Coproducing leadership: a reason to resist destructive leaders

Abstract

Purpose

Follower's individual differences have been receiving increased attention in studying destructive leadership because followers may enable or disable it. One of these yet under-researched features is the role of followers' leadership coproduction beliefs (a role construal) in explaining their resistance to destructive leaders. Departing from the proactive motivation theory, this paper explores the robustness of coproduction beliefs by testing its ability to predict followers' resistance to destructive leaders across four situations – abusive supervision, exploitative leadership, organization directed behaviors and laissez-faire.

Design/methodology/approach

With a sample of 359 participants that answered a scenario-based survey, the present study tests the relationship between coproduction beliefs and resistance behaviors in the four mentioned groups, while controlling for alternative explanations. A multigroup analysis was conducted with PLS-SEM.

Findings

Constructive resistance is always favored by coproduction beliefs independently of the leader's type of destructive behavior. Dysfunctional resistance, however, is sensitive to the leader's type of destructive behavior.

Originality/value

This paper extends knowledge on the role of coproduction beliefs as an individual-based resource against destructive leaders.

Citation

Almeida, T., Ramalho, N. and Esteves, F. (2023), "Coproducing leadership: a reason to resist destructive leaders", Leadership & Organization Development Journal, Vol. 44 No. 2, pp. 189-204. https://doi.org/10.1108/LODJ-09-2021-0427

Introduction

Research on the dark side of leadership has been gaining momentum (Zhu et al., 2019), in part because of its nefarious consequences (Mackey et al., 2021; Schyns and Schilling, 2013). For a long time, the focus of destructive leadership was single placed on leaders (Thoroughgood et al., 2018). However, followers, are now considered an intrinsic part of the leadership process (for a review, see Uhl-Bien et al., 2014) and play an active role in trying to curb destructiveness (Almeida et al., 2021; Wee et al., 2017). There are many ways for followers to express disapproval about the leader's behavior (e.g. resistance, voice and retaliation), and there are differences between them (Brett et al., 2016). This study focuses on resistance behaviors (Tepper et al., 2001), and places special attention on constructive resistance as an expression of ethical followership (Carsten and Uhl-Bien, 2013).

Research has uncovered some resistance roots (Brett et al., 2016; Greenbaum et al., 2013; Tepper et al., 2001). Among these, followers' beliefs on the coproduction of leadership (Carsten and Uhl-Bien, 2012) are attracting researchers' attention. Although the relationship between coproduction and constructive resistance has been established (Carsten and Uhl-Bien, 2013), this study goes a step further as it examines the role of coproduction in light of Parker et al. (2010) model of proactive motivation. According to this model, efforts to produce work-related changes are driven by three motivational states: “can do”, “energized to” and “reason to”. Whereas the first and second are more volatile and context-dependent, the latter may have a more stable nature as it is linked to the individual's sense of personal responsibility (Curcuruto et al., 2019). Accordingly, we propose coproduction beliefs as a “reason to” motivation and study its role in explaining both constructive and dysfunctional resistance while controlling for affect (“energized to”) and personal sense of power (“can do”). We contend these beliefs gain relevance when analyzed within a context where a leader displays destructive behaviors, because holding a sense of participation in the leadership process can be conceived as an important individual-level organizational resource.

Destructive leadership studies are known both to show a lack of clarity in operationalizing destructiveness (Schyns and Schilling, 2013; Tepper, 2007) and to remain focused on a single sort of profile (e.g. abusive supervision). However, previous research shows that different destructive leader's behaviors lead to different followers' reactions (Schmid et al., 2018). Accordingly, we designed an experimental study to test whether coproduction beliefs explain resistance across different destructive profiles (abusive, exploitative, organization directed and laissez-faire).

Literature review

Followers' resistance

Destructive leaders' behaviors lead to many consequences related to the job, the organization, the individual and the leader (Schyns and Schilling, 2013). Among these consequences, followers' actions that help to balance power in the leader-follower relationship can be considered especially relevant as they may contribute to change (Wee et al., 2017).

Tepper et al. (2001) introduced the study of followers' resistance mechanisms. The authors defined two resistance strategies: Constructive resistance has a negotiation-based nature, while dysfunctional resistance comprises passive-aggressive responses. Constructive resistance expresses nonconformity and dysfunctional resistance is retaliatory, aiming to re-establish a sense of justice. The latter contributes to a conflict spiral that can be stopped using constructive resistance behaviors.

Both strategies lead to important consequences to the individual as well as to the job, and the leader. Accordingly, dysfunctional resistance is negatively associated with performance and job satisfaction (Vecchio et al., 2010) and positively related to the leader's emotional exhaustion, while constructive resistance results in lower levels of the leader's emotional exhaustion (Brett et al., 2016). However, these relationships are not context-free and Tepper et al. (2006) show that the leader–follower relationship plays a critical role in how leaders perceive resistance. Another study reveals that showing dysfunctional resistance to a leader that displays downward hostility, lessens negative consequences (Tepper et al., 2015).

Supervisors' abusive behaviors are an important antecedent of resistance behaviors (Haggard and Park, 2018; Tepper et al., 2001), and it is enough to witness abuse to activate resistance (Greenbaum et al., 2013). There are, additionally, follower-related features explaining resistance behaviors: goal orientation is positively associated both to constructive and dysfunctional resistance, and learning goal orientation is positively related to constructive resistance and negatively related to dysfunctional resistance (Brett et al., 2016); aggressive humor is associated to dysfunctional resistance (Goswami et al., 2015); low levels of consciousness and agreeableness strengthen the relationship between abusive supervision and dysfunctional resistance, while high levels of consciousness heighten the relationship between abusive supervision and constructive resistance (Tepper et al., 2001), which also occurs when followers have high levels of moral identity (Greenbaum et al., 2013); coproduction beliefs are positively associated to constructive resistance (Carsten and Uhl-Bien, 2013). Based on what has been described, one can assert that individual differences, most of them ethic-based, play a key role in routing different resistance types.

How do followers resist? The role of coproduction beliefs

When trying to change something, people need to have a reason to engage and persist in risky behaviors (Urbach et al., 2020) such as resistance (Carsten and Uhl-Bien, 2012). In line with this, coproduction beliefs deserve special attention because they inform about followers' commitment to leadership, which becomes critical when accounting for what makes followers resist or defer a destructive leader (Carsten and Uhl-Bien, 2013).

Shamir (2007) introduced the idea of coproduction referring to the process of leadership being actively developed by both the leader and followers as active contributors. This approach marks an important turning point in the study of followership and highlights the importance of followers own role perception (Carsten et al., 2010; Knoll et al., 2017). Uhl-Bien et al. (2014) present a role-based approach where the followers' role is influenced by the way they perceive their duties and responsibilities. If they believe they should actively contribute to leadership together with the leader, then they hold coproduction beliefs. These are neither a state nor a trait (Carsten et al., 2018) but a cognitive mechanism – role construal – representing how individuals build different social roles (Biddle, 1986; Vial et al., 2021). Research has been showing that individual differences in role construal explain how individuals behave in the organizational context (Zellars et al., 2002). Among these studies, Simpson and Laham (2015) showed that individual differences in moral role construal (cognitive moral motives) are linked to moral judgment. This finding provides evidence on the cognitive roots of ethical decision-making, encouraging the study of coproduction beliefs as an individual difference. Accordingly, as predicted by the Theory of Planned Behavior (Ajzen, 1991), these beliefs underlie followership attitudes that will help predict followers' behaviors. Carsten and Uhl-Bien (2012) operationalized coproduction beliefs and research has been showing these cognitive construals explain constructive resistance and voice (Carsten and Uhl-Bien, 2013; Carsten et al., 2018), which can be conceived as problem solving strategies (Mitchell and Ambrose, 2012), contrasting with retaliatory behaviors (e.g. Skarlicki and Folger, 2004) such as dysfunctional resistance (Tepper et al., 2001, 2015).

Resistance can be thought of as a mechanism that allows the expression of discontentment. Parker et al. (2010) introduced a motivation model whereby self-efforts to change the work environment depend on the activation of three motivational states: “reason to”, “can do”, and “energized to”. Finding a “reason to” is related to the individual's will to engage in certain behaviors. The “can do” driver is linked to the perception of self-efficacy, cost and control. Feeling “energized to” relates to the emotional boost underlying proactive behaviors. Both “can do” and “energized to” can be conceived as more context-dependent as they result from a situation's assessment (i.e. “can I?”/“Is it risky?” and “how do I feel?”). Conversely “reason to” refers to a stable individual difference as it is related to the individuals' integrated and identified motivation (Deci and Ryan, 2000) and sense of personal responsibility toward achieving a goal related to a constructive change (Curcuruto et al., 2019; Lebel and Patil, 2018; Parker et al., 2010). Moreover, as highlighted by Parker and Wang (2015), individuals may feel they are able to do something but will not do anything unless they have a reason to. This draws our attention to how followers define their roles, and the specific contribution of a moral role construal (Simpson and Laham, 2015), and is in line with previous research showing that beliefs may operate as dispositional variables helping to predict compliance with unethical requests (Blass, 1991). In light of this, we focus our attention on the “reason to” driver (coproduction beliefs), as we aim to understand individual-based differences that make people resist destructive leaders (Carsten and Uhl-Bien, 2013). According to the same authors, followers who hold deferential roles will show compliance with unethical requests, while those who report higher levels of coproduction beliefs engage in constructive resistance. However, it is timely to question whether these beliefs are able to guide followers' behaviors regardless of the nature of the leaders' destructive behaviors, i.e. it is yet unclear how independent are coproduction beliefs effects from context.

Destructive leader's behaviors: the role of context

Although it is well-established that destructive leaders have an important impact on followers' behaviors (Schyns and Schilling, 2013), there is a lack of knowledge regarding how the different leader's destructive behaviors impact on followers (for an exception see Schmid et al., 2018). Leaders' destructive behaviors can be classified regarding their intensity and target (Einarsen et al., 2007). According to this classification, and in line with previous research (Schmid et al., 2018) there are three main leaders' destructive profiles: abusive (Tepper, 2000), exploitative (Schmid et al., 2019) and organization directed (Thoroughgood et al., 2012). Schmid et al. (2018) have not considered the controversial laissez-faire behavior, and there is no consensual agreement on whether laissez-faire leadership is destructive or not (Schyns and Schilling, 2013). However, its negative effects (Judge and Piccol, 2004; Skogstad et al., 2007) suggest that destructive behaviors can be displayed in a more active or passive way and a destructive leader can be both related to unethical and ineffective behaviors (Kellerman, 2012).

The universality of coproduction effects has been under scrutiny. Although leaders can condition how followers express their own characteristics (Zhang et al., 2020) and distinct destructive profiles are known to have different consequences for followers (Schmid et al., 2018), coproduction beliefs-as a construal role-are taken to be stable and central in decision making (e.g. Vial et al., 2021), which received empirical support with the Carsten and Uhl-Bien (2012) findings that showed context had no moderating effects on the relationship between coproduction beliefs and constructive resistance. This reinforces the claim that coproduction beliefs, as a relatively stable individual difference, function independently of their context. Thus, the present study builds on the idea that coproduction is an organizational resource nurtured by followers, who use these beliefs as a motivational driver to choose between resistance behaviors.

Constructive resistance is consistent with coproduction beliefs as they foster the willingness to participate. Conversely, dysfunctional resistance expresses the will to undermine the leader (Tepper et al., 2006) and is therefore inconsistent with believing in co-production. We thus hypothesize that:

H1.

There is a positive relationship between coproduction beliefs in leadership and constructive resistance regardless of the type of leader's destructive behavior.

H2.

There is a negative relationship between coproduction beliefs in leadership and dysfunctional resistance regardless of the type of leader's destructive behavior.

The conceptual model is depicted in Figure 1.

Method

Procedure and participants

To test the proposed hypotheses, 359 participants (M_age = 39.95, SD_age = 13.68; 68.2% female) took part in an online experimental study. First participants responded to the coproduction scale and demographic variables. Then, individuals were randomly assigned to one adapted scenario: abusive supervision, exploitative leadership, organization-directed and laissez-faire (Christie et al., 2011; Hughes and Harris, 2017; Kelloway et al., 2003; Schmid et al., 2018). Then, participants answered the remaining scales.

Controlling for alternative explanations

As detailed in the literature review section, three important motivational drivers can help explaining followers' resistance. This study aims to analyze the “reason to” motivation (i.e. co-production), and to make sure we are able to test its impact on followers' behavior, we controlled the other drivers (“can do” and “energized to”). The “can do” comprehends the perception that one is able to attain goals, balancing its risks/feasibility (Urbach et al., 2020). So, individuals must own a sense of self-efficacy that allows them to pursue goals (Parker and Wang, 2015). The leader-follower interaction can be conceived as an interdependent and asymmetric power-based relationship (Collinson, 2005). To balance this relationship, the power gap should be reduced (Wee et al., 2017), and the follower needs to believe he or she can influence the leader. Moreover, it is known that covert behaviors (i.e. present in dysfunctional resistance) may occur due to power imbalance situations (O'Reilly and Aquino, 2011). The personal sense of power is the individual's perception of his or her ability to influence others (Anderson et al., 2012), and, therefore, we asked participants their perceived power to influence the leader. Feeling “energized to” is mostly linked to positive affect (Parker et al., 2010), especially activating positive affect (Bindl and Parker, 2010). However, negative affect is also informative as it is known to make people display defensive behaviors (Fredrickson, 2001), which may be critical to understand dysfunctional resistance. Accordingly, deactivating negative affect may inhibit risky behaviors (Urbach et al., 2020). For this reason, we asked participants how they would feel at work, having the described supervisor.

Measures

Destructive leadership. Abusive supervision was measured using three items from Tepper (2000) scale (α = 0.86; e.g. Puts me down in front of others). Exploitative leadership was assessed with three items from Schmid et al. (2019; α = 0.84; e.g. Sees employees as a means to reach his or her personal goals). Organization directed behaviors were measured with three items from Thoroughgood et al. (2012) anti-organizational behaviors subscale (α = 0.89; e.g. Violates company policy/rules) and laissez-faire leadership was assessed with Avolio et al. (1999) subscale (α = 0.87; e.g. e.g. Avoids deciding).

Co-production of leadership was assessed with four items from Carsten and Uhl-Bien (2012) scale (α = 0.81; e.g. Followers should communicate their opinions, even when they know leaders may disagree).

The personal sense of power in the relationship with the leader was measured with four items from Anderson et al. (2012) scale (α = 0.82; e.g. Even if I voiced them, my views would have little sway(r)).

Activated positive affect was measured with two items (enthusiastic and inspired) from the positive affect subscale of PANAS (Galinha et al., 2014; Spearman-Brown coefficient = 0.85). Deactivated negative affect was measured with two items (frightened and tormented) from the negative affect subscale of PANAS (Galinha et al., 2014; Spearman-Brown coefficient = 0.75).

Constructive and dysfunctional resistance were measured with four items each from Tepper et al. (2001) scales (α_constructive = 0.92; e.g. I would explain that it should be done in a different way; α_{dysfunctional} = 0.83; e.g. I would disregard what my supervisor says)”.

All items were rated on a six-point Likert scale (1 = strongly disagree to 6 = strongly agree) with the exception of the resistance scale (1 = never and 6 = always).

Control variables

Following previous research, we controlled for age and gender (e.g. Brett et al., 2016; Haggard and Park, 2018). We also controlled “can do” (sense of power) and “energized to” (affect) motives. As the perception of leaders' behaviors is related to followers' resistance (Vecchio et al., 2010), we controlled destructive leadership for each scenario.

Manipulation check

One-way ANOVA and Tukey HSD post-hoc comparisons were conducted to test the manipulation check. There was a significant effect of abusive supervision manipulation [F_(3, 353) = 61.72, p < 0.001] and, abusive supervision rates (M = 5.42, SD = 0.95) were higher than organization-directed (M = 3.61, SD = 1.43), exploitative (M = 2.96, SD = 1.44) and laissez-faire rates (M = 3.63, SD = 1.22). Similarly, the exploitative scenario showed a significant manipulation effect [F_{(3, 355)} = 6.48 p < 0.001] exploitative leadership rates were higher (M = 5.16, SD = 1.04) than abusive supervision (M = 4.43, SD = 1.46), organizational-directed (M = 4.63, SD = 1.42) and laissez-faire rates (M = 4.48, SD = 1.02). There was also a significant effect of the organizational-directed manipulation [F_{(3, 353)} = 72.53, p < 0.001], and organizational-directed rates were higher (M = 5.43, SD = 0.85) than abusive supervision (M = 3.25, SD = 1.44) exploitative (M = 2.89, SD = 1.54) and laissez-faire rates (M = 3.29, SD = 1.29). Finally, there was a significant manipulation effect for the laissez-faire condition [F_{(3, 351)} = 39.76 p < 0.001], laissez-faire leadership rates were higher (M = 5.21, SD = 1.00) than abusive supervision (M = 3.71, SD = 1.42), organizational-directed (M = 3.29, SD = 1.67) and exploitative (M = 3.07, SD = 1.49).

Results

The hypotheses were tested using PLS-SEM via SmartPLS3 software (Ringle et al., 2015). PLS-SEM analysis comprises two steps: the measurement model analysis, and the structural model assessment (Hair et al., 2016).

Measurement model

The scales reliability and convergent validity (Bagozzi and Yi, 1988; Fornell and Larcker, 1981; Hair et al., 2014) are depicted in Table 1. The heterotrait-monotrait ratio (Henseler et al., 2015) corroborated discriminant validity as all values were below 0.85.

Table 2 present the correlations for each condition.

Structural model

Following Hair et al. (2016) recommendations, we started by assessing collinearity. The highest VIF value was 1.46 (total sample and each group), showing no collinearity issues. Then, we obtained the path coefficients to test the hypothesized relationships between reflective latent variables, using bootstrapping. After that, the model's predictive power was assessed (coefficient of determination). To evaluate the impact of each exogenous variable, the effect size was calculated. To close the structural model analysis, the model's predictive relevance was tested by examining the Stone-Geisser's Q² value. The described results are depicted in Table 3.

Multigroup analysis

To test the moderation effect of the conditions, we conducted a multigroup analysis. Measurement invariance was assessed through the MICOM three-step procedure (Henseler et al., 2016). Configural invariance is assumed as the same factor structure is represented in all groups. Compositional invariance was also attained as indicators have equal weights across groups. Once configural and compositional invariance have been established, we are able to assume partial measurement invariance that allows group comparisons. Although analyzing the pooled data was not our goal, we conducted the last step in invariance testing that assesses equality between mean values and variances. Our models failed this step, so we kept our initial purpose on not interpreting the global results.

After establishing partial invariance, we used permutation testing (Hair et al., 2017) to determine if the model's paths were different between the four groups. Significant differences (p < 0.05) were found only for dysfunctional resistance and between two cases: exploitative versus laissez-faire (Δβ = 0.31, p < 0.05), and organizational versus laissez-faire (Δβ = 0.38, p < 0.05).

Discussion and conclusion

This research aims to contribute to the study of coproduction as an individual-based organizational resource. Our results support the robustness of coproduction beliefs in explaining constructive resistance, as this effect occurs regardless of the leader's destructive behavior. Although the magnitude of the effects are mostly modest in size, they are in line with previous research (Carsten and Uhl-Bien, 2012). Moreover, there is no difference in the effects across groups. Unexpectedly, however, these beliefs reduce dysfunctional resistance only in the abusive and laissez-faire groups. We will discuss these findings while highlighting three ways in which this study extends previous research.

First, we studied the impact of followers' coproduction beliefs on resistance behaviors within a conceptual framework that explains proactive behaviors at work (Parker et al., 2010). Even when controlling for alternative explanations (“can do” and “energize to”) coproduction does make a difference, especially for constructive resistance. Second, by examining this relationship across four scenarios, we answer the call for more conceptual clarification as regards the types of destructive leader behaviors in line with Schmid et al. (2018) while testing the robustness of these beliefs (i.e. its universal nature). Our findings support the claim that differences in coproduction beliefs define ethical followership (Carsten and Uhl-Bien, 2013). Finally, extending previous research (Carnevale et al., 2018; Greenbaum et al., 2013) this study analyzed both dysfunctional and constructive resistance. Studying both types is useful as these behaviors hold divergent natures (Tepper et al., 2001) but are not mutually exclusive (Haggard and Park, 2018). Having found no association between constructive and dysfunctional resistance (except for the abusive condition), it seems unlikely one can infer one from the other. Against expectation, coproduction beliefs do not play a role in both exploitative and organization-directed conditions (when explaining dysfunctional resistance). According to Schmid et al. (2018), these two styles convey low hostility when compared with abusive supervision. Thus, hostility may activate the expression of coproduction beliefs in decreasing dysfunctional resistance. We should highlight that the items assessing dysfunctional resistance have a subversive nature mirroring passive resistance (Almeida et al., 2021) and therefore: when high hostility (abusive supervision) or no leading behavior (laissez-faire) is observed, believing in one's own active role as a follower reduces chances of behaving in a concealed manner. Hence, when destructive behaviors are more adverse, followers need to go subversive. For the laissez-faire case, follower who hold coproduction beliefs will not need to resist as they most likely will take the leading role in the relationship.

Limitations and future research

Findings must be interpreted at the light of the limitations of this study. First, it is a scenario-based research assessing behavioral intentions. However, this is a widely used approach in organizational studies, especially in ethics-related topics (e.g. Cianci et al., 2014; Tseng, 2019), and it is presented as a valuable alternative research strategy where large samples are not required (Ehrhart and Naumann, 2004).

Another limitation concerns the four scenarios, which do not allow room for mixed contexts as well as the occurrence of opposing behaviors. However, destructive behavior may damage the organization while protecting followers and vice-versa (Einarsen et al., 2007). Therefore, future research could explore resistance behaviors using scenarios that include these variations.

Another limitation concerns the exclusive individual-level focus of this study. Although findings are informative on the relevance of coproduction for constructive resistance regardless of the leader's destructive profile, the role of team as context may be considered in future research, e.g. by testing coproduction beliefs against group conformity.

Implications

Despite limitations, findings from this study offer a novel view and more extensive insight on resistance behaviors and coproduction by bridging conceptual frameworks (role theory and motivation for proactive behavior) and testing its stable nature. We trust our findings also have practical relevance by offering guidance for organizational design. Decision-makers interested in preventing destructive leadership may conclude that some of the strategies inspired in extant literature may not suffice. Namely, uncovering the roots of leaders' destructive behaviors (e.g. Zhang and Bednall, 2016) so that leaders become aware of such behaviors (Goswami et al., 2015), and acknowledge and apologize (Basford et al., 2014; Haggard and Park, 2018) to rebuild the relationship with followers. This may explain why destructive behaviors are prevalent (Aasland et al., 2010), and organizations do attract destructive profiles to leading positions (Chamorro-Premuzic, 2013). Therefore, a shift in focus – i.e. giving stage to followers instead of perpetuating a leader-centric perspective – may offer novel solutions for this enduring problem. Organizations should counter the widespread negative connotation associated with the word follower and should formally acknowledge followers' agency and influence in the organizational dynamics. A step toward this can be precisely the adoption of the coproduction of leadership concept. From an organizational perspective, this allows employing governance mechanisms that endorse co-determination, thus avoiding power centralization (Crane et al., 2004; Padilla et al., 2007). From the follower's standpoint, the formal recognition of their agency role may give more latitude for performing different behaviors, namely those that express discontent (e.g. resistance). As evidenced by the present study, coproduction beliefs can foster these behaviors. Although cognitive beliefs hold a stable nature, it is also known that changing or adjusting them is possible (Albarracin and Shavit, 2018). Accordingly, organizations may opt to hire people with marked coproduction beliefs and invest in developing it on current employees. The latter is of special relevance as it is critical to teach what “to follow” really means, as followership is not common sense. Resembling Challef's (2015) intelligent disobedience idea, organizations can develop based on a representation of the ideal follower as someone that combines a set of characteristics such as courage to challenge (Chaleff, 1995), critical thinking (Kelley, 1988) and coproduction (Carsten and Uhl-Bien, 2012). To foster this followership culture, organizations can, for example, provide psychologically safe climates (e.g. Peng et al., 2014).

All in all, the present study highlights the importance of coproduction beliefs, which decision-makers may enhance to establish governance mechanisms, thus avoiding power centralization (e.g. Padilla et al., 2007) and conversely, building on this role construal. This may ultimately foster built-in organizational immunity that prevents and breaks any destructive leadership cycle.

Figures

Figure 1

Research model

Table 1

Reliability, convergent validity and items' loadings

Construct	Items	Total (N = 359)			Abusive (N = 89)			Exploitative (N = 93)			Org-directed (N = 93)			Laissez-faire (N = 84)
		Load	CR	AVE	Load	CR	AVE	Load	CR	AVE	Load	CR	AVE	Load	CR	AVE
Coproduction	Followers should be on the lookout for suggestions they can offer to superiors	0.72	0.88	0.64	0.78	0.86	0.60	0.75	0.89	0.68	0.70	0.88	0.65	0.65	0.86	0.61
	Followers should proactively identify problems that could affect the organization	0.88			0.84			0.91			0.92			0.82
	Followers should be proactive in thinking about things that could go wrong	0.86			0.76			0.89			0.88			0.88
	Followers should communicate their opinions even when they know leaders may disagree	0.73			0.73			0.74			0.69			0.77
Constructive Resistance	I would explain that it should be done in a different way	0.89	0.94	0.79	0.89	0.94	0.79	0.91	0.94	0.68	0.88	0.94	0.78	0.90	0.95	0.81
	I would convince my supervisor to reassess whether or not the task is worthwhile	0.90			0.90			0.91			0.89			0.91
	I would explain that the task would not yield the expected benefits	0.89			0.92			0.84			0.92			0.89
	I would present logical reasons for doing the task differently or at a different time	0.88			0.84			0.90			0.86			0.90
Dysfunctional Resistance	I would ignore my supervisor	0.74	0.89	0.66	0.71	0.86	0.60	0.78	0.91	0.71	0.74	0.85	0.60	0.71	0.90	0.70
	I would pay no attention to my supervisor	0.82			0.77			0.82			0.86			0.85
	I would act as if I had never asked to do so	0.80			0.83			0.88			0.62			0.85
	I would disregard what this supervisor says	0.88			0.80			0.90			0.84			0.93
Sense of power	My wishes would not carry much weight	0.71	0.87	0.63	0.58	0.85	0.58	0.91	0.90	0.69	0.66	0.86	0.62	0.84	0.77	0.47
	Even if I voice them, my views would have little sway	0.80			0.73			0.87			0.86			0.32
	My ideas and opinions would be often ignored	0.74			0.78			0.67			0.71			0.84
	Even when I tried, I would not be able to get my way	0.91			0.93			0.84			0.89			0.63
Negative affect	Frightened	0.94	0.88	0.79	0.93	0.90	0.81	0.95	0.87	0.77	0.92	0.90	0.82	0.93	0.84	0.72
	Tormented	0.83			0.87			0.80			0.89			0.77
Positive affect	Enthusiastic	0.90	0.93	0.87	0.96	0.93	0.87	0.96	0.97	0.93	0.86	0.91	0.83	0.96	0.88	0.78
	Inspired	0.96			0.91			0.97			0.96			0.81
Destructive leader	Breaks promises he/she makes/Sees employees as a means to reach his or her personal goals./Falsifies documents/Reacts to failure	–	–	–	0.69	0.91	0.78	0.96	0.96	0.90	0.84	0.85	0.66	0.84	0.90	0.75
	Puts me down in front of others/Uses my work to get himself or herself noticed./Steals company property and resources/Avoids deciding	–			0.96			0.96			0.92			0.88
	Ridicules me/Puts me under pressure to reach his or her goals/Violates company policy/rules/Delays responding	–			0.97			0.92			0.66			0.87

Table 2

Correlations per condition

	Abusive					Exploitative					Organization					Laissez-faire
	CP	CR	DR	NA	PA	CP	CR	DR	NA	PA	CP	CR	DR	NA	PA	CP	CR	DR	NA	PA
CR	0.23*					0.26*					0.29**					0.33**
DR	−0.26*	−0.21*				−0.15	−0.08				−0.05	−0.06				−0.34**	−0.06
NA	−0.01	−0.26*	−0.10			−0.11	−0.27*	0.44***			0.02	−0.20	0.32**			−0.07	−0.21	−0.01
PA	−0.13	0.05	0.17	−0.21*		−0.02	0.14	−0.20	−0.21*		0.00	0.18	−0.03	−0.23*		0.17	−0.12	0.12	−0.26*
SP	−0.05	0.33**	−0.14	−0.42***	0.28**	−0.12	0.08	−0.09	−0.28**	0.28**	0.02	0.08	−0.45***	−0.28**	0.25*	−0.04	−0.23*	−0.08	−0.20	0.29**

Note(s): CP = coproduction; NA = negative affect; PA = positive affect; SP = sense of power; CR = constructive resistance; DR = dysfunctional resistance

^*p < 0.05 ^**p < 0.01 ^***p < 0.001

Table 3

Structural model results

	Total		Abusive		Exploitative		Organization		Laissez-faire
	β	f2	β	f2	β	f2	β	f2	β	f2
Gender→CR	−0.07	0.01	−0.16	0.03	−0.13	0.02	−0.04	0.02	0.00	0.00
Age→CR	0.03	0.00	0.03	0.00	0.00	0.00	0.02	0.00	0.00	0.00
DL→CR	0.13*	0.02	−0.10	0.01	−0.01	0.00	0.28*	0.07	0.22*	0.05
CP→CR	0.23***	0.05	0.25*	0.07	0.25†	0.06	0.23*	0.06	0.24*	0.06
SP→CR	0.08	0.00	0.32*	0.10	0.04	0.00	0.07	0.01	−0.17	0.03
NA→CR	−0.20**	0.03	−0.06	0.00	−0.18	0.03	−0.19	0.04	−0.24*	0.06
PA→CR	0.01	0.00	−0.05	0.00	0.11	0.01	0.09	0.01	−0.13	0.02
R2	0.15		0.22		0.15		0.21		0.28
Q2	0.11		0.14		0.08		0.12		0.08
Gender→DR	−0.10†	0.01	−0.07	0.01	−0.19†	0.05	−0.11	0.01	−0.05	0.00
Age→DR	0.09	0.01	0.14	0.02	0.14	0.02	−0.08	0.01	0.11	0.01
DL→DR	0.09	0.01	0.10	0.01	0.10	0.01	−0.01	0.00	0.07	0.01
CP→DR	−0.20***	0.04	−0.29*	0.09	−0.10	0.01	−0.03	0.00	−0.42***	0.16
SP→DR	−0.12†	0.00	−0.25	0.06	0.08	0.01	−0.42***	0.19	−0.15	0.02
NA→DR	0.15*	0.02	−0.16	0.02	0.48***	0.24	0.21†	0.05	0.04	0.00
PA→DR	0.03	0.00	0.17	0.03	−0.10	0.01	0.11	0.02	0.25	0.06
R2	0.09		0.18		0.28		0.28		0.19
Q2	0.05		0.06		0.17		0.11		0.01

Note(s): ^†<0.10;^*p < 0.05;^**p < 0.01;^***p < 0.001