The influence of moral leadership (in)congruence on negative workplace gossip is explained by perceived hypocrisy Authors: Qingling Chen, Zhaohui Liu, Guanfeng Shi Published: 2026 ABSTRACT Extant research positions moral leadership as a cornerstone of virtuous leadership, emphasizing its favorable impact on employees’ assessments of leaders. Nevertheless, such studies often overstate its advantages by conflating the personal (moral person) and managerial (moral manager) dimensions of moral leadership, while neglecting its potential drawbacks. This research bridges leadership identity theory and ethics scholarship to investigate both aspects. Employing polynomial regression and response surface analysis, it demonstrates that alignment between a leader’s moral person and moral manager identities is associated with lower negative workplace gossip. Leaders proficient in both dimensions experience the lowest levels of gossip, whereas those with a high moral person-low moral manager profile face less gossip than the inverse scenario. Additionally, this study examines perceived leader hypocrisy as a mediating mechanism, illuminating the potential drawbacks of moral and virtuous leadership. Introduction In the realm of value-based leadership research, while servant leadership has gained prominence for its emphasis on altruistic motives and follower growth, moral leadership offers a more robust theoretical framework for examining the intricacies of identity alignment1. In contrast to the service - oriented emphasis of servant leadership, moral leadership presents a structurally rigorous dual - pillar framework. This framework differentiates a leader’s internal virtues, which pertain to the moral person dimension and reflect integrity and fairness, from their external managerial functions, which fall under the moral manager dimension and signify the proactive guidance of organizational ethical conduct via communication and reinforcement2,3. This structural deconstruction not only facilitates a more granular analysis of the interplay between personal character and professional roles but also reinforces the necessity of functioning as a unified moral entity4,5. Specifically, this resonates with the concept of “Xiuji-Anren” (Cultivate Oneself to Benefit Others), which emphasizes that “gentlemen” should focus on moral self-cultivation while guiding followers to adhere to organizational moral norms. To address the concern regarding incremental contribution, this study departs from the traditional “level-based paradigm”—which focuses solely on the absolute intensity of leadership behaviors—and introduces an “alignment-based paradigm”. However, despite this theoretical potential, extant research has frequently examined these dimensions in isolation, overlooking the critical role of their identity congruence and its collective impact on subordinates6. We contend that leadership effectiveness stems not merely from the intensity of individual dimensions, but from the alignment between a leader’s moral character and their managerial practices. Such identity congruence serves as a pivotal psychological signal that shapes employee evaluations—specifically regarding perceived leader hypocrisy—and subsequent behavioral reactions, such as negative workplace gossip about supervisor. By focusing on the “credibility gap” created by identity misalignment, this research offers a substantial theoretical advancement over unidimensional or purely additive models of ethical influence. Extant research in moral psychology underscores the inherent complexities between internal moral beliefs and enacted behaviors, noting that external actions are often shaped by reputational concerns or moral compensation rather than intrinsic values alone6–8. Because internal beliefs are unobservable, evaluators heavily rely on behavioral consistency to judge genuine morality9. Translating these mechanisms into organizational leadership, a leader’s internal virtues are conceptually captured by the moral person dimension, while their enacted ethical enforcement aligns with the moral manager dimension. However, current scholarship frequently treats these domains as isolated or purely additive constructs, fundamentally overlooking their structural congruence6. By evaluating leadership through a unidimensional lens, existing literature fails to capture how discrepancies between a leader’s personal character and actual management practices send conflicting signals to subordinates. To bridge this theoretical gap, this study introduces a “congruence-based paradigm.” We argue that subordinate evaluations are highly sensitive to the structural alignment between who a leader is and what they formally enforce. By examining this specific dimensional congruence, we clarify how moral incongruence breeds perceived leader hypocrisy and subsequently fuels negative workplace gossip about supervisor. Addressing a critical gap in moral leadership research, this study draws on leadership identity construction theory10 to examine how moral person-manager congruence shapes subordinate reactions. Because a leader’s moral identity is dynamically constructed through the interplay of personal values, managerial actions, and subordinate evaluations10, employees rely heavily on behavioral consistency. Consequently, moral incongruence fractures this identity construction, triggering perceived leader hypocrisy11,12. To capture the behavioral fallout of such hypocrisy, we specifically investigate negative workplace gossip about supervisor. While prior research typically evaluates overt outcomes—such as formal voice, trust erosion, or reduced organizational citizenship behavior1,13—these metrics fail to capture the unique threat of hypocrisy. Because hypocritical leaders maintain a facade of virtue while wielding formal authority, inherent power asymmetries render overt confrontation excessively risky14,15. Instead, employees strategically resort to negative gossip as a covert sense-making mechanism and informal social sanction16,17. Although this covert communication helps subordinates navigate uncertainty and bond over shared skepticism in the short term, it ultimately acts as a corrosive force that undermines the organization’s moral foundations18. By unpacking these hidden dynamics, this research elucidates how leadership (in)consistency dictates employee behavioral strategies and ethical climate formation. Based on leadership identity construction theory, this study analyzed 392 employee responses using polynomial regression and response surface analysis to examine how moral person-manager congruence relates to negative workplace gossip about supervisors, with perceived leader hypocrisy as a mediator. The data indicate that higher congruence is associated with lower gossip. Findings advance theoretical understanding of moral leadership’s “dark side” and provide practical insights for cultivating ethical organizational climates. Theoretical framework and hypothesis Moral person and moral manager congruence in relation to employee negative workplace gossip about supervisor Moral leadership comprises two dimensions: the “moral person” reflecting personal moral charisma and the “moral manager” representing managerial competence2. From a moral philosophy perspective, the former stems from moral self-internalization, while the latter functions as a management tool to achieve ethical goals19, with distinct mechanisms and effects. These two dimensions form four combinations along two levels: high person-high manager (Xiuji-Anren), high person-low manager (Xiuji-Unanren), low person-high manager (Unxiuji-Anren), and low person-low manager (Unxiuji-Unanren) (Fig. 1). High-high and low-low configurations represent moral consistency, whereas high-low and low-high reflect moral inconsistency. This study employs polynomial and response surface analyses to examine how these configurations impact employees’ negative gossip toward leaders. Fig. 1Match between moral person and moral manager. Match between moral person and moral manager. High person-high manager (Xiuji-Anren) leaders demonstrate both moral virtues and managerial ethics. Research indicates leaders’ personal traits (sincerity, fairness, trustworthiness) significantly shape followers’ moral behaviors2,20, with high-high moral leadership mitigating unethical follower behavior13. Firstly, leaders’ moral virtues convey non-coercive charisma21, fostering leadership identity and employee compliance. Bedi (2016) found superior moral character predicts employee moral orientations, including perceptions of integrity22, affective commitment23, and cognitive-affective trust24. Secondly, moral leaders shape behavioral standards through value communication and goal alignment, enhancing employee identification and leadership identity formation25. Walumbwa (2010) confirmed moral leadership correlates positively with organizational citizenship behaviors, mediated by employee leadership identity26. This congruent moral leadership fosters positive employee perceptions, promoting organizationally beneficial behaviors while reducing harm. Negative gossip—covert dissemination of unfavorable information in leaders’ absence—reflects implicit evaluation of reputation and norm adherence27. Leaders embodying moral principles and charisma inspire respect, value alignment, and reduced negative gossip. Moral managers communicate ethical values and set clear behavioral standards to embody moral identity, securing employee recognition of effectiveness28, strengthening acceptance of leaders’ moral identity and fostering trust29. Leadership identity construction theory highlights role modeling as critical for identity formation30: moral leaders demonstrate norms while using rewards and punishments to convey values6, fostering high-quality relationships, enhancing leader identification31, and improving moral efficacy. Beyond individual behavior, moral management strengthens leader-employee interactions, boosting evaluations32 and reducing negative behaviors like leader gossip. In summary, high person-high manager moral leaders significantly reduce employees’ negative gossip by minimizing unfavorable evaluations. The second congruent type is low person-low manager (Unxiuji-Unanren) leaders, lacking both moral virtues and stewardship responsibilities. Leadership identity construction theory posits leaders shape identity through behavioral modeling, guidance, and avoidance examples30. When leaders neglect standards like honesty and integrity without promoting such values, employees perceive them as having low moral standards. Under low-person low-manager leadership, employees with low morale may obtain fragmented information about leaders’ negative reputations due to status differences33. Negative gossip serves as a means to diagnose threats, assess negative reputations, and validate information34: employees may engage in negative leader gossip to gather moral leadership information while reinforcing group norms35. Sharing negative information about norm-violating leaders limits unethical behavior and protects group well-being36. Consequently, low-low moral leadership undermines moral leadership identity construction, reduces positive employee evaluations, and increases negative gossip toward leaders. In summary, the following hypotheses were proposed in this study: H1a High congruence between moral person and moral manager will be associated with lower employee negative workplace gossip about supervisor. H1b Employees’ negative workplace gossip about supervisor are lower when leaders’ moral person and moral manager are congruent and high compared to when they are congruent and low. Inconsistent moral leadership manifests as low person-high manager (Unxiuji-Anren) leadership, where leaders enforce ethical standards managerially while failing to embody societal moral norms personally. Leadership Identity Construction Theory explains how leaders construct organizational identities through identity statements10, requiring recognition from others to validate these statements10. Identity conferral solidifies leader roles by aligning personal claims with organizational acknowledgment10. For moral leadership identity, this necessitates congruent moral statements and follower recognition. Behaviors like verbal abuse, idea appropriation, sexual harassment, and bullying signal deficient personal moral leadership37. When leaders fail to articulate clear moral positions while exhibiting unethical conduct, employees perceive value inconsistency and erode trust in leadership standards38. Leaders lacking personal virtues may be deemed immoral despite enforcing ethical norms. Moral double standards—criticizing others’ misconduct to protect reputation while ignoring one’s own flaws39—render leaders vulnerable to hypocrisy perceptions, as condemning others’ immorality while concealing personal failings misrepresents moral integrity40, triggering employee moral indignation41. Moral indignation fosters negative evaluations of leaders’ moral reputation, increasing negative gossip17. Low person-high manager leaders—exhibiting personal moral laxity alongside strict managerial ethics—are perceived as unjust, undermining moral authority and exacerbating negative gossip. When leaders’ personal moral behavior conflicts with their moral manager role, they may adhere to personal moral norms but fail to apply them in organizational practices. This high person-low manager moral leadership (Xiuji-Unanren) embodies strict self-discipline alongside leniency toward others. Such leaders demonstrate exceptional qualities and charisma through moral behavior and grooming, enhancing non-authoritative influence21. Even without imposing strict limits on team misbehavior, this exemplar-based leadership fosters positive leadership image. A core feature of moral leadership—“leading by example”—aligns with social learning theory’s role modeling concept2, which posits individuals learn behaviors through observing others42. As organizational authority symbols, leaders’ moral behaviors exert strong modeling effects: high moral standards stimulate employees’ intrinsic motivation to follow norms without external coercion23. In high person-low manager contexts, employees emulate leaders’ compliance behaviors, conferring moral leadership status. This environment enhances ethical vigilance in decision-making and moral principle adherence43, thereby reducing negative gossip about moral leadership. In summary, the following hypothesis was proposed in this study: H1c Employees’ negative workplace gossip about supervisor are lower when leaders’ moral person is higher than their moral manager as compared to when moral person is lower than moral manager. Perceptions of leader hypocrisy as a mediator of the congruence effect on employee negative workplace gossip about supervisor As mentioned above, leaders who internalize an moral leadership identity exhibit personal traits and behaviors such as sincerity, fairness, and trustworthiness, which earn employees’ recognition of their leadership effectiveness2,20,28. From this perspective, scholars argue that leaders can serve as a potential source of competitive advantage for organizations by cultivating a high moral character44. Specifically, in the management process, leaders may openly communicate organizational moral norms and values to gain a competitive edge; however, their moral character may reveal personal traits such as hypocrisy and unfairness in their interactions with others. This inconsistency can lead to moral leadership behaviors that do not align with the moral personal qualities expected in their management practices45. In this context, research indicates that leaders’ behavioral performance may not consistently reflect their intrinsic qualities and values46. Similarly, moral leadership may involve leaders demonstrating a high level of moral managerial identity to shape their image among followers. Scholars have found that followers often perceive a leader’s display of humility as a positive trait, even if it is not an authentic representation but rather a strategic and deliberate projection47. Consequently, employees may associate negative information with a moral leader and view the leader as a hypocrite, perceiving the leader’s moral managerial behavior as inconsistent with their identity as a moral person. Furthermore, the definition of hypocrisy is shaped by its followers, as it is a socially constructed concept48. Consequently, this study posits that when leaders exhibit inconsistent moral persons and moral managers, employees will sense that this is a strategy used by leaders to gain more leadership and construct their leadership image, and employees will perceive leaders as both individuals and managers, employees will perceive this as a strategy employed by leaders to enhance their authority and cultivate a favorable leadership image, leading them to view these leaders as hypocritical. Conversely, when leaders demonstrate consistent moral behavior as individuals and managers, employees are likely to perceive them as genuine and not hypocritical. Based on these arguments, we propose the following hypothesis: H2a High congruence between moral person and moral manager will be associated with lower employee perceptions of leader hypocrisy. Scholarly discourse has posited that employees’ experiences of hypocrisy often elicit adverse emotional responses, including anger, disappointment, and a diminished sense of trust49. Some empirical investigations suggest that employees’ perceptions of leader hypocrisy do not necessarily lead to significant retaliatory actions, primarily due to employees’ apprehensions regarding potential repercussions for their careers50. In this context, the present study contends that negative gossip concerning leaders constitutes a relatively discreet and evaluative discourse regarding a leader’s unfavorable reputation27. This behavior may serve as a form of retaliation for employees who feel aggrieved by perceived leader hypocrisy. Specifically, it is posited that employees may mitigate their negative sentiments towards their leaders through the act of engaging in negative gossip. Research indicates that individual attitudes significantly influence gossiping behaviors51, with particular emphasis on the notion that when individuals perceive discussions about others’ moral conduct, they are more inclined to participate in gossip52. Furthermore, relevant studies affirm that when employees witness their colleagues breaching established norms or regulations, they are likely to engage in negative gossip as a mechanism of social sanction, thereby serving to alleviate the negative emotions provoked by the hypocritical actions of their leaders53, thus, alleviating the negative emotions triggered by the hypocritical behaviours of the leader. Based on the above arguments, we propose the following hypothesis: H2b Perceptions of leader hypocrisy are positively related to employees’ negative workplace gossip about supervisor. Combining Hypothesis 2a and Hypothesis 2b, this study suggests that when the moral manager of the leader and the moral person are more aligned with the moral leadership dimensions, employees are less likely to perceive hypocrisy in their leaders, and consequently, less likely to exhibit negative gossip behaviours about their leaders. This study posits that the alignment of the moral manager’s characteristics with the moral dimensions of leadership, as outlined in Hypothesis 2a and Hypothesis 2b, results in a decreased perception of hypocrisy among employees regarding their leaders. Consequently, this study suggests that when the moral manager of the leader and the moral person are more aligned with a reduction in negative gossip behaviors directed towards those leaders. H2c Leaders’ moral person and moral manager congruence has an indirect relationship with employee negative workplace gossip about supervisor through employee perceptions of leader hypocrisy. Therefore, for this paper, the research model shown in Fig. 2 was constructed. Fig. 2Theoretical model of this study. Theoretical model of this study. Materials and methods Measurements To assess the variables, a 5-point Likert scale was implemented, where participants were invited to express their level of concurrence with the statements, ranging from 1 (strongly disagree) to 5 (strongly agree). In an effort to guarantee the precision and accuracy of our measurements, we opted for scales that have been previously validated for their efficacy in capturing the intended concepts. For scales that had been utilized in China, we adhered to the existing versions. In cases where they had not been used in China, we engaged in iterative processes of translation and re-translation to achieve a unified agreement on the final version. Moral person. The moral person dimension of moral leadership was measured using an 8-item scale6. A sample item is “My leader reflects on the moral appropriateness of his/her decisions.” The Cronbach’s α for this scale was 0.88. Moral manager. The moral manager dimension was assessed with an 8-item scale6. A sample item is “My leader requires subordinates to learn and understand the codes of ethics.” The Cronbach’s α for this scale was 0.87. Leader hypocrisy. Leader hypocrisy was evaluated using a 4-item measure54. A sample item is “My supervisor tells us to follow the rules but doesn’t follow them himself or herself.” The Cronbach’s α for this scale was 0.82. Negative workplace gossip about supervisor. Negative workplace gossip directed at the supervisor was captured using a 5-item scale17. A sample item is “Told an unflattering story about your supervisor while talking to a work colleague.” The Cronbach’s α for this scale was 0.88. Data collection In this research, we applied a convenience sampling technique, leveraging alumni networks and the school’s corporate partnerships to reach out to a diverse range of companies for survey participation. The data was collected via an online survey conducted on Questionstar, a reputable survey platform in China. The survey was organized into two main parts. The initial part was designed to collect essential demographic details from the participants, including their gender, age, educational background, and the length of time they have been working with their current supervisors. It was clearly communicated to the participants that their information would be strictly for academic research, securely archived on the research team’s server, and accessible exclusively to the team. The second part of the survey included the scales used to assess the proposed theoretical framework. To mitigate potential biases associated with common methodological issues, we gathered data at separate time intervals. Initially, we obtained demographic details along with data on moral leadership, focusing on the aspects of “Moral Person” and “Moral Manager.” Subsequently, after a four-week interval, we acquired information regarding “Leader Hypocrisy” and “Negative Workplace Gossip about Supervisor.” The data sets from these distinct periods were correlated using the last four digits of participants’ phone numbers, and participants had the choice to submit or withhold their responses. In cases where participants opted not to submit, their responses were erased. At the first time point, a total of 523 electronic questionnaires were collected for this survey. At the second time point, a total of 460 electronic questionnaires were collected. After excluding invalid questionnaires with obvious errors or inconsistent responses, 392 valid questionnaires were obtained, with a valid response rate of 85.2%. Descriptive statistical information regarding the sample is provided in Table 1. Table 1Descriptive statistics.ProjectClassificationNumber of PeoplePercentageGenderFemale21655.10%Male17644.90%Age31–409022.96%Under 208922.70%≥ 517218.37%21–307218.37%41–506917.60%Education levelHigh school and below8621.94%Postgraduate8421.43%Junior College8020.41%Doctor or above8020.41%Undergraduate College6215.82%Time with Leaders≤ 1 year8621.94%5–10 years (inclusive)8020.41%Above 10 years7920.15%3–5 years (inclusive)7519.13%1–3 years (inclusive)7218.37%Participation time10–15 years (inclusive)8220.92%5–10 years (inclusive)8220.92%1–5 years (inclusive)7819.90%Above 15 years7719.64%≤ 1 year7318.62% Table 1Descriptive statistics.ProjectClassificationNumber of PeoplePercentageGenderFemale21655.10%Male17644.90%Age31–409022.96%Under 208922.70%≥ 517218.37%21–307218.37%41–506917.60%Education levelHigh school and below8621.94%Postgraduate8421.43%Junior College8020.41%Doctor or above8020.41%Undergraduate College6215.82%Time with Leaders≤ 1 year8621.94%5–10 years (inclusive)8020.41%Above 10 years7920.15%3–5 years (inclusive)7519.13%1–3 years (inclusive)7218.37%Participation time10–15 years (inclusive)8220.92%5–10 years (inclusive)8220.92%1–5 years (inclusive)7819.90%Above 15 years7719.64%≤ 1 year7318.62% Descriptive statistics. Analytic strategy The data analysis followed a two-stage procedure. First, strictly confirmatory factor analysis (CFA) was conducted using R 4.4.1 to verify the measurement model. Second, polynomial regression and response surface analysis were employed to test the hypothesized (in)congruence effects, followed by a block variable approach to evaluate the mediating mechanism. This study employed polynomial regression and response surface analysis to comprehensively investigate the mechanism behind the impact of the mechanisms of moral leadership (In)congruence on employees’ negative gossiping behavior about leaders. Previous work by55 utilized polynomial regression and response surface analysis. In this study, a polynomial regression analysis was performed on the mean-centered independent variables, leading to the development of the following equations (with simplified control variables):1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:M={b}_{0}+{b}_{1}X+{b}_{2}Y+{b}_{3}{X}^{2}+{b}_{4}X\times\:\mathrm{Y}+{b}_{5}{Y}^{2}+e$$\end{document} where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:M$$\end{document} represents “Leader Hypocrisy” or “Negative Workplace Gossip about Supervisor”, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:X$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:Y$$\end{document} represent “Moral person” and “Moral manager”, respectively, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{X}^{2}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{Y}^{2}$$\end{document} represent the square of “Moral person” and “Moral manager”, respectively, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:X \times\:Y$$\end{document} represents the product term of “Moral person” and “Moral manager”. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{0}$$\end{document} to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{5}$$\end{document} are unstandardized coefficients, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{0}$$\end{document} is a constant term, and the rest are regression coefficients. The rest are the regression coefficients of each item respectively, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:e$$\end{document} is the error term. When \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:X=Y$$\end{document}, Eq. (1) is deformed into (2):2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:M={b}_{0}+({b}_{1}+{b}_{2})X+({b}_{3}+{b}_{4}+{b}_{5}){X}^{2}+e$$\end{document} When \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:X=-Y$$\end{document}, Eq. (1) is deformed into (3):3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:M={b}_{0}+({b}_{1}-{b}_{2})X+({b}_{3}-{b}_{4}+{b}_{5}){X}^{2}+e$$\end{document} In Eq. (2): the slope of the change of M on the X = Y cross section is denoted as a1 (a1 = b1+b2)); the curvature is denoted as a2 (a2 = b3+b4 + b5). Similarly, in Eq. (3), the slope and curvature of the change of M in the X = − Y cross section are denoted as a3 (a3 = b1−b2) and a4 (a4 = b3−b4 + b5), respectively. The relationship between the four focal variables—the two predictors, moral person (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}$$\end{document}) and moral manager (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{Y}$$\end{document}), and the outcome variable, NWGS (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{Z}$$\end{document})—was analyzed using the five polynomial terms\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{b}}_{\mathrm{1}}\mathrm{to}{\mathrm{b}}_{\mathrm{5}}\mathrm{)}\text{}$$\end{document}described in Eq. (1) and visualized in a three-dimensional space spanning the response surface55. In the resulting graph (see Fig. 3), moral person and moral manager are displayed on perpendicular horizontal axes (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{Y}$$\end{document}), while NWGS is represented on the vertical axis (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{Z}$$\end{document})55. The response surface is characterized by two reference lines: the line of congruence (LOC), where moral person and moral manager identities are aligned \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}{=}\mathrm{Y}$$\end{document}, and the line of incongruence (LOIC), where they differ \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}{=-}\mathrm{Y}$$\end{document}55. Fig. 3Congruent and incongruent effects of asymmetry of moral person-moral managers on employee gossip behavior regarding leadership. Congruent and incongruent effects of asymmetry of moral person-moral managers on employee gossip behavior regarding leadership. Following established frameworks for assessing congruence effects, we evaluated several necessary conditions55. First, the ridge of the response surface should align with the line of congruence without significant rotation, which is verified by ensuring the first principal axis has an intercept of 0 and a slope of 1 (assessed via the significance of the fifth coefficient of the response surface,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{5}}\mathrm{)}$$\end{document}55. Second, for negative workplace gossip about supervisor, the surface along the line of incongruence should be curved upward—indicated by a significant positive\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{4}}\mathrm{)}$$\end{document} coefficient—suggesting that the dependent variable increases when a leader’s moral person and moral manager dimensions differ from each other in either direction55. Third, the slope of the line of incongruence at the origin\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{3}}\mathrm{)}$$\end{document} was examined to determine if the outcome is minimized at the point of perfect congruence55. To further explore the specific nature of the congruence effect, we examined the surface above the line of congruence by testing its slope\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{1}}\mathrm{)}$$\end{document} and curvature\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{2}}\mathrm{)}$$\end{document}55. A significantly negative\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{1}}\mathrm{)}$$\end{document} slope would indicate that negative workplace gossip about supervisor is lower when moral person and moral manager are congruently high compared to when they are congruently low55. Finally, the stability of the surface was verified by ensuring the first principal axis did not deviate significantly from the line of congruence, thereby confirming that the predefined parameters for a strict congruence effect were maintained55. The analysis strategies for each hypothesis are as follows: To test H1a and H2a (the congruence effect), we first examined whether the coefficients of the three second-order polynomial terms (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{X}^{2}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{Y}^{2}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:X\times\:Y$$\end{document}) were jointly significant. We also investigated whether the curvature along the inconsistency line \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:(X=-Y)$$\end{document} was significantly positive, suggesting a U-shaped curve in this region56. Additionally, we verified whether the fifth coefficient of the response surface (a₅) is significantly different from zero57. Furthermore, we assessed whether the first principal axis aligns with the concordance line, thereby providing additional support for H1a and H2a. This H1a and H2a posits that when “Moral person” and “Moral manager” are congruent, perceptions of leader hypocrisy and negative employee gossip about the leader is minimized. Therefore, support for H1a depends on both analyses of the response surface. In contrast, H1b was tested by examining the slope\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{(}{\mathrm{a}}_{\mathrm{1}}\mathrm{)}$$\end{document}Of the consistency line \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}\mathrm{=}\mathrm{Y}$$\end{document}, which determines whether “high-high” congruence yields more favorable outcomes than “low-low” congruence58. suggesting that when “Moral person” and “Moral manager” reached a high rather than a low level of employees’ negative gossip about their leaders decreases. To test H1c (asymmetric incongruence), we tested the surface by examining the lateral shift of the surface along the inconsistency (X = - Y) line -[\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{1}$$\end{document}- \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{2}$$\end{document}]/[2×(\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{3}$$\end{document}-\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{4}$$\end{document}+\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{5}$$\end{document})] to be significantly positive59. In addition, we combined the test H1c by testing that the slope of the inconsistency line (a3) is significantly negative58. suggesting that employees’ negative workplace gossip about supervisors are lower when leaders’ moral person is higher than their moral manager as compared to when moral person is lower than moral manager. To test H2b, we conducted a simple linear regression to examine the relationship between perceived leader hypocrisy and employees’ negative workplace gossip about supervisor. To test H2c, we constructed a “block variable” based on the weights in a polynomial regression model of “Perceived leader-hypocrisy” based on the matching of “Moral person” and “Moral manager”56. We constructed a “block variable” based on the weights in a polynomial regression model of “hypocrisy”56. The mediation test was then taken to MODEL 4 of the PROCESS plugin60. Results Confirmatory factor analyses In this study, the discriminant validity of the model consisting of four variables (a total of 25 question items) was tested using R 4.4.1. To verify the structural integrity of our measurement model, we performed a purely confirmatory factor analysis (CFA) using R 4.4.1. This analysis was strictly theory-driven, based on the a priori conceptualization of moral person, moral manager, leader hypocrisy, and negative workplace gossip. As summarized in Table 2, the four-factor model yielded a satisfactory fit to the data (χ2/df=2.17 < 3, RMSEA =0.055 < 0.06, CFI = 0.929 >0.9, TLI = 0.92 >0.9, SRMR = 0.049 < 0.06). In contrast, the three-factor model—in which the two moral leadership dimensions (moral person and moral manager) were combined—showed a significantly poorer fit (χ2/df = 4.88, Δχ2(3)=743.3, p <.001). Similarly, both the two-factor model (χ2/df = =7.75, Δχ2(5)=1539.3, p <.001) and the single-factor model (χ2/df = 9.65, Δχ2(6)=2069, p <.001)exhibited inadequate fit indices. These comparison results demonstrate that the four constructs are empirically distinct and that the measurement model is robust against common method bias. Table 2Confirmatory factor analysis results.ModelFactorchisqdfΧ2/dfCFITLISRMRRMSEAΔχ2/ΔdfFour-factor modelX, Y, M, Z 584.248 269 2.170.9290.920.0490.055-Three-factor modelX + Y, M, Z 1327.552 272 4.880.7610.7370.090.099743.3(3)***Two-factor modelX + Y+M, Z 2123.562 274 7.750.5820.5420.130.1311539.3(5)***Single-factor model X + Y+M + Z 2653.256 275 9.650.4620.4130.1490.1492069(6)***Note: + denotes the factorization of the merger. X= Moral Person, Y= Moral Manager, M= Leader Hypocrisy, Z= Negative Workplace Gossip about Supervisor. The same as below. Table 2Confirmatory factor analysis results.ModelFactorchisqdfΧ2/dfCFITLISRMRRMSEAΔχ2/ΔdfFour-factor modelX, Y, M, Z 584.248 269 2.170.9290.920.0490.055-Three-factor modelX + Y, M, Z 1327.552 272 4.880.7610.7370.090.099743.3(3)***Two-factor modelX + Y+M, Z 2123.562 274 7.750.5820.5420.130.1311539.3(5)***Single-factor model X + Y+M + Z 2653.256 275 9.650.4620.4130.1490.1492069(6)***Note: + denotes the factorization of the merger. X= Moral Person, Y= Moral Manager, M= Leader Hypocrisy, Z= Negative Workplace Gossip about Supervisor. The same as below. Confirmatory factor analysis results. Note: + denotes the factorization of the merger. X= Moral Person, Y= Moral Manager, M= Leader Hypocrisy, Z= Negative Workplace Gossip about Supervisor. The same as below. Common method variance Considering the cross-sectional nature of the data and their collection from a single source, the potential for common method bias (CMB) was examined following the recommendations of Williams and McGonagle (2016) through the application of two complementary statistical procedures61. First, Harman’s single-factor test was conducted using exploratory factor analysis (EFA). The results indicated that the first unrotated principal component accounted for 24.219% of the total variance. This value is well below the conservative 40% threshold62, suggesting that no single latent factor dominates the variance among the constructs. Second, we employed the unmeasured latent method factor technique to further evaluate the robustness of our measurement model. As shown in Table 2, a latent method factor was integrated into the baseline CFA model, with all observed items loading on both their respective theoretical constructs and the method factor. Although the chi-square difference test indicated a significant improvement in modelfit (Δχ2=222.54, Δdf=29, p<0.001), the incremental changes in the fit indices were minimal. Specifically, the CFI and TLI increased by 0.043 and 0.046, respectively, while the RMSEA and SRMR decreased by 0.019 and 0.012. Because these shifts in fit indices (ΔCFI, ΔTLI<0.1; ΔRMSEA, ΔSRMR < 0.05) remain within widely accepted conservative ranges61, we conclude that CMB does not substantively distort the structural relationships in this study. Furthermore, as our core analysis involves polynomial regression and response surface analysis, the complexity of these non-linear and interaction effects provides an inherent buffer against CMB. Methodological research suggests that CMB typically attenuates rather than inflates non-linear and interaction terms63; therefore, the significant congruence/incongruence effects reported in our study are unlikely to be artifacts of common method variance. Descriptive statistical and correlation analysis We summarized study variable averages, standard deviations, and correlations. Moral person positively correlated with both Moral manager (r = 0.462), leader hypocrisy (r = 0.219) and negative workplace gossip about supervisor (r = 0.321). Leader hypocrisy was positively correlated with negative workplace gossip about supervisor (r = 0.415) (see Table 3). The model aligns with our hypotheses and has low collinearity according to the variance inflation factor (VIF < 5), ensuring data reliability. Table 3Means, standard deviations and correlations.12345XYZMMeanSD1.Gender-1.550.502.Age0.03-2.911.413.Education level-0.020.03-2.981.464.Time with Leaders0.050.040.09-2.981.445.Participation time0.020.06-0.050.03-3.031.39X-0.070.030.000.010.02-3.140.98Y-0.060.000.030.03-0.030.462***-3.300.93Z0.05-0.105*0.060.03-0.02-0.321***-0.08-2.820.82M-0.02-0.010.070.010.00-0.219***-0.080.415***-2.970.99*p < 0.05, **p < 0.01, *** p < 0.001. Table 3Means, standard deviations and correlations.12345XYZMMeanSD1.Gender-1.550.502.Age0.03-2.911.413.Education level-0.020.03-2.981.464.Time with Leaders0.050.040.09-2.981.445.Participation time0.020.06-0.050.03-3.031.39X-0.070.030.000.010.02-3.140.98Y-0.060.000.030.03-0.030.462***-3.300.93Z0.05-0.105*0.060.03-0.02-0.321***-0.08-2.820.82M-0.02-0.010.070.010.00-0.219***-0.080.415***-2.970.99*p < 0.05, **p < 0.01, *** p < 0.001. Means, standard deviations and correlations. *p < 0.05, **p < 0.01, *** p < 0.001. Hypothesis testing We performed polynomial regression analyses to evaluate the proposed hypotheses, with results summarized in Table 4. As indicated, Model 1 serves as the baseline model for negative workplace gossip about supervisor (NWGS). Model 3 includes the quadratic and interaction terms of moral person (MP) and moral manager (MM), resulting in a significant increase in explanatory power as R2 rose from 0.018 to 0.167. Similarly, Model 4 serves as the baseline model for leader hypocrisy (LH). Model 5 evaluates these polynomial terms in relation to leader hypocrisy (LH), with the R2 increasing from 0.005 to 0.077. These improvements in the goodness-of-fit indices suggest that the polynomial regression framework provides a more nuanced understanding of the relationships between moral leadership dimensions and their outcomes than linear models (Table 5). Table 4Polynomial regressions of leader hypocrisy and negative workplace gossip about supervisor on moral leadership (in)congruence.Model1Model2Model3Model4Model52.747***1.761***2.709***2.883***2.833***1.Gender0.0820.0960.062-0.040-0.0652.Age-0.063*-0.060-0.049-0.0080.0013.Education level0.0340.0180.0310.0450.0404.Time with Leaders0.0170.0140.0210.0070.0135.Participation time-0.007-0.009-0.0070.0050.003M0.342***Polynomial coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{1}X$$\end{document} -0.239***-0.195** \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{2}\:Y$$\end{document} 0.016-0.022 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{3}{X}^{2}$$\end{document} 0.00020.028\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{4}(\mathrm{X}\times\:\mathrm{Y}$$\end{document})-0.154***-0.117* \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{5}{Y}^{2}$$\end{document} 0.1280.128* Congruence (X = Y) line) Slope (a1 = b1 + b2) -0.223***-0.217*** Curvature (a2 = b3 + b4 + b5) -0.0260.039 Incongruence (X = − Y) line Slope (a3 = b1−b2) -0.255**-0.173 Curvature (a4 = b3 – b4 + b5) 0.283***0.273** a5 = b3-b5 -0.128-0.100 Lateral shift 0.450*0.316χ20.0180.187***0.167***0.0050.077** F for the 3 quadratic terms 6.577***3.099**p < 0.05, ** p < 0.01, ***p<0.001. Unstandardized coefficients are shown in the table. Table 4Polynomial regressions of leader hypocrisy and negative workplace gossip about supervisor on moral leadership (in)congruence.Model1Model2Model3Model4Model52.747***1.761***2.709***2.883***2.833***1.Gender0.0820.0960.062-0.040-0.0652.Age-0.063*-0.060-0.049-0.0080.0013.Education level0.0340.0180.0310.0450.0404.Time with Leaders0.0170.0140.0210.0070.0135.Participation time-0.007-0.009-0.0070.0050.003M0.342***Polynomial coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{1}X$$\end{document} -0.239***-0.195** \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{2}\:Y$$\end{document} 0.016-0.022 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{3}{X}^{2}$$\end{document} 0.00020.028\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{4}(\mathrm{X}\times\:\mathrm{Y}$$\end{document})-0.154***-0.117* \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{b}_{5}{Y}^{2}$$\end{document} 0.1280.128* Congruence (X = Y) line) Slope (a1 = b1 + b2) -0.223***-0.217*** Curvature (a2 = b3 + b4 + b5) -0.0260.039 Incongruence (X = − Y) line Slope (a3 = b1−b2) -0.255**-0.173 Curvature (a4 = b3 – b4 + b5) 0.283***0.273** a5 = b3-b5 -0.128-0.100 Lateral shift 0.450*0.316χ20.0180.187***0.167***0.0050.077** F for the 3 quadratic terms 6.577***3.099**p < 0.05, ** p < 0.01, ***p<0.001. Unstandardized coefficients are shown in the table. Polynomial regressions of leader hypocrisy and negative workplace gossip about supervisor on moral leadership (in)congruence. *p < 0.05, ** p < 0.01, ***p<0.001. Unstandardized coefficients are shown in the table. Table 5Results of mediating effect test.Leader hypocrisyNegative workplace gossip about supervisorBlock variable1.0026***0.8967***Leader hypocrisy0.2790***Indirect effect0.2797 [0.1487, 0.4375]***p<0.001. Unstandardized coefficients are shown in the table. 95% confidence intervals in [ ]. Table 5Results of mediating effect test.Leader hypocrisyNegative workplace gossip about supervisorBlock variable1.0026***0.8967***Leader hypocrisy0.2790***Indirect effect0.2797 [0.1487, 0.4375]***p<0.001. Unstandardized coefficients are shown in the table. 95% confidence intervals in [ ]. Results of mediating effect test. ***p<0.001. Unstandardized coefficients are shown in the table. 95% confidence intervals in [ ]. Congruence and incongruence effects on NWGS (H1a–H1c) Regarding H1a, the three second-order polynomial terms in Model 3 were jointly significant (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{=6.577,}\text{}\mathrm{p}\mathrm{<0.001}$$\end{document}). Specifically, the curvature along the incongruence line (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}\mathrm{=-}\mathrm{Y}$$\end{document}) was significantly positive (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{4}}\mathrm{=0.283}$$\end{document}, p < 0.001 ), while the curvature along the congruence line (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}\mathrm{=}\mathrm{Y}$$\end{document}) was non-significant (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{2}}\mathrm{=-0.026}$$\end{document}, ns). This indicates a U-shaped response surface along the incongruence line, suggesting that as the discrepancy between MP and MM increases, the associated levels of NWGS rise significantly. Furthermore, the non-significant fifth coefficient (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{5}}\mathrm{=-0.128}$$\end{document}, ns) confirms that the response surface is properly centered without significant lateral displacement relative to the major axes. These findings, visualized in the response surface and the incongruence cross-section in Fig. 5. Figure 5 demonstrates that any perceived discrepancy between a leader’s MP and MM roles triggers a marked rise in NWGS. As the misalignment intensifies—evidenced by the movement toward the extremities of the X-axis—the frequency of NWGS escalates. Conversely, the results indicate that the more congruent these two moral dimensions are, the lower the prevalence of NWGS. support the congruence effect proposed in H1a. To test H1b, we examined the slope of the consistency line ((\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}\mathrm{=}\mathrm{Y}$$\end{document}). The slope (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{1}}\mathrm{=-0.223}$$\end{document}, p < 0.001) was significantly negative. As illustrated in the Line of Consistency (LOC) cross-section (Fig. 4), NWGS is notably higher when MP and MM are congruently low compared to when they are congruently high. This demonstrates that high-high moral leadership congruence is more effective in minimizing gossip than low-low congruence, supporting H1b. Fig. 4Moral person - moral manager concurrence. Moral person - moral manager concurrence. For H1c, the lateral shift along the inconsistency line (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{X}\mathrm{=-}\mathrm{Y}$$\end{document}) was significantly positive (lateral shift = 0.450, p < 0.05), and the slope was significantly negative (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{3}}\mathrm{=-0.255}$$\end{document}, p < 0.01). As shown in Fig. 5, NWGS reaches its maximum at the “low MP-high MM” endpoint (left side of the curve). In contrast, gossip levels at the “high MP-low MM” endpoint (right side of the curve) remain significantly lower. This asymmetric effect supports H1c, suggesting that employees are more tolerant of leaders who exhibit high personal virtue but lower managerial enforcement than vice versa. Fig. 5Moral person - moral manager (in)concurrence. Moral person - moral manager (in)concurrence. The mediating role of leader hypocrisy (H2a–H2c) In testing H2a, results from Model 5 showed that the three second-order terms were jointly significant (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{=3.099}$$\end{document}, p < 0.05). Similar to the gossip model, the curvature along the incongruence line was significantly positive (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{4}}\mathrm{=0.273}$$\end{document}, p < 0.01), while the curvature along the congruence line was flat (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathrm{a}}_{\mathrm{2}}\mathrm{=0.039}$$\end{document}, ns). This indicates a U-shaped response surface along the incongruence line, suggesting that as the discrepancy between MP and MM increases, the associated levels of LH rise significantly. Furthermore, the non-significant fifth coefficient (a5=-0.100, ns), confirms that the response surface is properly centered without significant lateral displacement relative to the major axes. As illustrated in Fig. 6. Figure 6 demonstrates that any perceived discrepancy between a leader’s MP and MM roles triggers a marked rise in LH. As the misalignment intensifies—evidenced by the movement toward the extremities of the X-axis—the frequency of LH escalates. Conversely, the results indicate that the more congruent these two moral dimensions are, the lower the prevalence of LH. support the congruence effect proposed in H2a. Fig. 6Moral person - moral manager (in)concurrence. Moral person - moral manager (in)concurrence. Regarding H2b, Model 2 confirms that perceived leader hypocrisy is positively associated with NWGS (β = 0.342, p < 0.001). Finally, to evaluate H2c, we employed a block variable approach to test the mediation effect. The indirect effect of moral leadership (in)congruence on NWGS through leader hypocrisy was 0.2797, with a 95% Bootstrap CI [0.1487, 0.4375]. Since this interval does not include zero, the mediating role of perceived hypocrisy is statistically significant, indicating a partial mediation effect and providing full support for H2c. Discussion and conclusions This research addresses a fundamental gap in value-based leadership literature by investigating how the congruence between a leader’s moral person and moral manager dimensions influences perceived leader hypocrisy and negative workplace gossip about supervisor. While prior studies have frequently conflated these two dimensions by treating moral leadership as a unidimensional construct2,3, in this sample disentangles their interactive effects. Utilizing polynomial regression and response surface analysis on data from 392 employees, we integrate leadership identity construction theory with moral psychology to propose an “alignment-based paradigm” for evaluating leadership consistency. Our empirical analysis yields three salient patterns. First, a high degree of congruence between moral person and moral manager attributes is associated with reduced negative supervisor gossip. This effect is driven by the structural alignment of personal virtues and managerial behaviors rather than absolute moral intensity. Second, while leaders exhibiting high levels of both dimensions elicit the lowest frequency of gossip, congruently low moral profiles still generate less gossip than incongruent configurations. This underscores that moral consistency intrinsically shapes employee reactions. Third, we identified an asymmetric incongruence effect: when a leader’s moral person behaviors exceed their moral manager behaviors, gossip levels remain comparatively lower than in the inverse scenario. This indicates a greater employee tolerance for unfulfilled moral promises than for strict ethical enforcement devoid of personal integrity. Theoretical implications This study advances moral leadership research beyond incremental contributions through three primary theoretical avenues: First, we introduce a congruence framework that pivots moral leadership research toward an “alignment-based paradigm.” While existing literature predominantly highlights the positive effects of moral leadership on employee attitudes64, it largely overlooks the critical alignment between a leader’s internal moral persona and external managerial roles. Furthermore, prior studies note an anomaly where moral leadership may inadvertently induce employee unethical behavior under certain conditions65,66. Our congruence framework resolves this paradox by demonstrating that such unintended negative outcomes often stem from dimensional imbalance rather than absolute moral intensity. We establish that employees evaluate leaders as unified moral entities, and leadership effectiveness is strictly contingent upon this dual-dimension congruence. Second, this study advances organizational congruence research by delineating the interactive dynamics of moral leadership’s dual pillars. Traditional congruence research posits that matched states generally yield superior outcomes67, yet it rarely differentiates between specific congruent and incongruent configurations. By integrating leadership identity construction theory, we test classical congruence effects alongside the “rising ridge” effect (verifying that high-high congruence yields more favorable evaluations than low-low congruence) and propose an asymmetric incongruence effect. This framework extends classical theory by elucidating precisely why specific misalignments (e.g., high moral person–low moral manager versus low moral person–high moral manager) trigger varying magnitudes of negative social evaluations. Third, we establish perceived leader hypocrisy as the critical mediating mechanism linking leadership inconsistency to harmful workplace behaviors. Prior research indicates that positive leader behaviors, such as humility, may paradoxically enhance hypocrisy perceptions68, reflecting negative cognitive responses to ostensibly positive leadership styles. Our study extends this logic by theorizing that perceived leader hypocrisy is fundamentally a symptom of identity fragmentation. We empirically demonstrate that inconsistencies between a leader’s professed moral person and enacted moral manager behaviors fuel these hypocrisy perceptions, creating a structural “credibility gap” that subsequently amplifies negative gossip. Practical implications Our findings offer actionable insights for managing ethical climates and interpersonal dynamics within organizations: First, organizations must leverage the dual dimensions of moral leadership to mitigate negative employee cognitions. Rather than focusing exclusively on character-building, leadership development programs should emphasize internal-external alignment. This entails integrating clear ethical standards, monitoring systems, and timely responses to violations directly with the cultivation of personal integrity. Such holistic strategies foster a positive organizational culture, ensuring adherence to ethical norms while minimizing the risk of leaders being perceived as hypocritical. Second, managing workplace gossip presents unique challenges compared to standard workplace deviance. Organizational research continuously aims to mitigate harmful interpersonal outcomes69. Although participating in negative gossip targeting supervisors may temporarily empower perpetrators, prior studies demonstrate that it ultimately depletes employees’ time, attention, and energy, thereby reducing overall productivity14. Consequently, when negative gossip emerges, managers must maintain composure, redirect discourse toward constructive channels, and systematically curtail the underlying structural inconsistencies instigating such behavior. Because subordinates’ behavioral norms are profoundly influenced by leadership consistency, managers must embody exemplary, congruent conduct to prevent the dissemination of detrimental rumors. Limitations This study presents several limitations that offer valuable avenues for future research. First, regarding the methodological design, although we employed a two-stage matched data collection procedure to mitigate potential biases, the data remain cross-sectional and single-source in nature. Because cross-sectional data capture variables at a specific point in time, this design inherently precludes definitive causal inferences; therefore, our findings can only demonstrate associations rather than causal relationships. Furthermore, while the inherent complexity of polynomial regression and our rigorous statistical checks provide some confidence, we acknowledge that CMB is only partially addressed due to our reliance on self-reported data. To establish robust causal directionality and further minimize self-report biases, future research should adopt longitudinal methodologies, controlled experimental designs, and multi-source data collection strategies. Second, our sample was drawn exclusively from Chinese organizations, which may limit the generalizability of our findings. The Chinese cultural context is characterized by specific values, such as high power distance, a strong emphasis on interpersonal harmony, and the concept of ‘face’ (mianzi). These cultural nuances might influence how employees perceive and react to a leader’s moral incongruence and hypocrisy. Consequently, future research should cross-validate our theoretical model in Western or other diverse cultural settings to determine whether these asymmetric congruence effects hold across different contexts. Third, from a theoretical standpoint, the current model does not explore the boundary conditions that may delineate the relationship between moral leadership congruence and leader-directed employee gossip. For instance, a leader’s structural position within an organizational network may significantly influence gossip dynamics. Extant literature suggests that leaders occupying central network positions garner heightened visibility and employee attention, which may subsequently amplify gossip behaviors70. Consequently, future research would benefit from incorporating leader network centrality as a moderating variable to elucidate the contextual boundaries and boundary effects of this congruence phenomenon.