Understanding the initiative paradox: the interplay of leader neuroticism and follower traits in evaluating the desirability of follower proactivity

ABSTRACT

This study investigates the relationship between leaders’ neuroticism and their evaluation of the desirability of followers’ proactive behaviour. We argue that leaders high in neuroticism are likely to evaluate follower proactivity as less desirable and that this relationship is amplified when followers are low in conscientiousness and agreeableness. Based on trait activation theory, we further propose that worry and threat act as underlying mechanisms of the interaction between these traits. We hypothesize that leaders high in neuroticism feel more threatened by the proactive behaviours shown by followers’ low in conscientiousness and agreeableness and rate the proactive behaviours of these followers as less desirable. In a field study with 205 leader–follower dyads (Study 1), we found the expected interaction effect. Leaders’ neuroticism interacted with their followers’ conscientiousness and agreeableness to predict their evaluation of the desirability of followers’ proactive behaviour. Study 2, an experimental vignette study, suggests a moderated indirect effect through the experience of threat, but not worry. We found no direct effects of leader neuroticism on the desirability ratings of followers’ proactive behaviour. This research emphasizes the value of investigating the interplay between leader and follower traits and the underlying cognitive-emotional processes for leader evaluations of followers’ proactivity.

KEYWORDS:

• Proactive work behaviour
• neuroticism
• conscientiousness
• agreeableness
• personality trait interactions

Proactive employees initiate change or speak up and alert leaders to problems, which can improve decision making as well as the products and services offered by the business, and may ultimately convey a competitive advantage to the organization (Burris, 2012; Grant et al., 2009). It is important for leaders to see their followers’ initiatives as desirable because leaders may show their positive reactions to followers. These positive reactions, in turn, may signal to followers that their actions, opinions and ideas are valued and effective, which will likely stimulate further proactivity (e.g., De Stobbeleir et al., 2010; Morrison & Phelps, 1999). While employers usually expect followers to behave proactively, research suggests that in practice, proactivity is not always seen as desirable. Followers may not be supported and may even be penalized if their initiatives do not align with the values, preferences or interests of the leader (e.g., Burris, 2012; Fuller et al., 2015; Grant et al., 2009; Nguyen et al., 2017). Campbell (2000) labelled this the “initiative paradox”. Why do some leaders desire proactive behaviour while others discourage such behaviour? Several studies have started addressing this question (e.g., Fast et al., 2014; Fuller et al., 2015; Grant et al., 2009; Urbach & Fay, 2018). For instance, evidence suggests that leaders’ characteristics, such as their power motives and self-efficacy affect their evaluations of follower proactive behaviours (Fast et al., 2014; Fuller et al., 2015; Urbach & Fay, 2018). The aim of this study is to add to this research by exploring the role of leader neuroticism (i.e., the predisposition to easily experience negative emotions such as worry and anger) (McCrae & Costa, 1997) as a trait that affects their evaluation of the desirability of followers’ proactive behaviour.

To date, there is little scientific understanding of how leader neuroticism affects evaluations of follower proactive behaviour. Drawing on personality theory (Costa & McCrae, 1992a; Matthews, 2018; Watson & Casillas, 2003) and previous research (e.g., Bajcar & Babiak, 2020; Niemann et al., 2014), we argue that when leaders high in neuroticism experience ambiguous and potentially threatening behaviours shown by their followers, they activate negative emotional-cognitive responses such as worry, which likely result in negative evaluations of the reaction-inducing person and their behaviours (Burris, 2012; Niemann et al., 2014; Veling et al., 2008). As followers’ proactive behaviour is future- and change-oriented, its consequences are uncertain (e.g., will the initiative be effective and achieve its goal? Will it trigger social resistance and conflict?). This can create insecurity about outcomes (Grant & Ashford, 2008), and leaders with a high level of neuroticism may be especially sensitive to the uncertainties inherent in proactive behaviour and will likely experience them as threatening (Bajcar & Babiak, 2020). We therefore propose that leaders with high neuroticism will tend to be less appreciative of such follower behaviour.

Like leader characteristics, follower characteristics, such as responsibility, prosocial values, self-efficacy and trustworthiness, were also found to play a role in how leaders evaluate proactive behaviour on the part of their followers (e.g., Burris, 2012; Fuller et al., 2015; Grant et al., 2009; Nguyen et al., 2017; Urbach & Fay, 2018). Recent research suggests considering the roles of leader and follower traits and their interplay on leaders’ evaluations of follower behaviour (e.g., Wang et al., 2019), which is a thread we pick up in the current study. Based on the literature on the relevance of the traits of conscientiousness and agreeableness to both task-related and social workplace functioning such as job performance and helping behaviours (e.g., Bowling et al., 2011; Penney et al., 2011; Witt et al., 2002), we propose that the strength of the relationship between leaders’ neuroticism and their evaluation of the desirability of proactive behaviour is affected by followers’ conscientiousness and agreeableness. Followers need to alleviate leaders’ concerns to receive a positive leader assessment of their proactivity and leaders high in neuroticism are likely to have more such concerns. We argue that highly conscientious and agreeable followers are better able to convince their leaders high in neuroticism that their initiatives and suggestions are well planned and organized (signalling conscientiousness) and that they have considered the social impact of their suggestions on others so that their initiatives will not cause unnecessary conflict (signalling agreeableness). In contrast, followers low in conscientiousness and agreeableness may signal that they lack the necessary task-related skills (that rely on conscientiousness) and interpersonal skills (that rely on agreeableness) necessary to come up with good initiatives. This in turn causes leaders with a high level of neuroticism who are sensitive to such signals to believe that the actions of these followers will likely not meet their needs for security, social order, and harmony (Denissen & Penke, 2008; McCrae & Costa, 1997). Thus, we expect a three-way interaction of leader neuroticism and follower conscientiousness and agreeableness in relation to leaders’ evaluations of the desirability of follower proactivity.

Drawing on personality theory (Costa & McCrae, 1992a; Matthews, 2018; Watson & Casillas, 2003) and trait activation theory (Tett & Guterman, 2000), we further propose that feelings of worry and threat are activated in leaders with high neuroticism when they encounter followers low in conscientiousness and agreeableness demonstrating proactive work behaviour. These negative cognitive-emotional responses in turn, are assumed to result in leaders’ negative evaluations of the desirability of follower proactive behaviour (Burris, 2012; Niemann et al., 2014; Veling et al., 2008). Our conceptual model is depicted in Figure 1. We test the three-way interaction in a multi-source field survey study among leader-follower dyads (Study 1). In Study 2, an experimental vignette study, we examine the roles of worry and threat as two underlying mechanisms for the proposed relationship.

Figure 1. Research model.

Our paper contributes to the literature in several ways. First, it introduces the concept of leaders’ subjective evaluation of the desirability of followers’ proactivity and investigates why some leaders welcome followers’ proactive behaviour while others negatively evaluate such behaviour. Second, it adds to the scarce literature on neuroticism in leaders and on how this trait affects evaluations of followers. Third, by studying the role of leader-follower personality trait interactions, our study recognizes that personality traits do not act in isolation and that leaders’ evaluations of proactivity are also informed by follower characteristics. Finally, our second study highlights the underlying cognitive-emotional mechanisms that help explain how leader neuroticism interacts with follower traits to predict perceived desirability of follower proactivity.

Proactive behaviour in followers and leaders’ evaluations of its desirability

Being proactive is about making things happen, seizing opportunities, and preventing problems, and it involves “self-initiated efforts to bring about change in the work environment and/or oneself to achieve a different future” (Parker et al., 2010, p. 827). Followers can be proactive in different ways (Parker et al., 2019). Here, we focus on speaking up, voicing concerns, recommending modifications to procedures and suggesting or making changes, in other words voice (Morrison, 2011) and taking charge (Morrison & Phelps, 1999). Due to their challenging and change-oriented nature (see Morrison, 2011), we expect these proactive behaviours to stimulate particularly strong reactions in leaders with high neuroticism.

Followers show proactive behaviour when they believe it is important for themselves or their environment (Parker et al., 2019). However, follower perceptions of the importance and usefulness of proactivity are not always in line with others’ perceptions. Leaders may not understand a follower’s commitment to change and proactive followers may encounter suspicion and resistance in their social environment (e.g., Bolino et al., 2010; Frese & Fay, 2001). Proactive behaviour may deviate from the prescribed work roles or challenge the status quo and individuals may speak up even when others disagree (Frese & Fay, 2001). Indeed, some research suggests that proactive behaviour can be risky for the individual and can lead to personal costs, such as conflicts and negative performance evaluations (for an overview see Bolino et al., 2010; Parker et al., 2019).

Several studies examine the conditions under which leaders endorse follower proactivity and how follower proactivity affects leaders’ evaluations of follower performance (e.g., Burris 2012; Grant et al., 2009; Fuller et al., 2015; Nguyen et al., 2017; Urbach & Fay, 2018). Preliminary evidence suggests that leaders may perceive follower proactivity as desirable and endorse it when it reflects their own motives, goals or attitudes. For example, Fast et al. (2014) showed that due to their ego defensiveness, managers with low self-efficacy were less likely to support employee proactivity than managers high in self-efficacy. Here, we propose that a leader’s trait neuroticism influences how s/he evaluate the desirability of a follower’s proactivity.

Leader neuroticism and evaluating follower proactive behaviour

Personality traits refer to aspects of individuals’ thoughts and behaviours that are relatively consistent across different situations and stable over time, yet malleable to some degree (McCrae & Costa, 1997; Roberts & DelVecchio, 2000). A dominant view is that the basic structure of personality consists of five factors: openness to experience, agreeableness, conscientiousness, extraversion and neuroticism (McCrae & Costa, 1997).

Personality theory suggest that individuals with heightened neuroticism adopt a negativistic cognitive style, interpret ordinary situations and events as negative and frequently experience aversive cognitive-emotional states, such as worry, anger, and threat across various contexts (Matthews, 2018; McCrae & Costa, 1997; Watson & Casillas, 2003). Also, individuals high in neuroticism are characterized by difficulties in coping with and a heightened intolerance of uncertainty (Bajcar & Babiak, 2020; Berenbaum et al., 2008; Hirsh & Inzlicht, 2008). Uncertain or ambiguous and potentially risky stimuli or situations are specifically likely to activate negative emotional-cognitive responses in people high in neuroticism (Matthews, 1999, 2018). These individuals tend to apply defensive strategies to cope with situations they perceive as uncertain or unfavourable (Bajcar & Babiak, 2020; Niemann et al., 2014). For instance, individuals high in neuroticism were found to show a higher tendency to dislike people who provided them with negative feedback than individuals low in neuroticism (Niemann et al., 2014).

Additionally, individuals high in neuroticism view inaction more favourably than action (Ireland et al., 2015). Action that challenges the status quo may create uncertainty about outcomes and may thus be perceived as worrisome. Followers who choose to challenge the status quo by offering suggestions for change or taking initiative may generate outcome insecurity, leading to appraisals of such situations as threatening and causing negative responses such as worry in people high in neuroticism (Hirsh & Inzlicht, 2008; Matthews, 1999, 2018). In line with previous research showing that negative emotional-cognitive reactions that are likely in people high in neuroticism result in negative evaluations of the reaction-inducing person and their behaviour (Burris, 2012; Niemann et al., 2014; Veling et al., 2008), we hypothesize that leaders high in neuroticism will generally be less appreciative of proactive follower behaviour. Conversely, leaders low in neuroticism are less likely to be concerned about maintaining safety; they appraise new situations in a more favourable light and feel less stressed when confronted with uncertainty or ambiguity (Hirsh & Inzlicht, 2008). Consequently, we propose that they are generally more favourable towards change and evaluate follower proactive behaviour as more desirable.

Hypothesis 1: Leader neuroticism is negatively related to leader-rated evaluations of the desirability of proactive follower behaviour.

Leader and follower traits and the desirability of proactive behaviour

Past research has mostly focussed on the separate effects of leader or follower characteristics on evaluations of follower proactivity (e.g., Fuller et al., 2015; Grant et al., 2009; for an exception, see De Stobbeleir et al., 2010). Here, we follow recent recommendations to study the interplay between leader and follower characteristics in relation to leader evaluations (Lam et al., 2019; Wang et al., 2019). Previous work has shown that interactions of conscientiousness and agreeableness are important for workplace functioning and explain incremental variance in outcomes such as individual job performance ratings, team performance and helping behaviours (Bowling et al., 2011; Burke & Witt, 2004; for an overview, see Penney et al., 2011). We argue that leader neuroticism interacts with follower conscientiousness and agreeableness and that these jointly predict leaders’ evaluations of the desirability of followers’ proactive behaviour.

The trait of conscientiousness reflects the ability of an individual to be dutiful, responsible and thorough (Costa & McCrae, 1992a; Witt et al., 2002). High conscientiousness represents a tendency towards precision and organization, while individuals low in conscientiousness tend to be less organized, accurate and dependable. Consistent with the findings of Frei and McDaniel (1998), we expect that the behavioural results of conscientiousness (e.g., precision, being on time and organized) will be relatively concrete and readily observable by supervisors. Conscientiousness is a valid and consistent predictor of job performance (Barrick & Mount, 1991), and leaders see it (next to cognitive ability) as the most important trait when hiring new employees (e.g., Dunn et al., 1995).

Agreeableness reflects people’s interpersonal orientation and refers to how friendly, trusting, honest and cooperative they are (e.g., Costa & McCrae, 1992a). Individuals high in agreeableness value the maintenance of positive social relationships and are respectful and sensitive to the needs of others (Kalshoven et al., 2011). Individuals low in agreeableness are more self-focussed and self-interested; getting along with others well is less important to them; and they can be perceived as suspicious, uncooperative and unfriendly.

Individuals high in both conscientiousness and agreeableness are characterized as diligent, reliable and dutiful while also valuing communion and harmony at work. As noted, the interactive effect of these two traits in predicting work outcomes is widely acknowledged in the literature (e.g., Bowling et al., 2011; Burke & Witt, 2004; Schippers, 2014; Witt et al., 2002). For example, research shows that the positive association between conscientiousness and performance disappears for individuals low in agreeableness (Witt et al., 2002). Additionally, Burke and Witt (2004) show that conscientiousness strongly relates to job performance when agreeableness is high.

Similarly, we expect the interplay of follower conscientiousness and agreeableness to influence how leaders evaluate the desirability of proactive follower behaviour. We propose that low levels of both conscientiousness and agreeableness in followers could be especially detrimental when leaders with high neuroticism evaluate their proactive behaviour and that this is less so for leaders who score low in neuroticism. Individuals who are low in both conscientiousness and agreeableness might not think through proactive behaviour adequately and may base their behaviour on self-interest rather than communal needs (Bowling et al., 2011; Grant et al., 2009; Penney et al., 2011; Witt et al., 2002). Leaders high in neuroticism are more likely to perceive the behaviour shown by these followers as negative and worrisome, and they are less able to manage the ambiguity and uncertainty (Bajcar & Babiak, 2020; Berenbaum et al., 2008) relating to follower behaviours if they expect a lack of responsibility, dutifulness and low interpersonal competence. The combination of low conscientiousness and low agreeableness may affect desirability ratings of leaders low in neuroticism to a lesser extent, as these individuals tend to focus less on negative signals in their environment and are less likely to experience negative cognitive-emotional responses.

High levels of both follower conscientiousness and agreeableness may buffer the negative relationship between a leader’s neuroticism and evaluation of the desirability of proactive behaviour. Leaders with high neuroticism are likely to evaluate proactive behaviours less negatively when these behaviours are recognized in followers who tend to value performance standards, responsibility, carefulness (McCrae & Costa, 1997) (high conscientiousness), as well as harmony, cooperation and the proper treatment of others at work (high agreeableness). Hence, the negative association between leader neuroticism and the judgment of desirability is likely to be less pronounced or may even disappear when followers have high conscientiousness and agreeableness as through these traits the concerns of a leader with heightened neuroticism are addressed.

We anticipate a stronger negative relationship between leader neuroticism and the desirability ratings of follower proactive behaviour for the other two combinations of the follower traits of conscientiousness and agreeableness, although likely weaker than when both traits are low in a follower. For instance, when a follower is highly conscientious but low in agreeableness, leaders high in neuroticism might still worry about whether this follower takes the values and needs of others into account when engaging in proactive behaviour. If the follower is high in agreeableness but low in conscientiousness, the leader may still question whether this person’s proactive behaviour follows rules and is sufficiently organized. In sum, we propose a three-way interaction of leader neuroticism and follower conscientiousness and agreeableness in predicting the desirability of proactive behaviour as rated by leaders.

Hypothesis 2: The negative relationship between leader neuroticism and leader-rated desirability of follower proactive behaviour is moderated by follower conscientiousness and agreeableness. Compared to all other conditions, the negative relationship is most pronounced when both conscientiousness and agreeableness are low in followers.

Leaders’ feelings of worry and threat as mediators

In Study 1, we test the hypotheses presented above, and in Study 2, we further examine the role of two cognitive-emotional mechanisms that may explain this relationship. Specifically, we propose that for leaders with high neuroticism, feelings of worry and threat occur when they perceive followers low in both conscientiousness and agreeableness engaging in proactive behaviour, which leads them to evaluate these behaviours as less desirable. Worry is a feeling of concern about imagined or real future dangers. Threat is a person’s cognitive evaluation that a certain situation is potentially harmful or dangerous to them. Both worry and threat are trait-consistent cognitive-emotional manifestations of neuroticism (Matthews, 2018; Perkins et al., 2015; Watson & Casillas, 2003).

According to trait activation theory (Tett & Guterman, 2000), personality traits are conveyed as a reaction to trait-relevant situational cues. A trait is aroused by trait-relevant situational cues to manifest in behavioural expression and reaction (see also Urbach et al., 2016). We argue that observing followers low in consciousness and agreeableness engage in self-initiated, change-oriented behaviour, which challenges the status quo, offers cues for trait expressions in leaders because low levels of these traits reflect a lack of constructiveness and effectiveness of these followers’ suggestions and initiatives (Penney et al., 2011), and this results in threat and worry for leaders high in neuroticism. These negative cognitive-emotional reactions in turn are likely to affect leaders’ evaluation of the desirability of follower proactivity (Burris, 2012; Niemann et al., 2014; Veling et al., 2008).

We argue that worry and threat are less likely experienced by leaders high in neuroticism if proactive behaviour is shown by followers high in both traits because there is less reason for leaders to believe that behaviour is harmful when it is shown by followers who value responsibility and discipline and who consider others’ needs at work. Proactive behaviours should then be evaluated less negatively. We assume that when followers show any other combination of conscientiousness and agreeableness, leaders experience negative feelings of worry and threat, albeit to a weaker extent than when followers score low on both. Therefore, we hypothesize a mediated moderation in which the three-way interaction of leader neuroticism and follower conscientiousness and agreeableness predicts leaders’ feelings of worry and threat and, in turn, their rating of the desirability of proactive behaviour.

Hypothesis 3: The interaction of leader neuroticism with follower conscientiousness and agreeableness on leader ratings of the desirability of follower proactive behaviour is mediated by feelings of a) worry and b) threat. Leaders high in neuroticism are especially worried about and feel more threatened by the proactivity of followers who show both low conscientiousness and low agreeableness, resulting in lower evaluations of the desirability of follower proactivity.

We conducted a multi-source field study to test Hypotheses 1 and 2. Next, we ran an experimental vignette study to additionally examine the role of worry and threat as stated in Hypothesis 3. To maximize the internal validity of the results, it is important to exclude alternative explanations. The desirability ratings of proactivity may be influenced by the relationship quality between leaders and followers. Leader-member exchange theory (LMX) argues that leaders develop relationships of varying quality with different followers (Graen & Uhl-Bien, 1995). Leaders tend to see followers with whom they have developed a positive LMX relationship more favourably, which results in better evaluations of these followers (Erdogan & Enders, 2007). This general evaluation bias of leaders might also affect their perception of the desirability of proactive behaviour. Leaders may like follower suggestions only when these come from followers they have a good relationship with. To exclude this alternative explanation, we control for the influence of relationship quality in our studies.

Study 1

Method

Procedure and sample

We contacted leaders in various organizations in the Netherlands through established contacts and with the help of alumni students from a Business School. Leaders were provided with a printed survey, a stamped return envelope, and a letter explaining the study and the confidential and voluntary nature of participation. The leaders were asked to give a second survey package with a separate stamped return envelope and letter they received to one of their followers. Specifically, we asked them to randomly select one of their followers who was willing to fill in the follower survey. The researchers were available to answer any questions. Leaders and followers sent their responses directly to the researchers. Only one follower per leader participated. The respective leader and follower survey packages were marked with a unique dyad-specific ID to match the data. The raw data was then separated from the participant identifying information.

The final sample consisted of 205 leader–follower dyads (42% response rate). Everyone participated voluntarily and did not receive anything in return for participation. Respondents worked for a wide range of employers including consultancies, retailing companies, financial service providers, and the government. Of the followers, 53% were men, mean age was 30.7 years (SD = 8.8), and average organizational tenure was 4.4 years (SD = 5.2). Of the participating leaders, 67% were male, their average age was 40.1 years (SD = 8.2). Leaders and followers worked together on average 3.3 years (SD = 3.3).

Measures

Both leaders and followers completed some of the scales. All measures used a 7-point scale (1 = “completely disagree” to 7 = “completely agree”) except for the demographics. In the general instruction of the leader survey, it was emphasized that the leaders had to complete some items referring to themselves and some items with reference to the specific employee they selected. The items were translated from English to Dutch using the translation and back-translation procedure.

Leader neuroticism

We used four items by the Mini-IPIP scales (Donnellan et al., 2006) to measure neuroticism in leaders. Sample items are: “I have frequent mood swings” and “I am relaxed most of the time” (reverse coded).

Evaluation of the desirability of proactive behaviour

Leaders were provided with a definition and five descriptions of proactive behaviours from the taking charge measure by Morrison and Phelps (1999) and the voice measure by Van Dyne and LePine (1998) to give them concrete examples of proactive behaviours. Subsequently, we asked the leaders to evaluate to what extent they felt that such proactive behaviour shown by the specific follower they were rating is desirable for different entities based on five items. The instruction read: “Being proactive is about making sure that things happen, about anticipating and preventing problems and seizing opportunities. It concerns self-initiated efforts to bring about change in the work environment and/or to realize a different future. Examples of proactive behaviours are: Your employee tries to bring about improved procedures for the work unit or department. Your employee tries to eliminate redundant or unnecessary procedures. Your employee tries to implement solutions to pressing organizational problems. Your employee communicates his/her opinions about work issues even if his/her opinion is different and others disagree with him/her. Your employee speaks up and encourages others to get involved in issues that affect the organization. To what extent do you consider the proactive behaviour that your employee shows desirable for a) the organization? b) for the relationship between you and your employee? c) for the department or the team? d) for the relationship between colleagues? e) for the personal development of your employee?” We ran a confirmatory factor analysis (CFA) using Mplus 8.1 to test the factor structure of this new scale. Reasonable model fit was assumed for the comparative fit index (CFI) with values exceeding 0.90 and the standardized root mean square residual (SRMR) with values less than 0.08 (Hair et al., 2010; Hu & Bentler, 1998). The CFA confirmed the expected one-dimensional solution of the scale (χ2[5] = 20.258, p < .01; CFI = .972; SRMR = 0.027); all factor loadings exceeded .72 (all p = .000).

Follower conscientiousness

Followers rated their conscientiousness based on four items from the Mini-IPIP scales (Donnellan et al., 2006). Example items are “I make a mess of things” (reverse coded) and “I like order”.

Follower agreeableness

Followers self-rated their agreeableness based on four items from the Mini-IPIP scales (Donnellan et al., 2006). Example items are “I am not really interested in others” (reverse coded) and “I sympathize with others’ feelings”.

Control variables

We controlled for leader and follower age and gender (0 = male, 1 = female) (Fuller et al., 2015; Grant et al., 2009), and the length of time (in years) they worked together. The longer they worked together, the better the leader knows the follower, which might affect the desirability ratings (Blickle et al., 2013). In addition, we controlled for LMX assessed by the leader which was measured using the six items with the highest factor loadings by Liden et al. (1993). The reliability analysis showed that the corrected item-total correlations for two items were <.35 indicating that they did not correlate well with the overall scale. Similarly, the inter-item correlations of these items with the other four items ranged between .00 and .40. The scale reliability was relatively low (alpha = .70) and alpha increased from .70 to .78 by dropping these two items. Consequently, we decided to delete them and measure LMX using the remaining four items, for example: “I think that I understand my subordinate’s problems and needs.”

Results

First, we conducted an overall CFA. Results showed an adequate fit of the hypothesized five-factor structure (χ²[179] = 317.978, p < .001; CFI = .924; SRMR = 0.062). This model performed significantly better than the 4-factor model with the evaluation of the desirability of proactive behaviour and LMX loading on one factor (Δχ² [4] = 132.509, p < .001) and a two-factor model with leader neuroticism, leaders’ evaluation of the desirability and LMX loading on one factor and followers’ self-rated conscientiousness and agreeableness loading on a separate factor (Δχ² [9] = 580.708, p < .001). This suggests that the scales used in Study 1 are distinctive. The standardized factor loadings were significant (all p = .000) and ranged from .57 to .79 for leader neuroticism, .71 to .85 for desirability of proactive behaviour, .64 to .74 for LMX, .45 to .81 for follower conscientiousness, and .67 to .85 for follower agreeableness. Descriptives, inter-correlations, and Cronbach’s alphas are presented in Table 1. LMX as rated by the leader was positively related to the desirability of proactive behaviour (r = .47, p < .001). Leader neuroticism was not significantly associated with the desirability of proactive behaviour (r = −.05, ns).

Table 1. Means (M), Standard Deviations (SD), and correlations between the variables (Study 1)

	M	SD	1	2	3	4	5	6	7	8	9	10
1. Age leader	40.1	8.2	-
2. Gender leader	0.3	0.5	−.15*	-
3. Age follower	30.7	8.8	.30**	.14*	-
4. Gender follower	0.5	0.5	−.06	.28**	.18*	-
5. Years leader and follower working together	3.3	3.3	.34**	−.19**	.22**	−.09	-
6. Leader-member exchange	5.5	0.8	−.04	−.02	−.01	−.03	−.13	(.78)
7. Leader neuroticism	2.9	1.1	−.10	−.09	−.14*	−.04	−.04	−.19**	(.78)
8. Evaluation of the desirability of proactive behaviour	5.7	0.9	−.06	−.13	−.13	.08	−.13	.47**	−.05	(.89)
9. Follower conscientiousness	5.0	1.1	−.03	.03	.08	.15*	.00	.22**	−.08	.21**	(.78)
10. Follower agreeableness	5.1	1.1	.04	.03	.04	.21**	−.14*	.19**	.00	.21**	.34**	(.85)

N = 205 leader–follower dyads. Gender was coded 0 = male, 1 = female. Age follower, gender follower, and follower conscientiousness and agreeableness were follower ratings. The other variables were leader ratings. Reliability estimates (α) are shown in parentheses on the diagonal. *p < .05. ** p < .01.

To test our hypotheses, we applied multiple regression analysis using the PROCESS macro in SPSS (Hayes, 2017). The independent variable, moderators, and control variables were mean-centred prior to the analyses to better interpret the results (Cohen et al., 2003). Hypothesis 1 states that leader neuroticism is negatively related to the desirability of follower proactive behaviour. The results show that when the covariates were controlled, the coefficient for leader neuroticism on the desirability of proactive behaviour was non-significant (B = 0.01, SE = 0.05, t = 0.293, p = .770). Thus, Hypothesis 1 was not supported.

In line with Hypothesis 2, the results indicate a significant three-way interaction effect of leader neuroticism, follower conscientiousness and agreeableness on the desirability of proactive behaviour (B = −0.08, SE = 0.04, p= .021). Table 2 presents the results of this analysis. The variables explain an overall of 32.81% in the variance of desirability of proactive behaviour with the three-way interaction explaining 1.92% of variance above and beyond the main effects and the two-way interaction effects, which is statistically significant (F (1, 191) = 5.4511, p = .021). The significant three-way interaction is plotted in Figure 2. A simple slopes analysis revealed that the relationship between leader neuroticism and the desirability of follower proactive behaviour was negative when leaders evaluated followers who scored low (one standard deviation below the mean) in conscientiousness and agreeableness (B= −0.17, SE = 0.08, p = .046; slope 4 in Figure 2). Leader neuroticism was unrelated to the desirability ratings in followers with high (one standard deviation above the mean) conscientiousness and agreeableness (B = 0.05, SE = 0.08, p = .504; slope 1 in Figure 2), or with either high conscientiousness or agreeableness and low values on the other (B = 0.18, SE = 0.12, p = .123; slope 2 and B = 0.08, SE = 0.10, p = .431; slope 3). Thus, the negative effect of leader neuroticism on desirability ratings is only evident when followers are low in both conscientiousness and agreeableness, which partially supports Hypothesis 2. The study results were equivalent (i.e., non-significant relationship between leader neuroticism and significant three-way interaction at p < .05) when we did not control for leader and follower age and gender, the length of time they worked together, and LMX. We kept the control variables to offer a comprehensive and conservative presentation of the results.

Table 2. Results of Regression Analysis with the Evaluation of the Desirability of Proactive Behaviour as Dependent Variable Testing Hypothesis 2 (Study 1)

Predictor variable	B	SE	95% CI	t	p
Age leader	0.00	0.01	[–0.01, 0.02]	0.25	.800
Gender leader	–0.33	0.12	[−0.57, −0.09]	-2.70	.008
Age follower	–0.01	0.01	[−0.02, 0.00]	-1.49	.138
Gender follower	0.22	0.12	[−0.00, 0.45]	1.97	.051
Years leader and follower working together	–0.02	0.02	[–0.06, 0.01]	-1.18	.241
Leader-member exchange	0.47	0.08	[0.32, 0.67)]	6.25	.000
Leader neuroticism	0.04	0.05	[–0.06, 1.40]	0.76	.448
Follower conscientiousness	0.06	0.06	[–0.05, 0.17]	1.14	.138
Follower agreeableness	0.03	0.06	[–0.08, 0.14]	0.53	.599
2-way interaction effects
Leader neuroticism x Follower conscientiousness	0.08	0.05	[–0.02, 0.17]	1.61	.109
Leader neuroticism x Follower agreeableness	0.03	0.05	[–0.07, 0.13]	0.55	.584
Follower conscientiousness x Follower agreeableness	0.01	0.05	[–0.08, 0.10]	0.29	.770
3-way interaction effect
Leader neuroticism x Follower conscientiousness x Follower agreeableness	-0.08	0.04	[–0.15, –0.01]	-2.34	.021
Total R^2	.328

Note. N = 205. B = Unstandardized regression coefficient; SE = Standard error. Gender was coded 0 = male, 1 = female. Total R² = Proportion of variance explained in the criterion associated by all the variables in the model. Variables that are used to calculate the interaction terms and the control variables are mean-centred. Bootstrap sample size = 5,000. * p < .05. ** p < .01.

Figure 2. Three-way interaction of leader neuroticism, follower conscientiousness and agreeableness predicting the evaluation of proactive behaviour in followers (Study 1).

Study 2

Method

Sample

We used the crowd-working platform Prolific to recruit participants for this experimental vignette study. Previous research showed that this platform is suitable for participant recruiting, delivers high data quality, and has some advantages to other platforms (Palan & Schitter, 2018). We applied the pre-screening option and invited 385 participants from the UK who work part- or full-time and hold a leadership position. Participants were reimbursed £1.15 for survey completion. One item was added to check whether respondents pay attention to the instructions in the survey. Five participants failed the attention check and were removed from further analyses (Oppenheimer et al., 2009), leaving a final sample of 380 participants. Of these, 196 (51.6%) were female, their mean age was 39.8 years (SD = 10.8), ranging from 19 to 71 years, and most participants (71.3%) held a university degree. Participants worked on average 38.8 hours (SD = 8.2) in various industries, including education, finance, production, and health and social welfare. Participants had been supervisor on average 7.3 years (SD=6.9) and supervised on average 10.2 employees (SD = 23.5).

Procedure and manipulations

Participants started answering some demographic items, rated their neuroticism, and then read a description of a work situation in which they had to imagine that they were working as a team leader who is preparing for the annual employee performance appraisal interviews. One fictitious male follower, Martin, was described as a follower who has been very proactive at work in the past. Specifically, the description read:

“You noted that during the past 12 months, Martin has been very proactive at work. He communicated his opinions and concerns about work issues even if other colleagues disagreed with him. For example, Martin suggested some modifications to the customer database system the company uses with the aim to increase the company’s customer orientation and, ultimately, customer satisfaction. Some colleagues were not pleased with these suggestions though as they implied extra work for them.”

We manipulated conscientiousness and agreeableness of the fictitious follower described in the vignette and used a two-way between-subjects factorial design (conscientiousness: high vs. low; agreeableness: high vs. low). Participants were randomly assigned to one of the four conditions and each condition comprised 93–97 participants. Leader neuroticism was treated as a continuous factor (for a similar approach, see Urbach & Fay, 2018).

In the high conscientiousness condition, Martin was described as: “In general, Martin is an organized person. He is always well-prepared and prefers making a thorough plan before acting. He is a diligent, hard-working employee who is attentive and persistent in working on his tasks.” In the low conscientiousness condition, participants read: “In general, Martin tends to be a disorganized person. He is flexible and prefers spontaneous behaviour to planned behaviour. While working on his tasks, he is easily distracted and tends to be less persistent.” The high agreeableness condition read: “Besides, he is characterized as a friendly and compassionate person who enjoys cooperating with others. He is eager to maintain positive social relationships and is sensitive to the needs and interests of others.” The low agreeableness condition read: “Besides, he is characterized as a self-focussed and competitive person who is sometimes perceived by others as being unfriendly. He is hardly interested in and sensitive to other people’s needs and interests. Getting along with others well is not very important for him.”

We then measured the level of perceived threat and worry that supervisors experienced as reactions to the behaviour shown by Martin and asked them to evaluate the desirability of Martin’s behaviour. Then they were debriefed and thanked for participating.

Measures

Manipulation checks

We used the Mini-IPIP scales (Donnellan et al., 2006) to check whether our manipulation influenced participants’ assessment of Martin’s conscientiousness and agreeableness. After reading the scenarios, we asked participants “Based on this short description, how would you characterize Martin?”. An example item for agreeableness is “Martin is not interested in other people’s problems” (recoded; Cronbach’s alpha = .84); and for conscientiousness: “Martin likes order” (Cronbach’s alpha = .71). The items were rated on a 5-point scale ranging from 1 (“completely disagree”) to 5 (“completely agree”).

Leader neuroticism

We applied the same four items based on the Mini-IPIP scales (Donnellan et al., 2006) as in Study 1 to measure self-rated neuroticism in leaders.

Worry

Two items based on the Worry Penn State Questionnaire (Meyer et al., 1990) were used to assess how leaders feel about the behaviour shown by Martin. Participants indicated how much each of these statements applies (1 = “does not apply at all”; 5 = “fully applies”). The items were “I feel worried about Martin’s proactive behaviour” and “Martin’s proactive behaviour makes me worry”.

Threat

We used two self-developed items based on Sijbom et al. (2015) to assess leaders’ level of perceived threat. The items were “I feel threatened by Martin’s proactive behaviour” and “Martin’s proactive behaviour undermines my position as a leader” (1 = “does not apply at all”; 5 = “fully applies”).

Evaluation of the desirability of proactive behaviour

We used the same 5 items as in Study 1 which were adapted to align with the context in the experimental vignette. The items read: “To what extent do you feel that Martin’s proactive behaviour is desirable for the following entities? Martin’s proactive behaviour is desirable a) for the organization, b) for the department or the team, c) for the relationship between team members, d) for the relationship between Martin and me, e) for Martin’s personal development (1 = “completely disagree” to 7 = “completely agree”)?. A CFA was conducted to test the factor structure. The results confirmed its one-dimensional solution (χ2(5) = 85.623, CFI = 0.902, SRMR = 0.052). The factor loadings ranged between .60 and .86 (all p = .000).

Control variables

We controlled for leader gender (0 = male; 1 = female), age and tenure as leader in the current organization (in years). We further measured the extent to which participants liked Martin as this might influence the leaders’ evaluation of the behaviour Martin shows. Leaders may be more likely to desire follower ideas and suggestions when these come from followers they like more. Participants indicated how much they agreed (1 = “not at all” to 5 = “extremely”) with: “How much do you like Martin?”, “How much do you sympathize with Martin?”, “Would you like working together with Martin?”.

Results

Participants in the high conscientiousness conditions perceived the follower in the vignette as higher in conscientiousness (M = 4.2, SD = 0.5) compared to participants in the low conscientiousness conditions (M = 2.6, SD = 0.7), F (1, 378) = 649.497, p = .000, η2 = .632. Participants in the high agreeableness conditions rated the follower as higher in agreeableness (M = 3.8, SD = 0.7) than those in the low agreeableness conditions (M = 2.1, SD = 0.7), F (1, 378) = 532.913, p = .000, η2 = .585. Thus, the manipulation of conscientiousness and agreeableness was successful. Descriptives, inter-correlations, and Cronbach’s alphas are presented in Table 3. The evaluation of the desirability of proactive behaviour was strongly and positively related to liking (r = .62, p < .001) and negatively to worry (r = −.43, p < .001) and threat (r = −.30, p < .001).

Table 3. Means (M), Standard Deviations (SD), and correlations between the variables (Study 2)

	M	SD	1	2	3	4	5	6	7	8	9	10
1. Age	39.8	10.8	-
2. Gender	0.5	0.5	.12*	-
3. Tenure leader position (in years)	7.3	6.9	.61**	.03	-
4. Liking	3.1	0.8	−.04	−.08	−.01	(.84)
5. Leader neuroticism	2.7	0.9	−.08	.13**	−.08	−.08	(.82)
6. Worry	1.8	0.9	.04	−.01	−.02	−.39**	−.06	(.91)
7. Threat	1.5	0.7	−.07	−.03	−.10*	−.25**	.13*	.55**	(.80)
8. Evaluation of the desirability of proactive behaviour	4.8	1.1	−.07	.05	−.04	.62**	−.00	−.43**	−.30**	(.85)
9. Follower conscientiousness^a	-	-	−.05	.02	−.02	.35**	−.04	−.21**	−.06	.26**	-
10. Follower agreeableness^b	-	-	−.06	−.08	−.10*	.43**	−.00	−.21**	−.09	.36**	.02	-

N = 380. Gender was coded 0 = male, 1 = female.

^aExperimental factor 0 = low conscientiousness, 1 = high conscientiousness.

^bExperimental factor 0 = low agreeableness, 1 = high agreeableness. Reliability estimates (α) are shown in parentheses on the diagonal.

*p < .05. ** p < .01.

To test the main effect of leader neuroticism on the evaluation of the desirability of Martin’s proactive behaviour as proposed in Hypothesis 1, we ran a multiple regression analysis controlling for gender, age, tenure as leader in the current organization and liking. The effect was not significant (B = 0.03, SE = 0.05, t = 0.694, p = .488). Thus, Hypothesis 1 was not supported.

Next, we tested the mediated moderation outlined in Hypothesis 3, namely that there is a three-way interaction effect between the leader and follower traits that is mediated by leaders’ feelings of worry and threat, by running a conditional process analysis (see Tables 4 and 5). PROCESS generates 95% bootstrap confidence intervals for the indirect effect using 5,000 bootstrap samples. The index of moderated mediation for feelings of worry was not significant, the 95% confidence interval contained zero (B = −0.03, SE = 0.05, 95% CI [−0.14, 0.06]). The non-significant conditional indirect effects for different values of conscientiousness and agreeableness and worry as mediator are presented in Table 4. The results showed a significant indirect effect of the three-way interaction through leaders’ feelings of threat (B = −.10, SE = 0.05, 95% CI [−0.19, −0.02]). Specifically, threat was predicted by the three-way interaction effect of leader neuroticism, follower conscientiousness and agreeableness (B = 0.41, SE = 0.16, p = .010), and threat was significantly related to leaders’ evaluation of the desirability of proactive behaviour (B = −0.24, SE = 0.06, t = −3.71, p < .001).

Table 4. Results of the Conditional Process Analysis Testing Hypothesis 3 with Worry as Mediator (Study 2)

Predictor variable	B	SE	95% CI	t	p
DV: Worry
Age leader	0.00	0.01	[–0.02, 0.01]	0.80	.426
Gender leader	–0.07	0.08	[−0.23, 0.10]	-0.78	.434
Tenure leader position (in years)	–0.01	0.01	[–0.00, 0.01]	-1.13	.258
Liking	–0.36	0.06	[–0.48, –0.25]	–6.01	.000
Leader neuroticism	–0.08	0.04	[–0.16, 0.01]	–1.79	.075
Follower conscientiousness	–0.15	0.09	[–0.32, 0.03]	–1.67	.096
Follower agreeableness	–0.11	0.09	[–0.29, 0.07]	1.22	.224
2-way interaction effects
Leader neuroticism x Follower conscientiousness	0.13	0.09	[–0.04, 0.30]	1.46	.145
Leader neuroticism x Follower agreeableness	0.13	0.09	[–0.05, 0.30]	1.43	.154
Follower conscientiousness x Follower agreeableness	–0.10	0.16	[–0.42, 0.22]	–0.63	.529
3-way interaction effect
Leader neuroticism x Follower conscientiousness x Follower agreeableness	0.12	0.18	[–0.23, 0.47]	0.68	.499
Total R^2	.186
DV: Evaluation of the desirability of proactive behaviour
Age leader	–0.00	0.01	[–0.01, 0.01]	–0.70	.482
Gender leader	0.21	0.09	[0.03, 0.38]	2.27	.024
Tenure leader position (in years)	–0.00	0.01	[–0.02, 0.01]	–0.37	.716
Liking	0.76	0.06	[0.65, 0.88]	12.82	.000
Leader neuroticism	0.01	0.05	[–0.09, 0.11]	0.23	.821
Worry	–0.28	0.06	[–0.39, –0.17]	–5.06	.000
Total R^2	.439
Conditional indirect effect at +/- 1 SD
Moderators	Unstand. boot indirect effect		Boot SE	Boot LLCI	Boot ULCI
Low C., low A.	0.05		0.03	–0.02	0.13
Low C., high A.	0.03		0.02	–0.01	0.08
High C., low A.	0.03		0.03	–0.02	0.09
High C., high A.	–0.02		0.02	–0.06	0.02

Note. N = 380. B = Unstandardized regression coefficient; SE = Standard error. Gender was coded 0 = male, 1 = female. DV = dependent variable. C. = follower conscientiousness, A. = Follower agreeableness. Total R² = Proportion of variance explained in the criterion associated by all the variables in the model. Variables that are used to calculate the interaction terms and the control variables are mean-centred. Bootstrap sample size = 5,000. * p < .05. ** p < .01.

Table 5. Results of the Conditional Process Analysis Testing Hypothesis 3 with Threat as Mediator (Study 2)

Predictor variable	B	SE	95% CI	t	p
DV: Threat
Age leader	–0.00	0.00	[–0.11, 0.01]	–0.50	.621
Gender leader	–0.11	0.08	[−0.25, 0.04]	–1.45	.149
Tenure leader position (in years)	–0.01	0.01	[–0.02, 0.01]	–1.12	.264
Liking	–0.25	0.05	[–0.36, –0.14]	–4.61	.000
Leader neuroticism	0.09	0.04	[0.01, 0.16]	2.19	.029
Follower conscientiousness	0.06	0.08	[–0.09, 0.22]	0.80	.425
Follower agreeableness	0.03	0.08	[–0.13, 0.20]	0.41	.682
2-way interaction effects
Leader neuroticism x Follower conscientiousness	0.06	0.08	[–0.09, 0.22]	0.81	.421
Leader neuroticism x Follower agreeableness	0.01	0.08	[–0.15, 0.16]	0.08	.937
Follower conscientiousness x Follower agreeableness	0.10	0.15	[–0.18, 0.39]	0.74	.482
3-way interaction effect
Leader neuroticism x Follower conscientiousness x Follower agreeableness	0.41	0.16	[0.10, 0.72]	2.59	.010
Total R^2	.109
DV: Evaluation of the desirability of proactive behaviour
Age leader	–0.01	0.01	[–0.02, 0.01]	–0.98	.325
Gender leader	0.20	0.09	[0.02, 0.38]	2.19	.029
Tenure leader position (in years)	–0.00	0.01	[–0.02, 0.01]	–0.37	.716
Liking	0.83	0.06	[0.72, 0.94]	14.49	.000
Leader neuroticism	0.05	0.05	[–0.04, 0.15]	1.10	.270
Threat	–0.24	0.06	[–0.36, –0.11]	–3.71	.000
Total R^2	.423
Conditional indirect effect at +/- 1 SD
Moderators	Unstand. boot indirect effect		Boot SE	Boot LLCI	Boot ULCI
Low C., low A.	–0.05		0.02	–0.09	-0.01
Low C., high A.	0.01		0.02	–0.03	0.05
High C., low A.	0.00		0.02	–0.04	0.03
High C., high A.	–0.05		0.02	–0.10	–0.01

Note. N = 380. B = Unstandardized regression coefficient; SE = Standard error. Gender was coded 0 = male, 1 = female. DV = dependent variable. C. = follower conscientiousness, A. = Follower agreeableness. Total R² = Proportion of variance explained in the criterion associated by all the variables in the model. Variables that are used to calculate the interaction terms and the control variables are mean-centred. Bootstrap sample size = 5,000. * p < .05. ** p < .01.

The three-way interaction explained 1.62% of the variance in threat above and beyond the main effects and the two-way interaction effects. The interaction effect on threat is depicted in Figure 3. A simple slopes test revealed that in line with our assumptions, the relationship between leader neuroticism and threat is positive when the fictitious follower was described as low (one standard deviation below the mean) in both conscientiousness and agreeableness (B = 0.16, SE = 0.08, p = .045; slope 4 in Figure 3). As an unexpected result, leader neuroticism was also positively related to threat when the follower was described as high in both conscientiousness and agreeableness (B = 0.22, SE = 0.08, p = .006; slope 1 in Figure 3). Relationships were non-significant in the other two conditions (slopes 2 and 3).

Figure 3. Three-way interaction of leader neuroticism, follower conscientiousness and agreeableness predicting leaders’ feelings of threat (Study 2).

The conditional indirect effect via threat was negative and significant when the follower was described as low in both conscientiousness and agreeableness (B = −0.05, SE = 0.02, 95% CI [−0.09, −0.01]) as well as when the follower was described as high in conscientiousness and agreeableness (B = −0.05, SE = 0.02, 95% CI [−0.10, −0.01]). Thus, leaders high in neuroticism feel more threatened by followers who are either low in both conscientiousness and agreeableness or high in both, and in turn rate the desirability of these followers’ proactive behaviour as lower. These results tentatively and partly support Hypothesis 3 regarding threat as a mediator.

Additional analyses

We ran mediation analyses to explore the role of feelings of worry and threat as potential mechanisms in the relationship between leader neuroticism and desirability ratings. Results reveal that leader neuroticism was positively related to the perception of threat (B = 0.08, SE = 0.04, t = 2.11, p = .035) and threat predicted a lower evaluation of the desirability of proactive behaviour shown by the follower as described in the vignette (B = −0.24, SE = 0.06, t = −3.71, p < .001). The indirect effect through threat perceptions was negative and significant (B = −0.02, SE = 0.01, 95% CI [−0.05, −0.01]). The indirect effect for worry was not significant (B = 0.02, SE = 0.02, 95% CI [−0.00, 0.05]). The results for all analyses in Study 2 did not change if the control variables were removed.

General discussion

Summary and research implications

There has been increasing interest in research on proactive work behaviour and how it is evaluated by leaders. The so-called “initiative paradox” (Campbell, 2000) suggests that while employee proactivity is often expected in organizations, leaders may not always appreciate this form of self-initiated change- and future-oriented behaviour (e.g., Belschak et al., 2010; Bolino et al., 2010). Our study builds on this research by investigating why some leaders desire proactive behaviour while others evaluate such behaviour negatively.

We addressed whether leaders’ neuroticism negatively affects their evaluation of the desirability of followers’ proactive behaviour and assessed the role that follower conscientiousness and agreeableness play in this evaluation. Contrary to our expectations, the results from both studies suggest that leaders high in neuroticism are not generally less appreciative of proactivity than leaders with low levels of neuroticism. One explanation for the non-significant direct effect is that follower proactivity may be ambivalent for leaders high in neuroticism. On the one hand, leaders might feel threatened by the uncertain outcome and the potential negative effects of follower proactivity, and threat may lower their evaluation of the desirability of follower proactivity. This is consistent with the results from the additional analyses from Study 2 and further aligns with arguments from personality theory (Costa & McCrae, 1992a; Matthews, 2018; Watson & Casillas, 2003) and previous research showing that the negative cognitive-emotional mechanism that are likely experienced by people high in neuroticism result in negative evaluations and attitudes (Burris, 2012; Niemann et al., 2014). On the other hand, leaders might sometimes rely on and appreciate followers’ initiatives in order to create needed constructive change because leaders high in neuroticism tend to score lower on proactivity themselves (Tornau & Frese, 2013).

Our findings suggest that the relationship between leader neuroticism and their desirability of follower proactivity evaluations is more complex. The results show that leaders high or low in neuroticism evaluated follower behaviour differently depending on follower characteristics. The negative relationship between leader neuroticism and desirability ratings was evident when the proactive behaviour was exhibited by followers who are less dutiful, careless (low conscientiousness) and show a lack of cooperation and sensitivity towards the needs and interests of others (low agreeableness) (Study 1). In Study 2, we found more specifically that this trait interaction particularly affected leaders’ feelings of threat, which were negatively linked to their evaluation of the follower’s proactive behaviour. In other words, in line with trait activation theory (Tett & Guterman, 2000), leaders high in neuroticism, compared to those low in this trait, are more likely to feel threatened by followers low in conscientiousness and agreeableness. Thus, threat acts as a trait-consistent, negative cognitive-emotional state that manifests in leaders high in neuroticism who experience proactive behaviour from followers low in conscientiousness and agreeableness.

The threat measure used in Study 2 included an item on perceived status threat due to power loss. While in our theoretical arguments we referred to feelings of threat due to the potential negative effects of “misguided” follower proactivity, leaders may also perceive threat to their status and power as a result of a direct challenge to their authority from a proactive follower. This is an aspect of threat to leaders that was not considered in our theoretical argumentation. Specifically, the results from Study 2 support our expectation that the relationship between leader neuroticism and perceived threat is positive when the follower is described as low in both conscientiousness and agreeableness. In addition, leaders with high levels of neuroticism also reported feeling threatened when a follower who was high in both conscientiousness and agreeableness undertook proactive behaviour. Albeit unexpected, this finding is still in line with trait activation theory (Tett & Guterman, 2000). Feelings of threat may manifest for different reasons, depending on the personality trait compositions of followers. In line with our hypothesis, followers low in both traits may trigger threat because they send signals that suggest low constructiveness and effectiveness of their proactive behaviours. Followers high in both traits might trigger feelings of threat because these followers are perceived by leaders high on neuroticism as challenging their competence and authority, that is, making the leader look passive and weak, or even as becoming potentially competent and serious rivals who may threaten the leader’s status and power (Burris, 2012; Lam et al., 2019; Urbach & Fay, 2018). In both situations, leaders seem to feel threatened, react defensively and evaluate followers’ proactive behaviours as less desirable. These findings offer a valuable contribution to the literature and could stimulate future research on the different sources of threat perceptions and leaders’ reactions to them.

While our findings in Study 2 supported the idea of neuroticism as a threat-sensitive trait and as a mechanism in the studied relationships, feelings of worry did not act as such a mechanism. This result aligns with the findings provided by Weiss and Deary (2020), which suggest that neuroticism has different facets. They found that high levels of neuroticism are not necessarily always reflected in high levels of worry but could also be driven by high scores in anxiety and tension instead, which aligns with our threat measure. Future research should therefore use more extensive measures of neuroticism, such as the Neo-PI-R (Costa & McCrae, 1992b), to assess its various lower-order facets (e.g., anxiety, hostility and depression) and to examine whether particular effects might be driven by more specific facets of neuroticism (Bowling et al., 2011; Weiss & Deary, 2020).

Our paper contributes to previous research on the antecedents of leader evaluations of follower behaviour in various respects. First, it adds to the scarce literature on neuroticism in leaders and the consequences this trait has for leader evaluations of their followers. Despite some interest in the leadership literature in studying leader personality traits such as neuroticism (Judge et al., 2009) and several studies showing negative relationships of neuroticism with leader emergence, effectiveness and managerial performance (e.g., Cavazotte et al., 2012), there is a lack of research on how leaders’ neuroticism relates to the manner in which they evaluate their followers. Instead, research has mainly focussed on how neuroticism in followers determines their reactions towards leaders’ behaviours (e.g., De Hoogh & Den Hartog, 2009). Our results show that leader neuroticism, in interaction with follower traits, seems to play a role in leaders’ evaluation of their followers.

Second, by taking follower characteristics into consideration and studying the role of leader-follower personality trait interactions, the present study recognizes that personality traits do not act in isolation and that leaders’ evaluations of proactivity are informed by follower characteristics. By employing this perspective, our study helps explain why leaders evaluate and react differently to followers’ proactive behaviour. Our study further emphasizes the importance of the personality trait interaction of employee conscientiousness and agreeableness, which has so far, only been investigated in relation to work performance and well-being outcomes (Blickle et al., 2013; Bowling et al., 2011; Penney et al., 2011; Witt et al., 2002). Our findings suggest that these effects seem to generalize to a broader range of outcomes.

Finally, by studying the role of threat and worry as mediators, our research enhances understandings of the underlying cognitive-emotional mechanisms that may explain how leader neuroticism interacts with follower traits to predict desirability ratings. The results suggest that feelings of threat (rather than worry) play a role here and that leaders might feel threatened for different reasons, which is a promising area for future research.

Limitations and future research

In Study 1, we asked followers to provide self-report ratings of their traits to avoid common method bias. However, this approach could raise the concern that self-ratings of traits might differ from what leaders perceive because leaders’ perceptions of follower conscientiousness and agreeableness might not match with followers' self-reports (Lam et al., 2019). In Study 2, we addressed this shortcoming and expanded our measure of follower traits. Specifically, we manipulated follower traits by providing descriptions of tangible behaviour that can be easily observed. Leaders use these cues of behavioural manifestations to form judgements about follower personality traits (Funder & Sneed, 1993).

Even though we used different approaches and designs across the two studies and controlled for the role of LMX and liking in desirability ratings to examine our hypotheses, further work is required to replicate the findings. For instance, although the data for our two studies were collected in different countries (the Netherlands and the UK), both countries are highly individualistic, which might affect people’s reactions to proactivity (e.g., Urbach et al., 2021). Replication studies should be conducted to investigate potential differences and similarities in samples that vary in terms of organizational and socio-cultural contexts. As proactivity might be rewarded in some contexts and punished in others (Parker et al., 2019), contextual and cultural differences may affect how desirable leaders rate follower proactivity.

We focussed on voice and taking charge behaviours that might be more assertive and threatening to supervisors than other proactive behaviours such as job crafting or feedback-seeking. Future research could therefore examine other proactive behaviours to further understand the factors determining leader evaluations of follower proactivity. Besides, future research is necessary to further validate our measure of desirability of follower proactivity. Finally, the consequences of leaders’ evaluations of proactivity for performance-related follower outcomes, such as rewards and development opportunities for these followers (De Stobbeleir et al., 2010), and changes in proactive behaviour shown by followers over time, are other important areas for future research.

Practical implications

In dynamic organizational environments, employees who voice concerns, improve work processes and take charge to resolve issues, bring value to the organization (Morrison & Phelps, 1999; Parker et al., 2019). Yet, our results tentatively suggest that leaders high in neuroticism may feel threatened and evaluate proactive behaviour as less desirable when shown by followers who are low in conscientiousness and agreeableness. In Study 2, we find that this may also occur when followers are high in both traits. These results point to the importance of informing and sensitizing leaders high in neuroticism about their tendency to provide biased ratings in evaluating the desirability of follower proactivity. Encouraging leaders to support proactivity in followers is important because these behaviours have positive consequences for organizations (Grant et al., 2009). Our findings further suggest that followers can, to some extent, influence how their proactive behaviour is appreciated by leaders by managing the signals they send to their supervisors. Signalling low self-discipline, low thoroughness and less prosocial and trustful behaviours increases the chance that leaders evaluate followers’ initiatives negatively and should therefore be avoided by proactive followers.

Disclosure statement

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