Followers' unclear demands during the COVID-19 pandemic can undermine leaders' well-being: A moderated mediation model from an entrapment perspective.

Although effective leaders are important for reducing employee stress during the COVID-19, limited studies have examined how follower behaviors can influence leader stress and well-being during the COVID-19. This study draws on defeat-entrapment theory to examine how followers' unclear demands during the COVID-19 consequently impact leaders' psychological states and well-being. We conducted a three-wave time-lagged investigation with a sample of 281 leaders in the United Kingdom and found that followers' unclear demands could generate feelings of entrapment in leaders, leading to decreased levels of well-being outcomes in leaders. Importantly, we found that leaders who have higher levels of leadership responsibility during the COVID-19 are likely to feel trapped by followers' unclear demands. They are also likely to face higher levels of feelings of entrapment and impaired well-being compared with leaders who have lower levels of leadership responsibility. We discuss the implications for theories and practices, as well as directions for future research.

INTRODUCTION

Everyone wants them (leaders) to demonstrate empathy … Yet, for all their expertise, they are grappling with many new questions for which they do not have answers.

Since the COVID‐19 quickly began spreading across the world, employees worldwide have faced increased uncertainty associated with their jobs and lives. To cope with this uncertainty, employees are likely to become more dependent on leaders than usual (Mayseless & Popper, 2007; M. D. Mumford, Friedrich, et al., 2007; Yukl, 2002) as they seek assurance and protection. However, what followers demand from their leaders during this period can be ambiguous, inconsistent, or even volatile and infeasible due to the evolving situation as well as the level of uncertainty, threats, and disruption that COVID‐19 has caused. As Crayne and Medeiros (2020, p.12) indicated, “crises such as COVID‐19 are not static and the needs of the situation, and the people affected by the situation, may change over time ….” For example, followers may ask for different shifts on different days due to home care duties, request equipment for working at home that is not feasible to deliver, or seek to change tasks or job roles when the business is still uncertain.

Leaders, given that they play a key role in organising resources and shaping their followers' work conditions, are expected to take care of and respond to followers' needs, especially during times of crisis, as effective leadership is inseparable from meeting followers' demands and expectations (Burns, 1978). Specifically, leaders hold a structural position of power that enable them to control over valuable resources (French et al., 1959) in order to effectively influences others to act toward the achievement of group goals (Yukl, 2002). Due to their positions, during the crisis, leaders are expected to help followers by meeting their demands and easing their stress in order to keep everyone on track to achieve group goals. However, helping followers can bring negative implications on leaders' well‐being as interpersonal helping could generate costs for helpers (Bolino & Grant, 2016). For example, Lanaj and Jennings (2020) showed that leaders responding to followers' personal requests can increase leaders' negative affect. In light of the above, we seek to understand how unclear demands from followers during the COVID‐19 can affect leaders' well‐being and use the defeat‐entrapment theory (Gilbert, 2006; Taylor et al., 2011) to guide our theorising and empirical examination. In this paper, we focus on the leadership role of frontline leaders who have supervisory responsibilities, such as those working alongside staff and the day‐to‐day administration.

Defeat‐entrapment theory posits that defeat due to the failure to achieve important goals can give rise to feelings of entrapment. If all possible escape routes from a defeated situation are blocked, feelings of entrapment will consequently lead to increased stress and lower well‐being (Griffiths et al., 2018; Siddaway et al., 2015). Following the theory (Gilbert, 2006; Taylor et al., 2011), we propose that unclear demands during the COVID‐19 from followers can prevent leaders from effectively interacting with and influencing their followers to achieve collective goals. This defeats leaders' mission of leading and creates a sense of entrapment during the COVID‐19 for leaders that will impair their well‐being. Specifically, when leaders find themselves unable to either understand what their followers need and/or determine how to effectively address these needs to move things forward during the COVID‐19, leaders are likely to feel entrapped in their leader roles and to consequently experience higher levels of stress and lower well‐being. To better capture leader well‐being in the work context, we have included a broad range of positive and negative well‐being indicators, including COVID‐19‐relevant work valence, burnout, frustration at work, and job satisfaction.

Defeat‐entrapment theory also contends that one's judgment of escapablity determines one's levels of being defeated and entrapped by stressful events (Galhardo et al., 2016). Individuals face increased risk of entrapment when they attach a stronger sense of value or commitment to solving stressful events, such as COVID‐19 (Piccinelli & Wilkinson, 2000). Following this, we propose that leaders who have a stronger sense of leadership responsibility during the COVID‐19 (i.e. the extent to which leaders have a sense of duty of leading others: Chan & Drasgow, 2001) are more vulnerable to feeling defeated and entrapped by followers' unclear and difficult demands. Although a sense of duty inspires leaders to take on leadership responsibilities during a crisis, it might make leaders feel more stressed when facing difficulties with responding to and leading followers. Research has shown that responsible individuals are more likely to experience induced tension and stress after experiencing failures (Boyce et al., 2010; Cianci et al., 2010). Because responsible leaders view taking care of followers and resolving their demands as their own duty, they tend not to ignore followers, even though their needs are difficult to manage during the COVID‐19, which makes them more likely to feel entrapped by followers' demands. Therefore, leadership responsibility strengthens the deleterious impact of followers' demands on leader entrapment, which undermines leaders' well‐being (see Figure 1 for the theoretical model). In sum, the aims of this study are twofold: first, to understand why followers' unclear demands during the COVID‐19 influence leader well‐being and second, to show which types of leaders are more likely to be influenced by this situation.

FIGURE 1

Hypothesised model. t1 = measured at Time 1; t2 = measured at Time 2; t3 = measured at Time 3

Our study contributes to the literature in three particular respects. First, previous leadership studies mainly considered a top‐down process regarding how leaders influence employees' behavior and well‐being (see Hu et al., 2020 in the COVID‐19 time specifically; Inceoglu et al., 2018, for a review). Effective leaders have been identified as being a key factor in responding to COVID‐19 disruptions (Sergent & Stajkovic, 2020; Zhao et al., 2020). Although some recent leadership studies began to apply a bottom‐up approach which “reverses the lens” of leadership (Uhl‐Bien et al., 2014, p. 97) and examine how follower behaviors—for example, their hostile behaviors (Camps et al., 2020) or proactive behaviors (Xu et al., 2019)—can influence leader behaviors and judgment, such a bottom‐up approach to leadership is still limited (Ahmad et al., 2020). Furthermore, leader well‐being has been investigated in only a few studies focusing on the role of leaders' own behavior in shaping their well‐being (Weiss et al., 2018; Zwingmann et al., 2016). Our research thus expands the scope of follower‐related factors to investigate that meeting followers' demands and needs can cost leaders their well‐being.

Second, by examining entrapment as a mediating mechanism, our study offers a new theoretical lens of defeat‐entrapment theory (Gilbert & Allan, 1998; Taylor et al., 2011) to show how followers' unclear demands influence leader well‐being during the COVID‐19. Past research on leader well‐being has mainly adopted a resources perspective to explain the process by which stressful factors (e.g. exhibition of positive leadership, answering followers' personal requests) consume leaders' personal sources, thus resulting in lower levels of well‐being (e.g. Lanaj & Jennings, 2020; Lin et al., 2019). We go beyond a resource perspective and focus on leaders' perception by showing that followers' unclear demands can cause leaders to cognitively view the situation as aversive and feel trapped in their responsibilities.

Third, our study uncovers the potentially negative implications of leadership responsibility. Drawing upon defeat‐entrapment theory, we challenge the general assumption that responsible leaders are always associated with positive outcomes, such as generating high levels of psychological safety for employees (Doh & Quigley, 2014). This is called for in COVID‐19 research (Tsui, 2020). Our study suggests that followers' unclear demands are more likely to entrap those who are higher in leadership responsibility during the COVID‐19. As such, we need to consider the potentially negative effects of increased leader responsibility; such responsibility may redouble leaders' stress because they want to take on their responsibilities but do not see a way forward.

THEORETICAL BACKGROUND AND HYPOTHESES: FOLLOWERS' UNCLEAR DEMANDS AND LEADERS' WELL‐BEING—AN ENTRAPMENT PERSPECTIVE

We use insights from defeat‐entrapment theory to explain how unclear follower demands lead to leader entrapment. This theory posits that entrapment arises when important goals are not met and people cannot disengage themselves from the associated failure (Wrosch et al., 2003). Burns (1978) argued that an essential goal is to “satisfy—or appear to satisfy—specific needs of the followers” (p. 294). Similarly, Bass (1985) stated that fulfilling employees' need is a central aspect of being an effective leader. When employees' needs are fulfilled, they are likely to develop high levels of commitment and to make an effort to fulfill collective interests. By contrast, unsatisfied employees can develop resentment and have an impaired ability to work for the collective (Lian et al., 2012). In our case, we argue that unclear demands from followers can trap leaders in feelings of defeat due to their failure to meet essential leader goals (i.e. leading by addressing followers' demands and engaging in effective social exchanges with followers).

The concept of unclear demands and their detrimental effects on individuals' well‐being has been examined in the service context. Dormann and Zapf (2004) suggested that unclear demands from customers bring ambiguity in customer interaction, which can jeopardise employee well‐being. For leaders, followers can be regarded as the internal customers (Wieseke et al., 2009) whom leaders serve to earn their contributions in return. During the COVID‐19, follower demands vary from one person to another (e.g. making an inquiry for information, asking for a different time shift, and requesting specific equipment to work from home). Even for the same person, needs can fluctuate in accordance with a quickly developing situation. However, when leaders are not able to address followers' needs, they cannot make followers reciprocate with input and performance (Gerstner & Day, 1997), especially when the followers' input is essential for an effective and flexible response to the changing COVID‐19 situation. Research has found that when leaders fail to meet the needs of their team members, followers tend to feel unaccepted by their leaders. This prevents the leader from achieving effective leadership, such as building group cohesiveness and promoting group performance (Boies & Howell, 2006; Cogliser & Schriesheim, 2000). Additionally, unlike during organisational crises, when leaders can still access information (James & Wooten, 2005; T. V. Mumford, Campion, et al., 2007), in a global pandemic such as COVID‐19, leaders, like others, experience great uncertainty and a lack of information and resources themselves. This makes it even more difficult for them to respond to followers' needs and requests, as well as to fulfill the role of leading people. In addition, their formal positions as leaders in organisations do not allow them to completely withdraw from their leadership roles. In such situations, where leaders have difficulty meeting followers' demands but cannot disengage from their leadership roles, leaders are likely to feel entrapped in their positions. The difficulty addressing unclear demands from followers can make leaders feel useless and powerless—typical feelings when individuals are in states of entrapment (Taylor et al., 2009).

Furthermore, we expect entrapment to impact leader well‐being. Perceptions of being trapped in stressful situations where no hope of changing the situations exists can lead to lower levels of well‐being (Ehlers & Clark, 2000; Taylor et al., 2011). Feelings of entrapment can cause individuals to feel incapable of defending themselves or escaping from their situations. Empirical evidence has shown that for those who are in caged conditions, where escaping or defending is impossible, mental health and well‐being are at risk (see a meta‐analysis by Griffiths et al., 2014, 2018; Siddaway et al., 2015). Prior well‐being scholars (e.g. Diener, 2006) suggested that well‐being is “not only the absence of mental disorder but also the presence of positive psychological resources” (p. 468) and recommended including both positive and negative indicators to provide a comprehensive account of well‐being. In our study, we focus on frustration at work and burnout to capture the negative side of the loss of interest and energy in work activities. We also focus on work valence and job satisfaction to represent leaders' impaired positive psychological states due to entrapment.

As for the negative side, frustration is defined as the interference with both goal attainment and goal‐oriented activity and with goal maintenance at work (Spector, 1978). It occurs when the realisation or maintenance of a goal is inhibited (Lazar et al., 2006). We expect that because unclear followers' demands inhibit leaders to achieve or maintain their leadership goals and therefore entrap them in their role responsibilities. Such feelings of interference with goal achievement would increase leaders' feelings of frustration at work. Burnout refers to “a syndrome of emotional exhaustion, depersonalisation, and reduced personal accomplishment that can occur among individuals who do ‘people work’ of some kind.” (Maslach & Jackson, 1986, p. 1). Prior empirical works showed that defeat and entrapment that generate feelings of powerlessness and loss of control lie at the core of burnout at work (e.g. Buunk et al., 2007). Drawing upon this research, we expect that leaders' entrapment generated by unclear followers' demands lead to increased burnout.

In terms of the impaired positive side, work valence is defined as individuals' beliefs regarding the degree of importance that work plays in their lives (Paullay et al., 1994). Individuals may choose to passively cope with defeat and entrapment by acting submissively and attaching less importance to their goals (Allan & Gilbert, 1997; Sloman et al., 2003). Accordingly, we expect that entrapment would make leaders see their work as less important as a form of harm avoidance. Finally, job satisfaction refers to a pleasurable emotional state that results from one's positive appraisal of one's experiences at work (Locke, 1969). We expect that when leaders feel difficult to fulfill their responsibilities and entrapped in their roles, they tend to see their job as stressful and negatively appraise their working experiences, leading to decreased levels of job satisfaction.

Unclear demands from followers are positively related to leader entrapment during the COVID‐19, which, in turn, is negatively related to leaders' work valence (Hypothesis 1a) and job satisfaction (Hypothesis 1b) and positively related to frustration at work (Hypothesis 1c) and job burnout during the COVID‐19 (Hypothesis 1d).

The moderating role of leadership responsibility

Defeat‐entrapment theory posits that perceived lower escapablity can strengthen the relationship between stressors and entrapment, which has been empirically supported (Galhardo et al., 2016; Minkler et al., 1997; Ng et al., 2016). In our research context, we suggest that leaders with more leadership responsibility are more likely to perceive lower escapablity relative to their leadership duties; thus, they are more likely to feel trapped by unclear follower demands.

Specifically, leaders who have higher levels of leadership responsibility tend to view providing care, guidance, and security for followers as their own duty (Chan & Drasgow, 2001). They also tend to engage in other‐centered actions, subordinate their priorities to the majority and not tolerate self‐indulgency or withdrawing from failure (Boyce et al., 2010; Jeong et al., 2020). One relevant study by Lanaj et al. (2021) found that self‐sacrifice behaviors demonstrated by leaders who strongly identify with their leader role can generate both positive (i.e. increased task performance and perceived prosocial impact) and negative consequences (i.e. increased depletion and conflict at home) for leaders. We argue that during the COVID‐19, as unclear demands grow, leaders with a stronger sense of responsibility are more likely to sacrifice themselves and keep directing their efforts toward supporting followers, even though followers' demands are unclear and difficult to address. This tends to prevent leaders from fulfilling their duty because they want to be reliable to their followers, but they face severe difficulties in actually addressing their demands.

Our reasoning is consistent with previous findings, though in different contexts. For example, Galhardo et al. (2016) found that people who perceived parenthood as a particularly important life goal had more difficulty escaping the negative effects of infertility and tended to report stronger feelings of entrapment due to infertility. Minkler et al. (1997) found that grandparents who perceived themselves as the primary caregivers for their grandchildren were more likely to experience entrapment in responding to the difficulty of raising grandchildren. Thus, we propose the following:

Leadership responsibility moderates the relationship between unclear demands from followers and leader entrapment during the COVID‐19, such that this relationship is more positive when leadership responsibility is high rather than low.

Taken together, we propose that leader entrapment mediates the relationship between unclear demands from followers and leader work and well‐being outcomes. In addition, this mediating effect is conditional on the moderator of leadership responsibility for the path from unclear demands to leader entrapment. As such, we propose a first‐stage moderated mediation model:

The indirect relationship between unclear demands from followers and leaders' work valence (Hypothesis 3a), job satisfaction (Hypothesis 3b), frustration at work (Hypothesis 3c), and burnout (Hypothesis 3d) via leader entrapment during the COVID‐19 is moderated by leadership responsibility, such that the indirect effect is stronger when leadership responsibility is high rather than low.

METHOD

We recruited 400 working professionals in the United Kingdom through Prolific Academic. We used screening criteria of having full‐time jobs during the COVID‐19 and having leadership responsibilities such as giving instructions to subordinates. Participants completed three surveys given at 1‐month intervals, and we used the participant IDs generated by the platform to match data over time. We sent out the first survey on May 18, 2020, approximately 2 months after the government had first imposed strict lockdown measures on March 26 and people had started to work from home. At Time 1, we asked participants to report their experiences with unclear demands from followers and follower proficiency during this period and to rate their levels of leadership responsibility and leader self‐efficacy in general. We sent out the second survey on June 8 (3 weeks after Time 1). At Time 2, we asked participants about their perceived levels of entrapment and negative affect. We sent out the third survey on June 29 (3 weeks after Time 2) when some lockdown measures had been eased and people were allowed to leave the house for any reason but were still asked to work from home if possible. At Time 3, we measured participants' levels of job satisfaction, work valence, burnout, and frustration at work. We received 358 responses at Time 1, and 328 of these respondents completed the survey again at Time 2 (response rate of 80.4%). Finally, we had a matched sample of 281 responses at Time 3 (response rate of 88.3%).

Of these respondents, 45.6 per cent were male, and the average job tenure was 10.34 years. Two per cent of them had received less than a high school diploma, 5.3 per cent had completed doctoral degrees, and the highest proportion had received bachelor's degrees (43.4% of the sample).

Measures

All variables were measured using 5‐point Likert scales (1 = strongly disagree to 5 = strongly agree), unless otherwise stated.

Unclear demands from followers during the COVID‐19

We adapted a four‐item scale of ambiguous customer expectations developed by Dormann and Zapf (2004), which was originally designed to measure employees' perceptions of unclear expectations and difficult requests from customers. This adaptation recognises followers as internal customers of the leaders, who need to satisfy followers' needs so the followers can produce improved outcomes (Huang & Rundle‐Thiele, 2014). This is also aligned with our theorisation of unclear followers' demands as a critical stressor for organisational leaders during the COVID‐19, where followers need to be supported in order to effectively collaborate with the leader to solve rapidly changing problems. We referent shifted the word “customers” to “followers.” For this scale, participants were instructed, “Think about your interactions with your followers during the COVID‐19. To what extent do you agree with the following statements about your followers?” Items included, “My followers' needs are often contradictory,” “It is not clear what my followers want from me,” “It is difficult to make arrangements with my followers,” and “My followers' requirements can complicate my work” (α = .82).

Leader entrapment during the COVID‐19

Entrapment was measured using a nine‐item scale developed by Allan and Gilbert (1997). They developed measures for internal (escape motivation triggered by internal feelings and thought) and external entrapment (escape motivation triggered by the perception of things in the outside world). We used external entrapment in this study because we aimed to study how the external situation, that of unclear expectations from followers during the COVID‐19 situation, induces escape motivation in leaders. A sample item was “I feel trapped by my obligations” (1 = does not describe my feelings to 5 = clearly describes my feeling), with the instructions, “To what extent do the following statements describe your feelings about performing leadership duties at work during the COVID‐19?” (α = .93).

Leadership responsibility

Leadership responsibility was measured using five items from the social normative motivation to lead scale developed by Chan and Drasgow (2001). We chose this scale to measure leadership responsibility because it implies leading out of a sense of duty or responsibility (Chan & Drasgow, 2001). A sample item was “It is my responsibility to lead others” (α = .83).

Leaders' work valence during the COVID‐19

Work valence was measured using a three‐item scale from Hirschfeld and Feild (2000), which they shortened from the scale originally developed by Lodahl and Kejnar (1965). Instructions for this scale read, “Over the past few weeks of the COVID‐19, to what extent do you view work as the following?” A sample item was “An important part of my daily life” (1 = not at all to 5 = a great deal; α = .86).

Leaders' frustration at work during the COVID‐19

We measured frustration at work with a three‐item scale developed by Peters et al. (1980). Instructions for this scale read, “Over the past few weeks of the COVID‐19, as a leader of a team, to what extent do you feel the following?” A sample item was “Being frustrated comes with the job,” and “Overall, I experience frustration on my job” (1 = not at all to 5 = a great deal; α = .92).

Leaders' burnout during the COVID‐19

We measured burnout based on three items from Hollet‐Haudebert et al. (2011), which were modified from the Maslach burnout inventory survey and have been validated in sales workers (Schaufeli et al., 1993). We used the three highest loading items from three dimensions, which were “Over the past few weeks, I feel burned out from my work” (emotional exhaustion), “I have become less interested in my work since I started this job” (depersonalisation), and “I have accomplished many worthwhile things in this job” (reverse coded; personal nonaccomplishment; α = .64).

Leaders' job satisfaction during the COVID‐19

We measured job satisfaction based on a three‐item scale developed by Cammann et al. (1983). A sample item was “All in all, I am satisfied with my job” (α = .92).

Control variables

As female and male leaders have reported different levels of pressure associated with leadership roles according to the gender ratios of the organisations or the industries (Gardiner & Tiggemann, 1999), we controlled for potential gender effect (0 = male, 1 = female) on leader well‐being. Furthermore, given that prior knowledge and personal experiences shape leaders' reactions toward followers' needs (Sieweke & Zhao, 2015), we controlled for education (0 = less than high school to 7 = professional degree) and job tenure (in years). We also controlled for participants' managerial self‐efficacy, as leaders with high managerial self‐efficacy are less likely to feel threatened or defensive as a result of follower‐initiated stressors (Fast et al., 2014). Self‐efficacy was measured using a four‐item scale developed by Fast et al. (2014), with a sample item being, “I will be able to successfully overcome many challenges” (α = .92). Next, we included a measure of follower task proficiency to ensure that it was followers' demands and not their task performance that affect leaders' well‐being. Follower task proficiency was measured on a three‐item scale developed by Griffin et al. (2007). A sample item was “During the COVID‐19, have your followers completed their core tasks well” (1 = very little to 5 = a great deal; α = .92).

Finally, as unclear demands from followers can evoke leaders' negative feelings, which can also affect leaders' well‐being, negative affect has been established as a significant mediator linking work stressors and well‐being outcomes (Meier & Semmer, 2013; Michel et al., 2016). To demonstrate the unique effect of feelings of entrapment on the relationship between unclear demands and leaders' outcomes, we included negative affect as an alternative mediator. Negative affect was measured using a 10‐item scale shortened and validated by Thompson (2007) from PANAS (Watson et al., 1988). Sample items included “(to what extent do you feel the following over the past few weeks) Upset,” “Hostile,” and “Ashamed” (α = .87).

RESULTS

Preliminary statistics

We first conducted a series of confirmatory factor analyses (CFAs) to examine the validity of the measures used in our model. In order to maintain reasonable degrees of freedom, we followed Kline (2015) and used item parceling for entrapment, which has nine items, and negative affect, which has 10 items. We formed three parcels for entrapment and four parcels for negative affect. Each parcel comprised two or three randomly assigned items. As shown in Table 1, the hypothesised 11‐factor model provided a good fit to the data (χ 2(610) = 1176.08, root mean square of approximation [RMSEA] = .06, comparative fit index [CFI] = .92, Tucker–Lewis index [TLI] = .91, standardised root mean square residual [SRMR] = .06). This result supports the distinctiveness of the variables used in this study. The means, standard deviations, and correlations among variables are shown in Table 2.

TABLE 1

Fit comparisons of alternative factor models

	χ 2	df	Δχ 2 (df)	RMSEA	CFI	TLI	SRMR
Hypothesised model	1176.08	610	‐	.06	.92	.91	.06
Model A	1437.86	620	261.78 ** (10)	.07	.89	.87	.07
Model B	1405.10	620	229.02 ** (10)	.07	.89	.87	.09
Model C	2040.02	620	863.94 ** (10)	.09	.80	.77	.13
Model D	1396.77	620	220.69 ** (10)	.07	.89	.88	.06
Model E	1809.40	629	633.32 ** (19)	.08	.83	.81	.08
Model F	2268.66	637	1092.58 ** (27)	.10	.77	.75	.09
Model G	3063.45	655	1887.37 ** (45)	.11	.66	.64	.11
Model H	4130.59	662	2954.51 ** (52)	.14	.51	.48	.14
Model I	5137.11	665	3961.03 ** (55)	.16	.39	.33	.13

Note: Model A: 10‐factor model combining entrapment and negative affect as one factor; Model B: 10‐factor model combining frustration at work and burnout as one factor; Model C: 10‐factor model combining entrapment and perceived impact of COVID‐19 as one factor; Model D: 10‐factor model combining work valence and job satisfaction as one factor; Model E: nine‐factor model combining entrapment, negative affect, and unclear demands as one factor; Model F: nine‐factor model combining entrapment, negative affect, unclear demands, and leader managerial self‐efficacy as one factor; Model G: five‐factor model combining entrapment, negative affect, unclear demands, and leader managerial self‐efficacy as one factor and work valence, job satisfaction, burnout, and frustration as another factor; Model H: entrapment, negative affect, unclear demands, and leader managerial self‐efficacy as one factor; follower proficiency, leadership responsibility, and perceived impact of COVID‐19 as one factor; and work valence, job satisfaction, burnout, and frustration as the last factor; and Model I: one‐factor model combining all variables.

Abbreviations: CFI, comparative fit index; RMSEA, root mean square of approximation; SRMR, standardised root mean square residual; TLI, Tucker–Lewis index.

p < .01.

TABLE 2

Variable, means, standard deviations, and correlations

Variables	Means	SD	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Gender_t1	0.54	0.50
2. Job tenure_t1	10.34	8.51	.04
3. Education_t1	4.68	1.53	.01	.09
4. Follower proficiency_t1	4.37	0.60	.02	.10	.05
5. Managerial self‐efficacy_t1	3.83	0.65	−.05	.04	−.01	.29 **
6. Perceived impact of COVID‐19_t2	3.73	0.90	.18 **	.04	.02	−.07	−.05
7. Unclear follower demands_t1	2.69	0.90	.11	.00	.06	−.26 **	−.20 **	.15 *
8. Leadership responsibility_t1	3.80	0.73	.01	.01	−.03	.10	.32 **	.09	−.03
9. Entrapment_t2	2.10	0.95	.08	−.05	−.05	−.23 **	−.37 **	.25 **	.23 **	−.05
10. Negative affect_t2	2.30	0.62	.09	−.15 *	−.02	−.26 **	−.46 **	.17 **	.25 **	−.23 **	.64 **
11. Job satisfaction_t3	3.85	1.07	.09	−.02	−.01	.20 **	.35 **	.01	−.16 **	.17 **	−.48 **	−.48 **
12. Work valence_t3	3.07	1.03	.13 *	.08	.02	.21 **	.32 **	.09	−.08	.21 **	−.36 **	−.43 **	.66 **
13. Frustration at work_t3	2.84	0.98	.06	−.02	.07	−.17 **	−.27 **	.22 **	.28 **	.03	.48 **	.42 **	−.39 **	−.25 **
14. Burnout_t3	2.65	0.94	.02	−.12 *	.02	−.23 **	−.38 **	.04	.18 **	−.17 **	.51 **	.62 **	−.63 **	−.60 **	.47 **

Note: N = 281. t1 = Time 1; t2 = Time 2; t3 = Time 3.

p < .05.

p < .01.

Hypotheses testing

We used path analysis in Mplus 8 (Muthén & Muthén, 2012–2020) and a maximum likelihood estimation to test our hypotheses. The demographics, follower proficiency, and leader managerial self‐efficacy were regressed on the mediators and outcomes in all analyses. To test Hypotheses 1 and 2, we specified the indirect effects of entrapment and negative affect linking unclear demands with outcomes. We allowed the disturbances of two mediators to covary because they were collected at the same time. First, we examined a full mediation model with no direct effects from unclear demands to outcomes specified. This model provides a good fit to the data (χ 2 = 8.24, df = 4, RMSEA = .06, CFI = 1.00, TLI = .93, SRMR = .01). We then tested a partial mediation model that included direct effects. This model was fully saturated with zero degrees of freedom and showed that unclear demands were not significantly related to job satisfaction (B = −.01, p = .82), work valence (B = .04, p = .51), or burnout (B = .00, p = .94) but were positively related to frustration at work (B = .16, p < .01). We concluded that entrapment fully mediated the relationship between unclear demands and job satisfaction, work valence, and burnout whereas it partially mediated the relationship between unclear demands and frustration. This model with one direct effect from unclear demands provides a superior model fit (χ 2 = .93, df = 3, RMSEA = .00, CFI = 1.00, TLI = 1.05, SRMR = .003). We hereafter report on the findings of this model (see Table 3).

TABLE 3

Results of a mediation path model (coefficients and standard errors)

	Entrapment	Negative affect	Job satisfaction	Work valence	Burnout	Frustration at work
Control variables
Gender_t1	.03 (.10)	.05 (.06)	.26 * (.11)	.30 ** (.11)	−.04 (.09)	−.03 (.10)
Job tenure_t1	.00 (.01)	−.01 (.00)	−.01 (.01)	.00 (.01)	.00 (.01)	.00 (.01)
Education_t1	−.03 (.03)	−.01 (.02)	−.02 (.03)	.00 (.03)	.03 (.03)	.05 (.03)
Follower proficiency	−.13 (.09)	−.09 (.06)	.08 (.09)	.11 (.09)	−.08 (.08)	−.01 (.09)
Managerial self‐efficacy_t1	−.46 *** (.08)	−.38 *** (.05)	.20 * (.09)	.20 * (.09)	−.12 (.08)	−.08 (.09)
Perceived impact of COVID‐19_t2	.22 *** (.05)	.09 * (.04)	.14 * (.06)	.18 ** (.06)	−.09 (.05)	.10 (.06)
Independent variable
Unclear demands_t1	.13 * (.06)	.09 * (.04)				.33 *** (.07)
Moderators
Leadership responsibility_t1
Two‐way interaction
Unclear demands × Leadership responsibility
Mediator
Entrapment_t2			−.34 *** (.07)	−.19 * (.07)	.21 ** (.06)	.33 *** (.07)
Negative affect_t2			−.46 *** (.12)	−.46 *** (.12)	.69 *** (.10)	.23 * (.11)
R 2	.22 ***	.28 ***	.34 ***	.27 ***	.43 ***	.30 ***

Note: N = 281. Unclear demands and leadership responsibility are grand‐mean centered. Unstandardised regression coefficients are shown. Bold numbers indicate significant results.

p < .05.

p < .01.

p < .001.

We found unclear demands were positively related to entrapment (B = .13, p < .05). Entrapment was positively related to burnout (B = .21, p < .01) and frustration at work (B = .33, p < .001) and was negatively related to job satisfaction (B = −.34, p < .001) and work valence (B = −.19, p < .001). A bootstrapping approach with 5000 resamplings showed a significantly indirect effect of entrapment linking unclear demands to work valence (effect size = −.02, confidence intervals [CIs] [−.07, −.001]), burnout (effect size = .03, [.002, .07]), frustration at work (effect size = .04, [.002, .10]), and job satisfaction (effect size = −.04, [−.10, −.002]). These results support Hypotheses 1a–1d.

We then included leadership responsibility as a moderator in the path model (see Table 4). The independent variable and the moderator were grand‐mean centered. This model provided a good fit to the data (χ 2 = 19.28, df = 13, RMSEA = .04, CFI = .99, TLI = .96, SRMR = .02). The interaction term of unclear demands and leadership responsibility was significantly related to leader entrapment (B = .14, p < .01). The interacting patterns are plotted in Figure 2. Unclear demands exerted a strong negative effect on entrapment when leadership responsibility was high (1 SD above the mean; B = .14, p < .01) but did not have a significant effect when leadership responsibility was low (1 SD below the mean; B = .00, p = .99), supporting Hypothesis 2.

TABLE 4

Results of a moderated mediation path model (coefficients and standard errors)

	Entrapment	Negative affect	Job satisfaction	Work valence	Burnout	Frustration at work
Control variables
Gender_t1	.02 (.10)	.05 (.06)	.26 * (.11)	.30 ** (.11)	−.04 (.09)	−.03 (.10)
Job tenure_t1	.00 (.01)	−.01 (.00)	−.01 (.01)	.00 (.01)	.00 (.01)	.00 (.01)
Education_t1	−.03 (.03)	−.01 (.02)	−.02 (.03)	.00 (.03)	.03 (.03)	.05 (.03)
Follower proficiency	−.16 (.09)	−.09 (.06)	.08 (.09)	.11 (.09)	−.08 (.08)	−.01 (.09)
Managerial self‐efficacy_t1	−.50 *** (.08)	−.38 *** (.05)	.20 * (.09)	.20 * (.09)	−.12 (.08)	−.08 (.09)
Perceived impact of COVID‐19_t2	.22 *** (.06)	.09 * (.04)	.14 * (.06)	.18 ** (.06)	−.09 (.05)	.10 (.06)
Independent variable
Unclear demands_t1	.10 (.06)	.09 * (.04)				.33 *** (.07)
Moderators
Leadership responsibility_t1	.18 ** (.06)
Two‐way interaction
Unclear demands × Leadership responsibility	.14 * (.06)
Mediator
Entrapment_t2			−.34 *** (.07)	−.19 * (.07)	.21 ** (.06)	.33 *** (.07)
Negative affect_t2			−.46 *** (.12)	−.46 *** (.12)	.69 *** (.10)	.23 * (.11)
R 2	.24 ***	.28 ***	.34 ***	.27 ***	.43 ***	.30 ***

Note: N = 281. Unclear demands and leadership responsibility are grand‐mean centered. Unstandardised regression coefficients are shown. Bold numbers indicate significant results.

p < .05.

p < .01.

p < .001.

FIGURE 2

The relationship between unclear demands and entrapment under conditions of low and high leadership responsibility

Finally, the indirect effect of entrapment was significant and stronger when leadership responsibility was higher (work valence: effect size = −.04 [−.10, −.006]; burnout: effect size = .04, [.01, .10]; frustration at work: effect size = .07, [.02, .14]; and job satisfaction: effect size = −.07, [−.15, −.02]) than when it is lower (work valence: effect size = .00 [−.04, .04]; burnout: effect size = .00 [−.04, .04]; frustration at work: effect size = .00, [−.06, .07]; and job satisfaction: effect size = .00, [−.07, .06]), with significant differences in the magnitude of the effect sizes (work valence: difference = −.04 [−.12, −.001]; burnout: difference = .04 [.004, .11]; frustration at work: difference = .07, [.003, .16]; and job satisfaction: difference = −.07, [−.16, −.003]). Hypotheses 3a–3d were supported.

DISCUSSION

Altogether, the results of this study show that leaders who have high leadership responsibility tend to feel entrapped when facing unclear demands from followers, which in turn leads to decreased leader well‐being. Our study has important implications for the literature and practices.

Implications

Existing COVID‐19 research has produced strong calls for leaders to act as promotors who listen to followers' needs and support them (Sergent & Stajkovic, 2020; Van Bavel et al., 2020; Zhao et al., 2020). Nevertheless, the process through which leader well‐being is impacted is less understood and therefore should be a subject for future research (e.g. Inceoglu et al., 2020). Past research has suggested that helping behavior results in negative consequences, such as role overload and fatigue (Bolino et al., 2015). More recent leadership studies show that responding to followers' personal requests generates costs for leaders (Lanaj & Jennings, 2020). Our research adds to this line by investigating a phenomenon in which followers place excessive demands on leaders, impairing leaders' well‐being. Our results show that although leaders are expected to take good care of their followers, the costs of managing followers' needs in an unpredictable environment such as during the COVID‐19 are high and therefore should not be ignored. Moreover, as an addition to the existing literature that focuses on leaders' affective states when explaining how and when leader well‐being is affected by followers (Lanaj & Jennings, 2020; Lin et al., 2019), we provide a new theoretical perspective of the defeat‐entrapment model to explain this process. Taken together, our research adds more evidence to the line of research on leader well‐being by showing that in a crisis context, followers' excessive demands significantly harm leaders' well‐being.

Our research contributes more empirical evidence to the leadership literature on bottom‐up perspectives of leadership. Scholars have long called for studies to go beyond a traditional leader‐centric view and adopt a bottom‐up perspective to examine the role of followers (e.g. Uhl‐Bien et al., 2014). Past research on followers has mainly focused on examining how follower psychological states (e.g. Lord & Brown, 2001), congruence between leaders and followers (Giessner & van Knippenberg, 2008), or followers' relationships with leaders (Tsai et al., 2017) act (i.e. mediate or moderate) in the processes through which leaders generate outcomes. Recently, research has begun to more directly “reverse the lens” and test the impact of follower behaviors on leaders (e.g. Camps et al., 2020), for example, the effects of followers' hostile behaviors on increases in abusive supervision (Camps et al., 2020) and of followers' proactive behaviors on increases in leaders' perceived access to resources (Xu et al., 2019). Given that followers are key to the leadership process, more needs to be known about how follower needs and expectations impact leaders (Lord et al., 2020; Nguyen et al., 2018). Our study contributes to this literature by expanding the scope of follower‐related factors and explaining how addressing followers' unclear demands can harm leaders' perception of their well‐being.

Our research also contributes to the literature on leader well‐being by providing a new theoretical perspective—the defeat‐entrapment model. Prior research mainly uses the defeat‐entrapment model in a clinical psychology context to explain why individuals become depressive and experience lower levels of well‐being (Panagioti et al., 2015), whereas many studies in work psychology have used resource theories (e.g. conservation of resources theory) to explain how workplace stressors influence well‐being (Debus & Unger, 2017; Hobfoll, 2001). The lack of the application of the defeat‐entrapment model in the work psychology literature does not imply that people at work do not experience feelings of entrapment. Prior research indicated that employees could feel “defeated” and “trapped” in work situations or in their job roles (Fernet et al., 2013; Fisk & Neville, 2011), whereas limited frameworks are able to specifically capture and explain how feelings of entrapment influence well‐being in the workplace. We found that the mediating effect of leader entrapment linking unclear follower demands to leader outcomes remains significant when leader negative affect is controlled. This provides theoretical and evidence‐based support for this new perspective to help with understanding how and why leader well‐being is impacted in the workplace. Further, our model has the potential to be extended to a wider working population and offers new insights into how and why workplace stressors in general impact employee health and well‐being (Pindek et al., 2019). Our study on the connection between defeat‐entrapment and the moderating effect of responsibility is a meaningful combination for organisations to identify those who are more likely to feel entrapped and experience impaired well‐being during crises, thereby reducing the negative impact of stressors on those individuals.

Finally, our study highlights the potential negative implications of leadership responsibility. Although higher leadership responsibility has been found to benefit followers (Doh & Quigley, 2014; Tsui, 2020), our finding suggests that high levels of responsibility can make leaders to consistently devote great effort to addressing followers' demands, even when these demands are unclear and difficult to address, thereby exacerbating their perceptions of defeat and entrapment. Our finding is in line with the recent literature of self‐sacrifice leadership, which shows that when leaders care about being effective as leaders, they experience depletion of personal resources as they sacrifice their own interests in order to promote the welfare of followers (Hoogervorst et al., 2012; Lanaj et al., 2021). Our study joins this line of research showing that positive leadership might generate costs for leaders (Lanaj & Jennings, 2020; Liao et al., 2020; Lin et al., 2019). It also extends prior research about the dark side of dutifulness (Dahm et al., 2017) by highlighting the negative impact of being responsible under adverse situations on leader well‐being.

Our research has several practical implications for both employees and leaders. First, because we find that unclear followers' demands can negatively impact leaders' well‐being, we suggest implementing progressive steps to establish a mutual understanding between leaders and followers and reduce the detrimental effects of followers consistently making demands of leaders. Organisations could help leaders better receive the messages and concerns from their followers by improving leaders' active listening skills (Lloyd et al., 2015). To reduce leaders' stress, organisations could facilitate interpersonal connections and support among leaders, improving the flow of messages and the quality of solutions (Detert et al., 2013). Organisations could also encourage followers to clarify their requests and concerns in a communal way. Providing employee training in nondefensive communication tactics (Thacker & Wayne, 1995) can be a good method to reduce leaders' burdens.

Second, organisations should pay special attention to their leaders' well‐being during the COVID‐19 as leaders are taking on more responsibilities and are experiencing significant pressure from a variety of sources. Organisations should make leaders aware that responding to followers' unclear demands may cause them to feel entrapped in their leadership role. Psychological distress among leaders can be identified early and treated effectively, for example, by encouraging them to allot specific times to respond to followers' requests (Lanaj & Jennings, 2020) and offering training to develop leaders' skills in cleansing their minds (Zhang et al., 2019). Organisations could offer practices to help leaders psychologically detach from work when appropriate (e.g. taking part in mindfulness training: Song et al., 2018; doing physical exercises: Toker & Biron, 2012), which would help promote leaders' health and well‐being and buffer the negative consequences for leaders as the targets of unclear demands.

Finally, we suggest that leaders with high levels of leadership responsibility would benefit from acting to fulfill their responsibilities in more strategic ways (e.g. when the requests or suggestions are clearly stated and useful) rather than trying to address every demand followers raised. Organisations could offer additional training or develop mentoring programs, especially for those leaders with less experience who are more vulnerable to the negative effects of leadership responsibility, to help them identify different types of requests and develop their skills is responding more strategically to their followers.

Limitations and future directions

Despite the strengths (e.g. longitudinal design and controlling for prior established predictors and mediators), our research has some limitations. First, we used leaders' self‐ratings for all variables, which raises concerns regarding common method variance (CMV). However, prior studies have suggested that self‐ratings are appropriate for the assessment of leaders' own behaviors and well‐being (McClean et al., 2019). Moreover, interaction effects are more difficult to detect through statistical tests due to the interaction terms being deflated (Siemsen et al., 2010). Finally, our independent variable of unclear follower demands, which is the most problematic in terms of same‐source bias as it was reported by the leaders in our study, had low correlations with the other study variables. Therefore, we concluded that CMV was less likely to bias our conclusions. Nevertheless, the use of multisource data in future research would be beneficial.

Another issue is that we cannot draw firm conclusions for causality and possible reciprocal effects among variables. For example, it is possible that lower well‐being increases feelings of entrapment, increasing supervisors' tendencies to see followers' demands as unclear and irritating. Moreover, our 1‐month time lag may not have been sufficiently long to capture the full development process of the outcome variables, such as burnout and frustration. However, evidence suggests that burnout and frustration can develop within a short period of time subject to role changes and life events (Friedman, 2000; Kurth et al., 2011). Nevertheless, we strongly recommend that future studies could use a longitudinal design and longer time lags to extend our research.

Next, although we used defeat‐entrapment theory to explain that responsible leaders are likely to attribute failure in meeting followers' demands to their personal inadequacy, we did not directly assess whether these leaders actually feel so. Future research could directly examine the potential mechanisms, for example, felt vulnerability and inadequacy or decreased sense of self‐worth, that underlie the moderating effect of leadership responsibility. Also, although fulfilling followers' needs has been viewed as an essential goal for leaders (Burns, 1978), leaders could have other goals (e.g. being responsible toward society and the growth of team performance; Hofstede et al., 2002), and leaders are expected to adapt or prioritise different goals according to the specific situations they face (Seah et al., 2014). Disentangling the effects of different types of goals interplaying with situational factors may provide valuable insights into their relative importance for leaders' feelings of entrapment, which is another fruitful area for future research.

CONFLICT OF INTEREST

No conflicts of interests.

ETHICS STATEMENT

The ethical approval was granted by Durham University Business School in May 2020.