effectiveness, there are several arguments related to competing leadership styles. Specifically, there are several typical dilemmas among different leadership types, such as strong versus networked leadership (Verheul & Schaap, 2010) and transformational versus transactional leadership (Nielsen et al., 2019), which can have varying effects on employee behaviours (Moldogaziev & Silvia, 2015; Moon & Park, 2019).

Complementary followership

Shared leadership implies the importance of followership that comprises the other side of leadership. The term “followership” tends to be defined as the passive, compliant, or obedient behaviour of organisation members (Berman et al., 2013). Similarly, followers have often been studied as part of the dependent variable in organisational studies, such as the subject of job satisfaction or performance (A. L. Blanchard et al., 2009; Hattke et al., 2018). On the other hand, in addition to the importance of leadership, followership is considered an active factor in organisational management (Vogel & Masal, 2015). For instance, Kelley (1992) defined effective followership as a combination of independent thinking and active engagement. Similar to leaders, followers also evaluate and judge their leaders’ intentions and competence (van Knippenberg et al., 2004). Moreover, how followers perceive their leaders’ behaviours may matter more than what leaders intend to do (Jacobsen & Andersen, 2015). Thus, followership can serve as a mediating factor between leadership and team performance (C. Kim & Schachter, 2015). Therefore, followers can significantly contribute to effective leadership (Baker, 2007). In this sense, leadership and followership work together and create synergistic dynamics that can enhance team effectiveness (Hassan & Hatmaker, 2015; Hassan et al., 2019; Oberfield, 2014; Ritz et al., 2012). Additionally, motivational factors such as public service motivation (PSM) can mediate the effects of leadership and followership on team performance (Jensen & Bro, 2018; Kroll & Vogel, 2014; Miao et al., 2018; Schwarz et al., 2016).

Table 1 shows several conceivable scenarios concerning the impact of leadership and followership on team effectiveness. In detail, a team can either have strong or weak leadership and followership. The combinations of leadership and followership can produce a beneficial or harmful impact on team effectiveness. Among the various scenarios, despite the significant role of followers, the idea that weak leadership can be better helped by strong followership (as shown in the highlighted cells in Table 1) has not been extensively studied. Therefore, this study examined when and how strong followership can complement weak leadership.

Table 1. Ambivalent impacts of the combinations of leadership and followership on Community-Driven Development (CDD) performances.

Combinations of leadership & followership		Impact on team effectiveness
		Beneficial	Harmful
Similar	S – S: Strong leadership, Strong followership	Reinforcing	Monolithic
	W – W: Weak leadership, Weak followership
Different	S – W: Strong leadership, Weak followership	Complementary, balancing	Competing, contradictory
	W – S: Weak leadership, Strong followership

Note: The highlighted part represents the focus of this study.

Figure 1 shows the research model consisting of three parts. The first part is the two response variables of team effectiveness in the context of CDD: (1) CDD performance during the official program and (2) CDD sustainability after the official program. The second part deals with the explanatory variables – the combination of leadership and followership, where we focus more on weak leadership and strong followership (W – S). The third part represents the dimensions of leadership where leadership and followership can be defined multidimensionally according to the MLQ factors.

Figure 1. Research questions at a glance.

The three-part research model described above corresponds to three research questions. In formulating these questions, we considered whether there were sufficient theoretical or practical arguments to help hypothesise the (directions and magnitudes of) associations between the variables. However, despite the abundance of studies on leadership and followership, due to a lack of research on the complementary patterns between them, it is difficult to derive a set of confirmatory hypotheses. Therefore, the research question for this study is expressed in exploratory terms as follows:

The first question is about whether and how the situation of weak leadership and strong followership (W – S) is associated with CDD performance during the official period of the CDD program. In other words, while being observed and rewarded by external overseers, can the village with weak leadership and strong followership perform better than other villages? The second question in this study is whether and how the pair of weak leadership and strong followership (W – S) is related to the sustainability of CDD. Even after the official period of the program (i.e., even without external material compensation for CDD practices), can a village with weak leadership and strong followership sustain its CDD efforts better than other villages? The last question is about how the impacts of W – S paired with CDD performances and sustainability would vary according to the multiple dimensions of leadership on the MLQ. The next section presents the research methods employed to answer the research questions.

RQ1.

Can a W – S state (for example, a combination of weak leadership and strong followership in terms of the MLQ leadership factors) lead to higher CDD performance (represented by, for instance, a CDD evaluation score during the CDD program) than other combinations of leadership and followership?

RQ2.

Can a W – S state (for example, a combination of weak leadership and strong followership in terms of the MLQ leadership factors) lead to higher CDD sustainability (such as the continuation of CDD activities after the CDD program has ended) than other combinations of leadership and followership?

RQ3.

For which MLQ leadership factors does the W – S state work best in terms of CDD performance? (In other words, for which MLQ leadership factors does a strong followership most complement a weak leadership?)

Methods

Case

We examined a rural CDD program in Myanmar implemented by the Korea International Cooperation Agency (KOICA) and the Myanmar Ministry of Agriculture, Livestock, and Irrigation (MOALI). One of the least developed countries in Southeast Asia, Myanmar had a GDP per capita of US $1,407 in 2019 (World Bank), with 32.1% of the population living below the national poverty line in 2015 (Asian Development Bank). Much higher poverty rates are found in rural areas. Rural development is critical to the country’s economic development, given the country’s large rural population and its dependence on natural resources and agriculture. In this context, the Korea International Cooperation Agency (KOICA) and the Myanmar Ministry of Agriculture, Livestock, and Irrigation (MOALI) implemented a three-year rural CDD program in 100 villages across nine regions (Ayarwaddy, Bago, Mandalay, Mon, Nay Phi Taw, Sagaing, Shan, Tanintharyi, and Yangon) between 2015 and 2019.

In this scenario, the CDD program presented three main objectives: (1) capacity building, (2) improving the living environment, and (3) income generation under which the various key performance indicators (KPIs) presented in Table 2 were measured and evaluated by the end of each program year. The CDD program in Myanmar differs from existing rural CDD programs in other countries in that participating villages compete with each other and receive different amounts of financial incentive based on their annual KPI ranking. During the first year (January 2016–January 2017), an amount equal to US $20,000 was provided equally to the 100 villages in which the CDD program was implemented. However, during the second (January 2017–January 2018) and third (January 2018–January 2019) years, the top-ranking 30% and the next 40% of the villages received amounts equal to US $40,000 and US $30,000, respectively. The bottom 30% of the villages received US $20,000. This differential incentive scheme based on inter-village competition makes the Myanmar CDD program unique compared to other CDD approaches (B. Kim et al., 2021).

Table 2. The CDD activities in Myanmar.

Objectives	Weight (%)			Key Performance Indicators (KPI) of CDD
	Year 1	Year 2	Year 3
Capacity building	10	10	10	Number of meeting
				Number of technical/educational training
				Number of trainees
				Number of villagers visit to other advanced villages
				Number of public information on project movement
Living environment improvement	55	40	30	% of project work completed in a year based on the original plan
				Villagers’ fund contribution
				Villagers’ labour contribution
				Number of participant households
				Villagers’ land/materials contribution
Income generation	35	50	60	Fund increment by means of interest
				Number of microfinance participant households
				Adoption of new business and technologies

Data collection

The final (third-year) evaluation of this CDD program was completed in January 2019, and we conducted a follow-up survey in August 2020 that focused on the CDD program leader and 20 randomly selected village residents from each of the 100 villages. Before the survey, we conducted preliminary interviews with those who were employed by the Korea International Cooperation Agency (KOICA) to oversee the CDD program. The KOICA interviewees shared a common observation that the success or failure of CDD in Myanmar significantly depends on the characteristics of the leader of the CDD program. Therefore, considering the significant influence of program leaders, the data for leadership in each village were collected via an interview with the CDD program leader.

The questionnaire for the survey and interview was originally designed in English not only because of the researchers’ familiarity with English but also because of the fact that the majority of theories and literature that contain the relevant keywords and models were written in English. The original questionnaire in English was translated into the Myanmar language by professional translators who are highly trained to accurately reflect the original meanings of messages. To conduct the translated survey, we used Survey Solutions, a computer-assisted personal interviewing (CAPI) program developed by the World Bank Group to improve data quality and initiate faster data collection. Professional survey enumerators from Myanmar Survey Research (a leading survey company in Myanmar) interviewed 100 CDD program leaders in addition to 1,996 villagers (either household head or another family member who knew the most about the household) face-to-face while strictly following the Myanmar government’s regulations concerning the coronavirus 2019 (COVID-19) epidemic. The enumerators conducted their interviews by asking individual interviewees each question and entering their responses into the CAPI system.

Appendix 1 presents descriptive statistics for the study sample. The CDD program leaders (in Panel A) were 52.4 years old on average, 99% were male, and 91% were married. Their education levels and annual incomes were much higher than those of the village residents. The randomly selected village residents (in Panel B) were 48 years old on average, and half of them were female. Only 4.3% had graduated from high school, and their annual income level (503,546 kyat) was approximately half that of the CDD program leaders (1,053,301 kyat). As for the village information (in Panel C), there are about 200 households in each of the 100 villages. The major profession of household heads is farmer or fisherman (48.6%). Approximately 64% of households in each village are able to use electricity. Most villages (84%) have public primary schools, and a few (5%) have public high schools. About 40% of the villages have a medical facility such as a health clinic, health centre, or hospital.

Ethical review and informed consent

Ethical aspects of this study were approved by the Myanmar Ministry of Agriculture, Livestock, and Irrigation (Admin/SMU/2016–360) and the KDI School of Public Policy and Management Ethics Review Committee (KDIS-IRB-2016-04). Regarding informed consent, we obtained verbal consent from all research participants before initiating the tablet-based mobile survey. Once verbal consent was given to a survey enumerator, the enumerator clicked the “yes” button for informed consent on the tablet screen, and this information was automatically saved in the survey database system.

Measurement and analysis

To measure two response variables (performance and sustainability of the CDD program), we used two sources of information. First, we took the three-year average ranking (percentile), which represents the overall CDD performance per village. Second, we asked respondents about whether the CDD committee and three subdivision activities continued to operate even after the end of the CDD program by using a five-point Likert scale: (1) not at all anymore, (2) rarely, (3) sometimes, (4) actively, and (5) very actively. In short, the response variables of interest, $Y_v$, are included in Equation (1):

$Y_v=\mathit{\alpha }+\mathit{\beta ML}{Q}_v^{SS}+\mathit{\gamma ML}{Q}_v^{SW}+\mathit{\delta ML}{Q}_v^{WS}+\mathit{\theta }{X}_v+{\epsilon }_v$

Regarding the set of explanatory variables (leadership and followership), we considered that leaders are often expected to exercise various roles in terms of task, relations, change, diversity, and integrity inside and outside an organisation (Fernandez et al., 2010). With this in mind, we used the MLQ-6S (Bass & Avolio, 1992), a widely used tool measuring various dimensions of leadership, for explanatory variables. The MLQ-6S consists of 21 five-point Likert scale questions to evaluate seven leadership factors: (1) idealised influence, (2) inspirational motivation, (3) intellectual stimulation, (4) individualised consideration, (5) contingent reward, (6) management by exception, and (7) laissez-faire. What is notable is that the seven factors can be divided into three categories that represent different leadership styles: transformational (factor 1 to 4); transactional (factor 5 and 6); and passive/avoidant (factor 6 and 7). Although various arguments have addressed the similarities and differences between leaders and managers since Zaleznik (1981), the MLQ does not clearly distinguish between them. Instead, the MLQ combines the common characteristics and roles that those in charge of groups or organisations are expected to have. The meanings and interpretations of the seven MLQ factors are specified below.

Factor 1. Idealized influence indicates whether you hold your subordinates’ trust, maintain their faith and respect, show dedication to them, appeal to their hopes and dreams, and act as their role model.

Factor 2. Inspirational motivation measures the degree to which you provide a vision, use appropriate symbols and images to help others focus on their work, and try to make others feel their work is significant.

Factor 3. Intellectual stimulation shows the degree to which you encourage others to be creative in looking at old problems in new ways, create an environment that is tolerant of seemingly extreme positions, and nurture people to question their own values and beliefs and those of the organization.

Factor 4. Individual consideration indicates the degree to which you show interest in others’ well-being, assign projects individually, and pay attention to those who seem less involved in the group.

Factor 5. Contingent reward shows the degree to which you tell others what to do in order to be rewarded, emphasize what you expect from them, and recognize their accomplishments.

Factor 6. Management by exception assesses whether you tell others the job requirements, are content with standard performance, and are a believer in “if it ain’t broke, don’t fix it”.

Factor 7. Laissez-faire leadership measures whether you require little of others, are content to let things ride, and let others do their own thing.

The MLQ consists of three standardised questions for each factor, which are presented in Table 3. As this study focuses on whether followers can complement (and even take on) their official leaders’ roles, the same measurement (MLQ) was used to measure both leadership and followership. Thus, we asked the CDD program leaders and the village residents to answer the MLQ-6S questions and defined the MLQ scores of program leaders and village residents as leadership and followership, respectively.

Table 3. Variables and measures.

Variables in the models		Measures (unit: village)
Response variables
	- Performance of CDD (i.e., CDD activities’ score during CDD program)	3-year average ranking (percentile)
	- Sustainability of CDD (i.e., CDD activities’ continuation after CDD program)	“How much actively does each of the following CDD activities continue after the end of CDD program?” Sum of the 5-point Likert scale for each of the following CDD activities CDD Committee Capacity building sub-division Living environment improvement sub-division Income generation sub-division
Dimensions of leadership (MLQ-6S measures with 5-point Likert scale of each)
	Factor 1- Idealised influence	Question number 1. I make others feel good to be around me. Q8. Others have complete faith in me. Q15. Others are proud to be associated with me.
	Factor 2- Inspirational motivation	Q2. I express with a few simple words what we could and should do. Q9. I provide appealing images about what we can do. Q16. I help others find meaning in their work.
	Factor 3- Intellectual stimulation	Q3. I enable others to think about old problems in new ways. Q10. I provide others with new ways of looking at puzzling things. Q17. I get others to rethink ideas that they had never questioned before.
	Factor 4- Individual consideration	Q4. I help others develop themselves. Q11. I let others know how I think they are doing. Q18. I give personal attention to others who seem rejected.
	Factor 5- Contingent reward	Q5. I tell others what to do if they want to be rewarded for their work. Q12. I provide recognition/rewards when others reach their goals. Q19. I call attention to what others can get for what they accomplish.
	Factor 6- Management-by-exception	Q6. I am satisfied when others meet agreed‐upon standards. Q13. As long as things are working, I do not try to change anything. Q20. I tell others the standards they have to know to carry out their work.
	Factor 7- Laissez-faire leadership	Q7. I am content to let others continue working in the same ways always. Q14. Whatever others want to do is OK with me. Q21. I ask no more of others than what is absolutely essential.
Explanatory variables
	Combinations of leadership and followership	Leadership (SMU Chair of each village) and followership (20 villagers randomly chosen from each village) values based on MLQ measures Strong: above the overall mean value Weak: below the overall mean value
		2 by 2 combinations of leadership and followership S – S: Strong leadership and Strong followership S – W: Strong leadership and Weak followership W – S: Weak leadership and Strong followership W – W: Weak leadership and Weak followership (reference in the regression analysis)
Control variables
	- Number of households	Statistics (average) of the village administrative data
	- Average age	Statistics (average) of the randomly selected 20 village residents
	- Percentage of female people	Same as above
	- Education level of village	Same as above (% of household where head of household is educated beyond primary school level)
	- Average monthly income	Same as above

For each MLQ factor, we calculated the village-level average for leadership and followership. We used the average value of villagers’ responses in each village if their responses were reliably consistent within each village. To address the inter-rater reliability problem, we estimated inter-rater agreement (IRA) by using multiple indices, as suggested by M. J. Burke et al. (1999) and James et al. (1984). The estimation results show that the degrees of consensus within villages were significantly high. The specific statistics of IRA indices are presented in Appendix 2.

Based on the MLQ factor scores, a strong leadership (followership) is then defined as above-average MLQ scores at the village level. Likewise, weak leadership (followership) is indicated by village-level MLQ scores that are below the average value. Considering the strong – weak categories of leadership and followership together, we generated four possible combinations: (1) strong leadership – strong followership (S – S), (2) strong leadership – weak followership (S – W), (3) weak leadership – strong followership (W – S), and (4) weak leadership – weak followership (W – W) for each MLQ factor per village. The mean values of MLQfactors for these four combinations are presented in Appendix 3. We examined these leadership – followership combinations not only in the same MLQ factor but also across different MLQ factors between the leaders and the followers and then generated a 7 × 7 matrix.

Using the aforementioned variables, we examined whether and how the pair of interests (weak leadership and strong followership, W – S) works better (or worse) for CDD performances and sustainability than the presumably worst situation (W – W) and other pairs. Therefore, while omitting W – W as the reference, we used the S – S ($\mathit{ML}{Q}_v^{SS})$, S – W ($\mathit{ML}{Q}_v^{SW})$, and W – S ($\mathit{ML}{Q}_v^{WS})$ as dummy variables. Regarding variable scale, as the continuous scale must be transformed into a categorical one to test simplified scenarios and hypotheses (Lee, 2021), the original data of MLQ with a five-point Likert scale was re-coded to the dummy variables. Among the three explanatory variables, we were more interested in W – S, where weak leadership was combined with strong followership. Thus, our main coefficient of interest was $\mathit{\delta }$ in Equation (1).

Finally, paying more attention to the unique impacts of the leadership – followership combination on CDD performances and sustainability, we controlled other situational contexts regarding the village-level characteristics, $X_v$, such as village size (number of households), average age, percentage of female people, education level of village, and average monthly income, in the regression equation. Table 3 summarises all variables and measures.

As this study utilises many regression models based on the same data set, the different tests can have a joint dependence structure. Thus, when it is conceivable that false judgement (such as a Type I error) can occur in a series of hypothesis tests, the familywise error rate (FWER) must be controlled (List et al., 2016). To address such a false rejection problem, we controlled the FWER and used adjusted p-values for analyses and interpretations.

Findings

A comprehensive summary of the findings is presented in Table 4 (see Appendices 4 and 5 for the regression results in detail). Table 4 shows how the combination of weak leadership and strong followership (such as the W – S) in the seven leadership factors is associated with team effectiveness either during or after the CDD program. In Table 4, the star symbol (★) represents the cases in which the CDD KPI score (during the CDD program) was highest in the W – S among the four types (S – S, S – W, W – S, W – W) and was also statistically significant. The arrow symbol (⇨) represents the cases in which the CDD activities continued (after completion of the CDD program) most actively in the W – S among the four combinations (S – S, S – W, W – S, W – W); this case was also statistically significant.

Table 4. Impacts of combination of weak leadership and strong followership on team effectiveness.

Weak leadership in …	Strong followership in …
	Factor 1: Idealised influence	Factor 2: Inspirational motivation	Factor 3: Intellectual stimulation	Factor 4: Individualised consideration	Factor 5: Contingent reward	Factor 6: Management-by-exception	Factor 7: Laissez-faire
Factor 1: Idealised influence
Factor 2: Inspirational motivation
	⇨	⇨	⇨	⇨	⇨
Factor 3: Intellectual stimulation				★	★
	⇨	⇨
Factor 4: Individualised consideration		★	★	★	★
Factor 5: Contingent reward				★
		⇨		⇨	⇨
Factor 6: Management-by-exception
Factor 7: Laissez-faire

Note. Star symbol (★) represents the case where the CDD KPI score was highest in W – S among the four situations (S – S, S – W, W – S, W – W) and also statistically significant. Arrow symbol (⇨) represents the case where the CDD activities continued most actively in W-S among the four situations (S – S, S – W, W – S, W – W) and also statistically significant. Highlighted cells represent the diagonal where weak leadership and strong followership of the same factors meet.

The results of the analysis can help answer the research questions. The first and second research questions, which asked about the impact of W – S on team effectiveness during and after the CDD program, found that W – S can be even better for CDD performances than S – S and S – W. Table 4 shows many cells filled with star and arrow symbols, which means that many cases exist in which weak leadership was significantly complemented by strong followership either during or after the CDD program. Paying attention to the diagonal on which weak leadership and strong followership of the same factors meet, it can be seen that strong followership with respect to some factors may best complement weak leadership with respect to the same factors for CDD performances either during or after the CDD program. Even in the non-diagonal cells, some other factors of followership still appear to complement leaders’ weaknesses. This finding indicates that in cases in which a leader lacks some leadership qualities, the leader can be complemented by followers who are skilled in any of the various factors.

Most notable about RQ1 and RQ2 is the observation that the two types of symbols (star and arrow) are in separate cells. In other words, the successful W – S combinations during the CDD program (star-symbol cells) seem to be quite different from those after the CDD program (arrow-symbol cells). The answers to RQ1 and RQ2 can be better specified by answering RQ3, which is about the specific conditions through which weak leadership can be effectively complemented by strong followership. The findings in Table 4 can be interpreted more deeply by focusing separately on the W – S conditions behind the successful achievement of CDD and the sustainability of CDD.

When it comes to the W – S combinations for successful achievement during the CDD, which are indicated by a star (★), the effect of complementary followership is most noticeable for weak leadership in terms of the individualised consideration factor. Table 4 shows that weak leadership in terms of individualised consideration combined with strong followership with respect to four factors (inspirational motivation, intellectual stimulation, individualised consideration, and contingent reward) led to more successful achievements during the CDD program. It may mean that even when a leader lacks individualised consideration for followers, team effectiveness can be better achieved as long as the followers are good at (self) supervising in terms of four attributes: (1) visionary (inspirational motivation), (2) innovative (intellectual stimulation), (3) inclusive (individualised consideration), and (4) incentivising (contingent reward).

However, such effects of complementary followership do not hold after the CDD program ends. As the arrow symbols (⇨) in Table 4 show, after the CDD program (when the external supports for CDD are terminated and each village should become self-sustaining), the successful W – S combinations changed significantly. For sustainable CDD activities, weak leadership in inspirational motivation needs to be supported by strong followership in terms of five factors: (1) idealised influence, (2) inspirational motivation, (3) intellectual stimulation, (4) individualised consideration, and (5) contingent reward. When the CDD program ends, more monetary or institutional support from outside villages will cease. In that case, each village should determine what the CDD activities require to continue on their own (and how to do so); therefore, how well the CDD spirit is internalised by villagers will be tested after the CDD program. In short, for the sustainability of CDD activities, villagers need to be motivated not by external support but by their own future visions for their villages. Such patterns can also be observed in weak leadership as a contingent reward. After the CDD program is complete (when a leader cannot guarantee material rewards to followers anymore), the continuity of CDD activities is possible as long as the followers are sufficiently visionary (inspirational motivation), inclusive (individualised consideration), and incentivising (contingent reward).

Summarising the aforementioned findings, we have two groups of W – S combinations that make a difference during and after the CDD program (as represented by star and arrow symbols). The first group of W – S represents a weak leader at an individual consideration level and followers good at self-supervising during the CDD program, and the second group represents a leader who is weak at inspirational motivation and followers who are good at self-motivating after the CDD program.

This pattern of W – S is also observed in the intellectual stimulation dimension, where weak leadership in intellectual stimulation can be effectively supplemented by different types of strong followership at different times. In specific, during the CDD program, a leader who is weak at intellectual stimulation can be complemented by followers who are strong at self-supervision (in terms of individualised consideration and contingent reward); however, after the completion of the CDD program, strong followers with respect to self-motivation (in terms of idealised influence and inspirational motivation) can better complement leaders who are weak at intellectual stimulation.

In addition, three leadership factors exist for which weak leaders do not appear to be complemented by any type of strong followership: (1) idealised influence, (2) management by exception, and (3) laissez-faire. First, when a leader lacks idealised influence, the followers’ complementary effects do not exist during or after the CDD program. Such a limited impact of strong followership for a leader’s weak idealised influence suggests that the leader’s trustworthiness (and followers’ respect and trust in their leader) cannot be replaced by any type of strong followership either during or after the CDD program.

Second, strong followership in management by exception can hardly complement weak leadership in all factors both during and after the completion of the CDD program. From the leader’s perspective, weak leadership in management by exception is seldom complemented by strong followership in all factors. Management by exception implies that providing members with a minimum job requirement is the unique job of leaders, not followers. However, it does not mean that strong leadership in management by exception can naturally lead to team effectiveness.

Last, as with management-by-exception, weak leadership in laissez-faire is seldom complemented by strong followership. However, the interpretation of the laissez-faire factor is somewhat tricky because strength in the laissez-faire factor actually means a weak propensity to directly influence others (and weak laissez-faire means strong influence on others). Therefore, weak leadership in laissez-faire that cannot be complemented by strong followership may imply that a leader’s strong propensity to control followers would adversely affect the achievement and sustainability of the CDD.

Discussion

Does leadership matter for team performance? If so, even when we cannot afford to have strong leadership, is there any room for team success? Despite the widely known beliefs and arguments that a leader is crucial for team effectiveness, having excellent leadership in an organisation is easier said than done. Rather, having unrealistically high expectations of leadership may result in disappointment and distrust of a leader. Therefore, instead of relying too heavily on a leader’s characteristics, we may want to consider many possibilities for leadership substitutes, such as well-defined tasks, cohesive groups, and competent followers (Avolio et al., 2009; Kerr & Jermier, 1978). Among such substitutes for leadership, this study focuses more on the importance of effective followership that can complement leadership.

The analyses in this study provide several discussion points on the importance of followership. First, W – S seems to be one of the necessary (even if not sufficient) conditions for successful team effectiveness. This finding implies that weak leadership can (not must) be an opportunity for manifesting and strengthening “complementary followership”, which will contribute to the leadership literature. It is widely known that empowering leadership can help followers learn and develop (Lorinkova et al., 2013). Weak leadership, for which empowerment can be done unintentionally, leaves space for being complemented by strong followership. In the context of the CDD, the program activities need to be conducted on a followers-driven basis, so it might be natural to observe such complementary relationships between leaders and followers. In other words, CDD programs are usually designed to promote the whole community’s engagement, and the process and outcomes of CDD are evaluated and rewarded according to the grass-roots efforts of the community. Thus, participants (especially common villagers and followers) in a CDD project will be mobilised by intrinsic and extrinsic motivations, which can generate an environment conducive to fostering complementary followership. In that sense, weak leaders may expect strong followers not only to follow the leaders but also to lead the leaders (becoming complementary followers). As a leader’s humility can beneficially influence followers’ collective behaviour (Owens & Hekman, 2015; Rego et al., 2017), followers can learn and grow by helping leaders (Yukl, 2013). In short, weak leaders can foster the followers’ leadership by being led by them intentionally and unintentionally.

Second, beyond the importance of followership, the fit of leadership and followership also matters, as argued in the existing literature (Tepper et al., 2018). For instance, beyond just passively helping leaders, followers can lead and help a leader in trouble through their initiative-taking behaviours (Wee et al., 2017). Authentic leadership accompanied by authentic followership can also lead to better follower performance (Leroy et al., 2012). Generally speaking, the findings of this study imply interchangeable roles among leader – follower positions. A team basically consists of formal (official) positions, namely, managers and subordinates. In reality, however, each formal position usually holds a (predetermined or emerging) informal role as either leader or follower. Considering this independent association between formal positions and informal roles, the substantive or actual roles of leadership can be (and sometimes should be) shared among managers and subordinates, especially when the manager’s leadership is weak with respect to certain leadership factors.

In this study, such interchangeable (informal) roles among (formal) leader – follower positions appear to differ according to the timing. Although the three leadership factors—idealised influence, inspirational motivation, and individualised consideration—are part of the transformational leadership style, each has a different pattern of follower complementarity. Specifically, what matters during the implementation of a CDD program is how leadership weak in individualised consideration can be complemented by followers’ strengths in self-supervision. On the other hand, what makes a difference for the sustainability of CDD is how leadership weak in inspirational motivation is supplemented by strong followership with respect to self-motivation. In addition, it is noteworthy to mention that leadership weak in idealised influence is not replaceable by any type of strong followership both during and after the CDD program, which means that no substitute exists for trustful and respected leadership, which is a unique and non-transferable role of the leader position. Still, such conditionality of the W – S combination’s beneficial impact on team effectiveness implies that W – S alone does not guarantee a good team performance. In other words, we need to be cautious in generalising about the positive effect of W – S, because such paradoxically good consequences of the W – S combination were, according to the analyses, limited to several aspects of leadership and followership among all MLQ factors.

Third, in contrast to the positive role of strong followership, even strong leadership can be detrimental to team effectiveness, as presented with the negative coefficients in Appendix 4. On several occasions, S – W (strong leadership and weak followership) seems to be somewhat associated with the worst performance during the CDD program. For instance, when strong leadership with respect to Factor 2 (inspirational motivation) and Factor 6 (management by exception) is paired with weak followership, the CDD performance turns out to be even worse than the W – W pair. Some supporting arguments exist for this irony: (1) transformational leaders can be endangered through exhaustion (Lin, 2019), (2) charismatic leaders can impair team performance through stressful demands on followers (Lepine et al., 2016), and (3) leaders’ excessive self-interest can lead to an exploitative leadership (Lorinkova et al., 2013; Schmid et al., 2019). In short, strong leadership may result in two events: (1) ignoring followers’ expectations, qualifications, or preparedness and (2) widening the gap between leader and followers in terms of vision setting/sharing and mobilisation of tangible/intangible resources. Such a gap would drive the loss of cohesive teamwork and team effectiveness.

Conclusion

The results from the analysis can be interpreted with respect to some practical implications in two ways: (1) from the “failure management” perspective, which indicates managing the bright side of organisational adversities (Lee & Miesing, 2017), weak leadership can be revalued as an opportunity to make the most of followers’ potentiality. Such paradoxical benefits of weak leadership might be possible only when the followers are physically and mentally prepared to take up the complementary followership role and (2) the “success management” perspective focuses on how to deal with the dark side of organisational strengths (Lee & Lee, 2018); “success management” exemplifies strong leadership that can be a double-edged sword because of its possible negative impact. Leader strengths should be prudently assessed and handled, because they can blind the leader through hubris or overconfidence and eventually lead to a loss of balanced leadership (Kaplan & Kaiser, 2013).

Considering the limitations of this study, several issues should be addressed in the future. First, regarding space-time boundaries, this study was conducted using a rural development program in an Asian country context. Therefore, more studies need to be conducted for other kinds of team projects in various spatial and temporal settings. Second, from the perspective of research themes, more research questions are worth pursuing through further studies: for instance, (1) What other impacts do the combinations of leadership and followership have? and (2) What are the similarities and differences between (good) leadership and (good) followership? Third, regarding the measurement of variables, two separate sets of a measurement instrument for leadership and followership, other than MLQ, need to be employed in order to capture not only the common but also the unique and distinctive characteristics of leadership and followership.

Acknowledgements

The authors appreciate the research funding from KDI School of Public Policy and Management. Lee and Kim gratefully acknowledge that this research was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2016S1A3A2924956 and NRF-2018S1A3A2075117, respectively). We are grateful to the excellent field survey administered by Myanmar Survey Research (MSR).

Disclosure statement

No potential conflict of interest was reported by the authors.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

The work was supported by the National Research Foundation, Republic of Korea [NRF-2016S1A3A2924956].

Notes on contributors

Junesoo Lee

Junesoo Lee is an associate professor at KDI School of Public Policy and Management.263 Namsejong-ro, Sejong, Republic of Korea. [email protected]

Jongwoo Chung

Jongwoo Chung an economist at Micro & Institutional Economics Team, Bank of Korea.3 Namdaemun-ro, Jung-gu, Seoul, Republic of Korea. [email protected]

Booyuel Kim

Booyuel Kim is an associate professor at Department of Environmental Planning, Graduate School of Environmental Studies, Seoul National University.1 Gwanak-ro, Gwanak-gu, Seoul, Republic of Korea. [email protected]

Appendix 1.

Descriptive Statistics on Village Members

Table

Variable	Mean	Std. Dev.	Min	Max
Panel A. CDD Program Leader (n = 100)
Age	52.4	11.4	22	73
Female ( = 1)	0.010	0.100	0	1
Belong to the Bamar ethnic group ( = 1)	0.850	0.359	0	1
Buddhist ( = 1)	0.960	0.197	0	1
Married ( = 1)	0.910	0.288	0	1
>High school graduate ( = 1)	0.160	0.368	0	1
Income (Kyat, monthly)	1,053,301	1,073,308	90,000	5,350,000
Number of Assets	7.9	1.5	5	12
Panel B. Villagers (n = 1,996)
Age	48.0	13.4	18	86
Female ( = 1)	0.499	0.500	0	1
Belong to the Bamar ethnic group ( = 1)	0.848	0.359	0	1
Buddhist ( = 1)	0.970	0.171	0	1
Married ( = 1)	0.744	0.436	0	1
>High school graduate ( = 1)	0.043	0.203	0	1
Income (Kyat, monthly)	503,546	737,658	1,250	15,900,000
Number of Assets	5.9	2.0	1	11
Panel C. Village (n = 100)
Number of households	199.7	88.0	49	449
% of household heads being farmer or fisherman	0.486	0.188	0.089	0.946
% of household heads being self-employed	0.101	0.066	0.000	0.321
% of household heads being wage-worker	0.157	0.105	0.000	0.429
% of households being able to use electricity	0.647	0.310	0	1
Have a public primary school	0.840	0.368	0	1
Have a public high school established	0.050	0.219	0	1
Have a health clinic, health centre or hospital	0.400	0.492	0	1

Appendix 2.

Estimates of Interrater Agreement (IRA) Indices

Table

Index	MLQ	Mean	Std. Dev.
$r_{\mathit{WG}\left(\mathit{J}\right)}$	Factor 1: Idealised influence	0.909	0.053
	Factor 2: Inspirational motivation	0.944	0.070
	Factor 3: Intellectual stimulation	0.979	0.072
	Factor 4: Individualised consideration	0.948	0.072
	Factor 5: Contingent reward	0.968	0.069
	Factor 6: Management-by-exception	0.976	0.061
	Factor 7: Laissez-faire	0.945	0.063
$\mathit{A}{D}_{\mathit{M}\left(\mathit{J}\right)}$	Factor 1: Idealised influence	0.792	0.161
	Factor 2: Inspirational motivation	0.910	0.170
	Factor 3: Intellectual stimulation	1.031	0.178
	Factor 4: Individualised consideration	0.926	0.183
	Factor 5: Contingent reward	0.984	0.186
	Factor 6: Management-by-exception	0.992	0.134
	Factor 7: Laissez-faire	0.864	0.183
$\mathit{A}{D}_{\mathit{Md}\left(\mathit{J}\right)}$	Factor 1: Idealised influence	0.677	0.180
	Factor 2: Inspirational motivation	0.827	0.213
	Factor 3: Intellectual stimulation	0.962	0.214
	Factor 4: Individualised consideration	0.842	0.226
	Factor 5: Contingent reward	0.893	0.230
	Factor 6: Management-by-exception	0.884	0.162
	Factor 7: Laissez-faire	0.687	0.208

$r_{\mathit{WG}\left(\mathit{J}\right)}$is the index suggested by James et al. (1984).

$\mathit{A}{D}_{\mathit{M}\left(\mathit{J}\right)}\mathit{\hspace{0.17em}}$is the index which is calculated using the mean value (M. J. Burke et al., 1999).

$\mathit{A}{D}_{\mathit{Md}\left(\mathit{J}\right)}$ is the index which is calculated using the median value (M. J. Burke et al., 1999).

J represents the number of items. As there are three items for each MLQ factor, J equals 3.

Appendix 3.

Mean values of MLQ factors

Table

MLQ factors		Leadership-followership combinations
		W-S	S-W	S-S	W-W
Factor 1: Idealised influence	Leader	10.250	12.000	12.000	9.800
	Follower	10.298	9.232	10.477	9.279
Factor 2: Inspirational motivation	Leader	9.842	12.000	12.000	9.917
	Follower	9.722	8.254	9.967	8.224
Factor 3: Intellectual stimulation	Leader	8.625	11.750	11.583	8.400
	Follower	9.374	7.105	9.252	7.020
Factor 4: Individualised consideration	Leader	8.667	11.703	11.800	8.889
	Follower	10.122	8.200	9.865	8.072
Factor 5: Contingent reward	Leader	9.222	11.677	11.656	8.368
	Follower	9.842	7.709	9.642	7.939
Factor 6: Management-by-exception	Leader	8.545	11.667	11.710	8.517
	Follower	9.590	8.228	9.545	8.246
Factor 7: Laissez-faire	Leader	5.800	11.320	11.326	6.947
	Follower	10.328	9.193	10.416	9.092

Appendix 4.

Impacts of the Combinations of Leadership and Followership on CDD Activities’ Evaluation Score during CDD Program

Table

Leadership		Followership
		Factor 1: Idealised influence		Factor 2: Inspirational motivation		Factor 3: Intellectual stimulatoin		Factor 4: Individualised consideration		Factor 5: Contingent reward		Factor 6: Management-by-exception		Factor 7: Laissez-faire
		Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak
Factor 1: Idealised influence	Strong	6.857	1.632	9.227	0.648	10.027	−0.931	12.399	−0.731	11.034	0.860	5.921	−4.323	3.757	−1.144
	Weak	7.557		8.648		8.028		12.535		14.529		3.100		2.373
Factor 2: Inspirational motivation	Strong	3.676	−5.344	6.050	−2.146	6.420	−5.775	9.057	−6.881	8.143	−5.214	3.673	−7.116	0.976	−4.369
	Weak	3.028		9.690		6.353		9.960		12.941		1.835		1.114
Factor 3: Intellectual stimulation	Strong	3.333	−1.512	6.055	1.587	6.900	0.257	9.592	−0.589	8.449	−0.816	3.415	−2.188	0.661	−1.248
	Weak	7.207		13.048		12.563		15.779*		16.307*		7.187		5.146
Factor 4: Individualised consideration	Strong	7.857	4.955	9.557	4.906	11.538	5.527	13.006*	3.317	11.112	1.445	6.130	1.322	2.188	−2.579
	Weak	12.951		16.724*		17.153*		19.258**		18.117*		10.429		1.574
Factor 5: Contingent reward	Strong	4.502	−3.083	7.988	3.480	9.576	3.515	12.477	4.604	11.314	2.036	6.383	4.698	2.242	−0.903
	Weak	3.966		15.695		15.86		21.557**		17.512		15.099		5.000
Factor 6: Management-by-exception	Strong	5.275	2.798	7.281	−7.593	9.073	−3.961	10.825	−3.216	9.323	−6.696	4.894	−1.719	2.489	−0.158
	Weak	10.301		2.870		6.114		12.120		9.940		7.567		5.697
Factor 7: Laissez-faire	Strong	5.208	−3.367	4.419	−11.167	6.931	−6.935	10.228	−4.706	8.119	−8.491	4.179	−5.425	4.346	2.068
	Weak	0.756		−9.339		−1.359		8.195		2.246		−1.458		7.654

Note: The numbers represent the coefficients of each of the three states (S – S, S – W, W – S) in the regression analysis where the three states are dummies while the W – W is omitted as the reference. * p < 0.1, ** p < 0.05, *** p < 0.01. The adjusted p-values were used to control for the familywise error rate.

Appendix 5.

Impacts of the Combinations of Leadership and Followership on CDD Activities’ Continuation after CDD Program

Table

Leadership		Followership
		Factor 1: Idealised influence		Factor 2: Inspirational motivation		Factor 3: Intellectual stimulatoin		Factor 4: Individualised consideration		Factor 5: Contingent reward		Factor 6: Management-by-exception		Factor 7: Laissez-faire
		Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak	Strong	Weak
Factor 1: Idealised influence	Strong	0.132	0.043	0.125	0.079	0.120	0.042	0.139	0.019	0.115	0.038	0.083	0.011	0.087	0.101
	Weak	0.107		0.138		0.098		0.107		0.116		0.050		0.110
Factor 2: Inspirational motivation	Strong	0.188*	0.177*	0.177*	0.195**	0.169*	0.168	0.175*	0.152**	0.166**	0.184*	0.158*	0.178	0.166*	0.165*
	Weak	0.202*		0.225**		0.197*		0.224*		0.227**		0.157		0.122
Factor 3: Intellectual stimulation	Strong	0.194*	0.172*	0.176*	0.186**	0.166	0.138	0.176*	0.122	0.154	0.138	0.142	0.119	0.121	0.077
	Weak	0.221*		0.234**		0.180		0.197		0.198		0.131		0.071
Factor 4: Individualised consideration	Strong	0.121	0.050	0.118	0.086	0.118	0.069	0.127	0.049	0.109	0.063	0.071	0.032	0.023	−0.050
	Weak	0.149		0.213		0.172		0.19		0.211		0.115		−0.018
Factor 5: Contingent reward	Strong	0.185	0.159	0.154	0.141	0.159	0.128	0.176	0.128	0.165	0.139	0.115	0.092	0.073	0.013
	Weak	0.228		0.225*		0.199		0.236**		0.228*		0.130		0.013
Factor 6: Management-by-exception	Strong	0.145	0.049	0.131	0.043	0.145	0.045	0.153*	0.017	0.125	0.040	0.099	0.056	0.071	−0.039
	Weak	0.094		0.084		0.078		0.084		0.097		0.069		−0.038
Factor 7: Laissez-faire	Strong	0.130	0.050	0.114	0.040	0.149	0.107	0.132	0.043	0.111	0.034	0.107	0.085	0.043	−0.033
	Weak	0.130		0.126		0.195		0.187		0.154		0.172		−0.058

Note: The numbers represent the coefficients of each of the three states (S – S, S – W, W – S) in the regression analysis where the three states are dummies while W – W is omitted as the reference. * p < 0.1, ** p < 0.05, *** p < 0.01. The adjusted p-values were used to control for the familywise error rate.