Does Support in Organizations Inhibit Power Harassment? An Analysis Based on Self-Esteem and Types of Narcissism

Abstract

In contemporary Japanese workplaces, interpersonal relationship problems have become increasingly serious, leading to heightened psychological stress and declining organizational functioning. One major contributing factor is power harassment (workplace bullying). This study surveyed 1621 Japanese workers to examine how support from supervisors and organizations influences power harassment, with particular attention to differences in self-esteem levels and narcissistic types. Hierarchical multiple regression analyses revealed that among individuals with high self-esteem, supervisor support tended to reduce power harassment in those characterized by the Need for Attention and Praise type, whereas organizational support tended to increase it. Additionally, for those classified as the Sense of Superiority and Competence type, the interaction between ego threat and both types of support showed a tendency to exacerbate power harassment. For individuals with low self-esteem, the interaction between ego threat and both types of support similarly tended to intensify power harassment in the Need for Attention and Praise type. These results suggest that the effects of support are not uniform; rather, they may inhibit or facilitate power harassment depending on individual psychological traits. Therefore, tailoring the method, timing, and source of support to workers’ psychological characteristics is essential for both preventing power harassment and promoting psychological adaptation.

1. Introduction

1.1. Background

In contemporary workplaces, interpersonal relationship issues have become increasingly serious, leading to elevated psychological stress among workers and the deterioration of organizational functioning. According to the Ministry of Health, Labour and Welfare (2023) 82.2% of workers in Japan reported experiencing strong anxiety, stress, or distress related to work or occupational life, with “interpersonal relationships (including sexual and power harassment)” cited as a major cause by 26.2% of respondents. Furthermore, the Ministry of Health, Labour and Welfare (2025) reported that among 267,755 consultations handled by Japanese Labour Standards Inspection Offices nationwide in FY 2024, cases related to “bullying and harassment” were the most frequent, totaling 54,987, indicating an increase in workplace power harassment.

Power harassment, a concept originating in Japan, is defined as “behavior in the workplace based on a superior relationship that goes beyond what is necessary and appropriate for business and harms the work environment of a worker” (Ministry of Health, Labour and Welfare 2012). Some forms of power harassment overlap with organizational dysfunctional behavior, a construct examined in Western research on antisocial behavior in organizations (Semba and Haraguchi 2014).

In Japan’s aging and shrinking workforce, it is increasingly essential to create workplaces where each individual can fully utilize their abilities. To this end, establishing support systems that prevent workers’ self-evaluations from being threatened—or that prevent such threats from leading to aggressive behavior such as power harassment—is crucial. To develop effective countermeasures against power harassment, we must first identify its determinants and construct a theoretical model to explain them. However, to the best of the author’s knowledge, no previous studies have comprehensively examined both the factors and the theoretical framework underlying the occurrence of power harassment. In contrast, numerous studies have accumulated findings on organizational dysfunctional behavior, a concept that encompasses some aspects of power harassment.

Among the determinants of organizational dysfunctional behavior, narcissism has been identified as one of the most powerful predictors (Tanaka 2008). To explain the relationship between narcissism and aggressive behavior, the Threatened Egotism Model of Baumeister et al. (1996) is widely cited. According to this model, individuals with threatened self-evaluations experience discrepancies between their self-concept and negative evaluations from others, perceiving these as ego threats. To prevent a decline in self-esteem, such individuals tend to engage in coping strategies aimed at compensating for or maintaining their self-evaluation (Tabata and Ikegami 2011), which may result in behaviors such as social withdrawal or aggression.

Oshio (1998) examined the relationship between narcissistic tendencies and self-esteem and found that the three subscales of the Narcissistic Personality Inventory (Raskin and Hall 1979) (hereafter NPI) exerted distinct influences on the subscales of a self-esteem measure. This suggests that the three NPI factors may play different functional roles in maintaining a positive self-evaluation. Building on this, Semba (2016a, 2018) investigated how these factors of narcissism influence coping behaviors and behavioral choices within organizations. Semba (2016b) incorporated both organizational support and narcissistic types—defined by the three NPI subscales—into the Threatened Egotism Model and analyzed the process by which perceived ego threat leads to organizational dysfunctional behavior. The findings indicated that supervisor support has a mitigating effect on such behavior, and that perceptions and seeking of support differ depending on narcissistic type. Furthermore, Semba (2018) quantitatively examined the effects of support under ego-threatening conditions, revealing that the type of organizational support and narcissistic type influence behavioral choices following threat perception.

Taken together, these findings suggest that power harassment—a concept adjacent to organizational dysfunctional behavior—can be understood as a form of defensive behavior triggered by perceived ego threat, and that the effectiveness of support can be discussed in this context. In particular, examining how supervisor and organizational support influence the process from ego threat to aggressive behavior (i.e., power harassment) is both theoretically and practically significant.

Therefore, this study, based on the Threatened Egotism Model, employs a quantitative approach to examine the effects of ego threat on power harassment, stratified by psychological factors, including narcissistic type. Furthermore, it divides support in organizations into supervisor support—defined as social support generally provided from supervisors to subordinates in Japanese management organizations—and organizational support, a comprehensive form of assistance that acts on all employees through systems, norms, and culture. Subsequently, we aim to analyze the effects of these types of support on power harassment, as well as their moderating effects on the relationship between ego threat and power harassment.

1.2. The Relationship Between Self-Esteem, Power Harassment, and Support in Organizations

Self-esteem, defined as an overall positive or negative attitude toward oneself (Rosenberg 1965), has been widely recognized as a core psychological factor influencing emotional stability and interpersonal behavior. Prior research has shown that even individuals with high self-esteem tend to express anger and hostility in response to negative information when their self-esteem is unstable (Kernis et al. 1989). In Japan, Shimizu (1999) found that individuals with consistently low self-esteem reported more experiences of physical violence and higher levels of hostility and irritability. Similarly, Kawamura and Toda (2014) demonstrated that low self-esteem was associated with strong hostile cognition. These findings suggest that both the level and stability of self-esteem are closely related to aggressive cognition and behavior, providing a useful framework for understanding workplace power harassment in Japan. Overall, high self-esteem tends to be linked with sensitivity to negative evaluations and tendencies toward anger and aggression, whereas low self-esteem is associated with self-negativity and tendencies toward resentment and hostility. Although the mechanisms differ, both conditions make individuals vulnerable to ego threats, which can trigger aggressive or defensive responses.

Furthermore, recent empirical studies have also examined the relationship between self-esteem and aggressive behavior. For example, Hu et al. (2023), using longitudinal data from Chinese adolescents, demonstrated that individuals with lower self-esteem tend to exhibit higher levels of aggressive behavior, suggesting a negative association between self-esteem and aggression. Furthermore, a large-scale meta-analytic study identified state self-evaluation as a key psychological mechanism through which workplace aggression influences employee behavior (Zhong et al. 2025). This finding indicates that temporary threats to self-evaluation mediate deviant behavior and aggressive responses, thereby supporting the importance of ego threat and self-esteem processes that constitute the focus of the present study. Taken together, these findings indicate that understanding how self-esteem moderates the influence of ego threat on behavior is crucial for clarifying the mechanisms underlying workplace power harassment.

Moreover, perceived support from supervisors and organizations has been found to reduce aggressive and deviant behaviors (Khan et al. 2015; Laurenza and Kristanto 2020). However, this effect is not uniform: individuals with low self-esteem may interpret support as a signal of their weakness or incompetence, potentially increasing defensive responses or aggression.

Note that although prior research has highlighted the role of unstable high self-esteem, the present study did not directly measure the stability of self-esteem. Therefore, this study focuses on the observable psychological state of self-esteem level. Accordingly, references to instability in the present study are positioned as part of the theoretical background, and the analyses and interpretations are conducted based on differences in the level of self-esteem.

1.3. The Relationship Between Narcissism, Self-Esteem, Power Harassment, and Support in Organizations

Narcissism is a psychological function aimed at maintaining a positive self-evaluation, defined as the function “to maintain the cohesiveness, stability, and positive affective coloring of the self-representation” (Stolorow 1975). It consists of three factors: Sense of Superiority and Competence, reflecting self-affirmation of being superior to others; Need for Attention and Praise, representing the desire to be noticed and praised; and Self-Assertion, denoting the tendency to assert one’s opinions (Oshio 2004, 2006). The relative strength of these factors characterizes individual behavior. Individuals high in Sense of Superiority and Competence are classified as the Sense of Superiority and Competence type, and tend to justify unilateral aggressive expressions toward others, such as causing discomfort or expressing sarcasm (Abe and Takagi 2006a). Those high in Need for Attention and Praise are classified as the Need for Attention and Praise type, expressing passive, dependent interpersonal tendencies (Oshio 2004); they are prone to experience anger yet often suppress its expression, perceiving it as inappropriate (Abe and Takagi 2006a, 2006b). Those high in Self-Assertion are classified as the Self-Assertion type, and actively project an exaggerated self-image, counteracting self-image instability (Kawasaki and Kodama 2007). They tend to be active, assertive, and self-reliant in interpersonal relationships and are less concerned with harmonious relationships (Oshio 2004, 2006). Empirical studies have shown that the strength of these factors significantly influences behavior after the perception of ego threat (Semba 2018).

Recent studies have further distinguished narcissism into vulnerable narcissism and grandiose narcissism, reporting that vulnerable narcissism is associated with hypersensitive and defensive reactions under ego threat as well as reactive aggression, whereas grandiose narcissism is linked to assertive and dominance-oriented tendencies and, depending on the situation, to aggressive behavior (Du et al. 2024). In the present study, narcissism was analyzed in terms of three dimensions: Self-Assertion, Need for Attention and Praise, and Sense of Superiority and Competence. The Need for Attention and Praise may involve evaluation anxiety and hypersensitivity to others’ judgments, and thus can be treated as reflecting aspects of vulnerable narcissism. At the same time, when it is grounded in high self-esteem or grandiosity, it may also encompass features of grandiose narcissism. Therefore, the Need for Attention and Praise can be positioned as a dual-faceted dimension that contains characteristics of both types. The other dimensions—Self-Assertion and Sense of Superiority and Competence—are considered to correspond to the prototypical features of grandiose narcissism recognized in the international literature. This conceptual clarification helps situate the narcissistic dimensions examined in this study within the broader context of global narcissism research.

In the present study, to clarify narcissistic typologies, only participants who showed the highest score on a single one of the three subscales were included in the analyses. While this procedure has the advantage of sharpening the conceptual distinctions among types, it also results in the exclusion of individuals with “mixed-type” profiles characterized by high levels on multiple traits. Because personality characteristics in reality are continuous and overlapping, the sample in this study may not fully represent the general working population. In particular, it should be noted that individuals with more balanced trait configurations may be underrepresented. Future research should incorporate mixed trait structures, for example, by employing latent profile analysis.

Regarding the relationship between these factors and aggression, while Sense of Superiority and Competence are positively correlated with verbal aggression during elementary to high school age, this correlation disappears by the time individuals reach university age (Sagara and Sagara 2006). Need for Attention and Praise indirectly promotes aggressive behavior through anger (Hibino et al. 2005). Individuals high in this factor tend to choose organizational dysfunctional behavior when they cannot accept negative evaluations from supervisors after perceiving an ego threat (Semba 2016b). For Self-Assertion, hierarchical regression analyses show positive effects on verbal aggression and verbal harassment in Japanese organizations (Tanaka 2008). Studies of Japanese junior high school students also indicate that Self-Assertion facilitates displacement of anger onto objects via anger-resolution cognition, since such forceful rationalization is not sufficient to suppress anger (Hibino et al. 2005).

Narcissism is closely linked to self-esteem. Oshio (1998) found that the three NPI subscales have distinct effects on the subscales of self-esteem, suggesting that each NPI factor serves distinct functions in regulating self-evaluation, and that coping strategies for regulating self-evaluation also differ across narcissistic types characterized by these factors.

Because maintaining a positive self-evaluation requires perceived acceptance from others (Leary and Baumeister 2000), social support contributes to maintaining self-evaluation by providing cues of acceptance (Genjida 2012). According to Cobb (1976), social support can be conceptualized as information signaling: (1) that one is cared for and loved, (2) that one is valued and respected, or (3) that one belongs to a network of mutual obligations. Individuals who perceive ego threats are likely to adopt maladaptive strategies, such as withdrawal from interpersonal relationships or aggressive responses, to compensate for the instability of their self-evaluation. In contrast, perceived social support is assumed to reduce defensive reactions to ego threats by stabilizing self-evaluation. Therefore, social support may fulfill the narcissistic need for positive self-evaluation and consequently decrease the likelihood of maladaptive coping strategies.

1.4. The Importance of Support in Organizations in Japanese Workplace Culture

Zhong et al. (2025) demonstrated that the psychological mechanisms underlying workplace aggression vary depending on cultural context. In particular, they reported that the effects of relationship quality and state self-evaluation are strengthened by cultural dimensions such as individualism and masculinity. Their findings underscore the significance of examining how ego threat is interpreted and functions within the Japanese cultural context.

Moreover, recent cross-cultural research has suggested that employees in Japanese cultural contexts tend to perceive negative evaluations and feedback not merely as task-related information but as interpersonal threats. For example, Kusakabe et al. (2024), in a comparative study of Japan and the United States, found that Japanese participants were more likely to overestimate the possibility of being negatively evaluated by others and showed a stronger tendency to avoid rejection. These findings suggest that employees in Japanese culture are more inclined to interpret negative feedback as a threat to interpersonal relationships.

Furthermore, Japan is characterized by an interdependent self-construal in which maintaining social face and interpersonal harmony is highly valued (Markus and Kitayama 1991; Heine et al. 1999). Within this cultural context, negative evaluations from others may be perceived as threats to interpersonal harmony, one’s social position, and relational stability within the group (Markus and Kitayama 1991). Consequently, the avoidance of negative evaluation in Japan carries personal, social, and relational implications.

Given these cultural norms, supportive behaviors may be interpreted in complex or unintended ways. Individuals may construe support as an implicit negative evaluation—for example, as a signal of inadequate ability or as evidence that they burden the group. In Japanese high-context communication environments, moderate levels of support are often viewed as consideration, whereas excessive support may be perceived as intrusive, controlling, or even as surveillance. In addition, the hierarchical nature of Japanese organizations makes it difficult for subordinates to express dissatisfaction even when supervisor support is experienced as disadvantageous. These considerations suggest that the effects of support are shaped not only by recipients’ psychological characteristics but also by culturally grounded interpretive norms.

Supervisors in Japanese organizations are generally expected to take an active role in developing their subordinates (Moro and Matsui 2009). However, because supervisory guidance can easily overlap with behaviors perceived as power harassment, supervisors must carefully consider how they provide support. Among various forms of supervisor support, this study focuses on negative evaluations, as negative feedback from supervisors tends to impose a particularly strong psychological burden and exerts a direct influence on subordinates’ self-esteem, job performance, and career prospects. Accordingly, this study examines whether modifications to the manner in which supervisors deliver negative evaluations can mitigate subordinates’ experiences of ego threat and, in turn, reduce power harassment.

Support does not necessarily function uniformly across individuals. Perceptions of support in ego-threatening situations vary depending on psychological characteristics (Kato et al. 2013). Some individuals may experience supervisory encouragement as reassuring and self-esteem enhancing, whereas others may interpret the same behavior as intrusive, patronizing, or demeaning, leading to further reductions in self-worth. Many legal disputes in Japan originate from conflicts over supervisory “guidance” (Japan Institute for Women’s Empowerment & Diversity Management 2011), suggesting that support does not always operate as intended. Recent research has also suggested that social support in the workplace does not necessarily produce uniformly positive effects. In particular, support that is unsolicited or that threatens the recipient’s sense of competence (competence-threatening support) may be perceived as a threat to self-esteem and can paradoxically lower self-esteem or intensify defensive reactions. For example, when workplace support is provided in an inappropriate manner, it may function not as a resource but as a stressor (i.e., a source of psychological strain) and has been shown to be associated with threats to self-esteem and stress responses (Semmer et al. 2025).

In other words, support does not always function in a protective manner; rather, it may operate promotively depending on the psychological and relational context. In situations of ego threat, support from supervisors may be interpreted as approval and acceptance in some cases, while in other cases it may be construed as implicit negative evaluation or surveillance. Particularly in hierarchical contexts, the meaning attributed to support varies according to individuals’ sensitivity to others’ evaluations and their self-regulatory tendencies. As a result, support may attenuate defensive reactions in some individuals, while amplifying aggressive responses in others. Based on these findings, the present study classifies employees according to narcissistic type and level of self-esteem to examine how these characteristics moderate the effects of supervisor support on power harassment.

Support also occurs at the organizational level. Perceived organizational support refers to employees’ “global beliefs concerning the extent to which the organization values their contributions and cares about their well-being” (Eisenberger et al. 1986). When such support is interpreted as a positive message from the organization, employees tend to enhance their self-esteem (Elloy and Randolph 1997). Furthermore, Zhong et al. (2025) demonstrated that relationship quality serves as a key mediating factor in the process through which workplace aggression influences citizenship behavior and deviant behavior. This relationship quality encompasses perceptions of trust and respect formed through interactions with supervisors and the organization, and is theoretically closely related to perceived organizational support—that is, the extent to which employees feel valued and supported by the organization. In other words, positive relationships with supervisors and the organization may foster the interpretation that the organization treats the self as a valued member, which in turn may contribute to the suppression of aggressive behavior. Building on these perspectives, this study incorporates both supervisor support and perceived organizational support as potential buffers against power harassment in Japanese workplaces.

Although the present study is grounded in Japanese workplace culture, the mechanisms examined here may not be limited to Japan but could also apply to other cultural contexts. In hierarchical social systems characterized by clear status differences and strong concerns for social face, support from authority figures may be interpreted not merely as assistance but also as evaluation, surveillance, or even an implication of incompetence. Within such contexts, organizational support can threaten relational face or social standing, thereby potentially amplifying defensive reactions among individuals who are highly sensitive to status and self-evaluation. From this perspective, Japanese workplace culture may be regarded as a salient example of a broader cultural dynamic in which hierarchical structures shape the psychological meaning and behavioral consequences of support.

Taken together, these considerations suggest that the psychological meaning of support is culturally contingent. In hierarchical contexts, support from supervisors may be experienced not purely as benevolent help but as implicit evaluation, particularly when individuals are already experiencing ego threat. Such interpretations may activate defensive self-regulatory processes and increase the likelihood of aggressive behavioral responses, especially among individuals with superiority-based narcissistic tendencies. Accordingly, in the present study, support is conceptualized not simply as a protective factor but as a culturally calibrated relational signal that can, under specific psychological and social conditions, amplify negative reactions.

1.5. Objective

The purpose of this study is to examine, among Japanese workers, the main effects of ego threat and support from supervisors and organizations on power harassment, as well as the moderating effects of both types of support on the relationship between ego threat and power harassment, according to different levels of self-esteem and types of narcissism. Furthermore, this study aims to derive practically effective measures for inhibiting power harassment and to discuss both theoretical and practical implications within the context of Japanese workplace culture.

1.6. Hypotheses

Based on this objective, the hypotheses are formulated. More specifically, the theoretical background and prior research reviewed above suggest that power-harassing behavior should be understood as emerging from the interaction between individuals’ experiences of ego threat and the support provided by others. Ego threat may elicit defensive reactions aimed at protecting self-worth and increase the risk of aggressive behavior, whereas support may function either suppressively or promotively depending on how it is interpreted. Moreover, such interpretations are likely to vary according to individuals’ narcissistic traits and levels of self-esteem. In other words, the same form of support may be received as approval or assistance by one individual, while being construed as evaluation or surveillance by another, thereby amplifying ego threat. Based on this framework, the present study examines the following hypotheses:

H1. 

Ego threat is positively associated with power-harassment tendencies.

H2a. 

Supervisor support exerts differential effects on power-harassment tendencies depending on psychological characteristics.

H2b. 

Organizational support exerts differential effects on power-harassment tendencies depending on psychological characteristics.

H3a. 

Supervisor support moderates the relationship between ego threat and power-harassment tendencies.

H3b. 

Organizational support moderates the relationship between ego threat and power-harassment tendencies.

H4. 

The moderating effects of support differ according to narcissistic type and level of self-esteem.

2. Materials and Methods

2.1. Overview of the Survey

This survey was conducted in May 2024 through an online survey platform operated by Company A. Company A recruited respondents and administered a questionnaire survey. The collected data were screened to select cases that met the quality control criteria (i.e., passed attention checks), could be classified by narcissistic type (i.e., among the three subscales of the Narcissistic Personality Inventory Short version (Oshio 1998) (hereafter NPI-S), only one had the highest score), and contained no missing values. Qualified cases were then distributed into each quota group. This process was repeated across all quotas—defined by gender, age, and narcissistic type—until the target sample size was achieved (see Figure 1).

Informed consent was obtained through Company A’s system by presenting participants with an explanatory statement describing the research purpose, voluntary nature of participation, right to withdraw, and procedures for handling of personal information. Participants provided consent by checking a box (responses could not be submitted without providing consent). All data were anonymized by Company A, and no personally identifiable information (e.g., IP addresses) was disclosed to the researcher.

To reduce evaluation concerns and social desirability bias, participants were assured that their responses would be anonymous and analyzed only at the aggregate level. Furthermore, by employing an independent research company, the direct relationship between researchers and respondents was separated, thereby alleviating perceived pressure to provide socially desirable responses.

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Kyoto Tachibana University (Approval No. 24-11, dated 13 May 2024).

2.2. Classification by Self-Esteem and Narcissistic Types

2.2.1. Classification by Self-Esteem

Based on the total score of the Japanese version of the Self-Esteem Scale (Yamamoto et al. 1982), participants were divided into two groups—a high self-esteem group and a low self-esteem group—using the median value of the sample distribution. This splitting was theoretically justified, as prior research has suggested that individuals with high versus low self-esteem may exhibit qualitatively different response patterns to ego threat (Rosenberg 1965; Kernis et al. 1993; Shimizu 1999; Kawamura and Toda 2014).

2.2.2. Classification by Narcissistic Types

This study was based on the premise that behavioral strategies for maintaining positive self-evaluation differ depending on the factors of narcissism. Accordingly, participants were classified using the three subscales of the NPI-S: Self-Assertion, Need for Attention and Praise, and Sense of Superiority and Competence.

For each participant, total scores on the three subscales were calculated, and the subscale with the highest score was identified as the dominant characteristic, determining the participant’s narcissistic type. This classification method follows prior studies (Semba 2018), which have demonstrated that individuals exhibit different response patterns under ego-threatening situations according to their dominant narcissistic factor.

2.3. Survey Items

The survey in this study consisted of a face sheet, scales assessing narcissism and self-esteem, power harassment, an ego threat scale, and supervisor and organizational support scales. Except for the face sheet, all instruments used were developed or modified in Japan to ensure that the psychological states of workers were accurately assessed within the Japanese linguistic and cultural context.

2.3.1. Face Sheet

This section asks respondents to indicate their gender, age, and occupation.

2.3.2. Narcissism Scale

The Narcissistic Personality Inventory Short version (Oshio 1998, 1999) (NPI-S) was used in this study. This instrument consists of 30 items and includes three subscales: Self-Assertion, Sense of Superiority and Competence, and Need for Attention and Praise (Oshio 2004). The NPI-S is an adaptation of the original NPI (Raskin and Hall 1979), modified to better reflect the Japanese cultural context. In this survey, participants rated each item on a five-point Likert scale ranging from 1 (does not apply at all) to 5 (applies very well). The total score for each subscale was calculated, and participants were classified into the narcissistic type corresponding to the subscale on which they scored the highest.

2.3.3. Self-Esteem Scale

The Japanese version of the Self-Esteem Scale (Yamamoto et al. 1982) was used in this study. This 10-item scale is structured as a single-factor measure. Participants rated their usual level of esteem on a four-point scale ranging from 1 (No) to 4 (Yes), with higher total scores indicating higher self-esteem.

2.3.4. Power Harassment Scale

As far as the author is aware, there is no existing standardized scale for measuring power harassment; accordingly, this study developed a scale based on the six types of workplace power harassment defined by the Ministry of Health, Labour and Welfare (2020): physical attack, psychological attack, social isolation, excessive demands, insufficient demands, and invasion of privacy. Behaviorally specific items were generated to reflect these categories in ways appropriate to Japan’s legal and cultural context (see

Table 1. 6 Types of Power Harassment and Related Items.

Type of Power Harassment	Item
Physical Attack	1. I lost my temper and ended up using physical force. Examples: shaking, pushing, or hitting.
	2. When I became angry, I ended up taking it out on objects. Examples: hitting the desk, throwing documents, or kicking a chair.
Psychological Attack	3. I reprimanded a worker loudly in front of others.
	4. I sent an email (or similar message) that disparaged or denied a worker’s abilities, addressing not only the person but also other workers.
Social Isolation	5. I ignored a single worker collectively as a group.
	6. I deliberately assigned a specific worker to a private room, preventing them from having contact with other workers.
Excessive Demands	7. I required a worker to perform menial tasks unrelated to their job, not allowing them to refuse, or forced them to participate in optional social gatherings or events.
	8. I assigned quotas or tasks that were impossible to accomplish.
Insufficient Demands	9. I deliberately withheld work assignments from a worker as a form of mistreatment.
	10. I assigned work that was significantly below the worker’s abilities or competence.
Invasion of Privacy	11. I disclosed a worker’s personal information—such as sexual orientation, gender identity, medical history, or infertility treatment—without their consent.
	12. I inquired into a worker’s private relationships or friendships.

).

The development process proceeded in several steps. First, a prior study (Japan Institute for Women’s Empowerment & Diversity Management 2011) and a government document (Ministry of Health, Labour and Welfare 2020) were reviewed in detail to generate an initial pool of candidate items covering all six types of power harassment. Second, two certified senior industrial counselors and the author (also certified) conducted an expert review. Each item was evaluated on four criteria—validity, clarity of meaning, and workplace appropriateness. Items were revised or removed based on the evaluation results. Third, cognitive interviews were conducted with a diverse sample of employees (n = 10) to examine item comprehension and response interpretation. A small pilot survey (n = 20) was then conducted to assess response burden and score distribution. These procedures produced the final set of items.

The final scale comprised 12 items, each rated on a 10-point Likert scale ranging from 1 (strongly disagree) to 10 (strongly agree) (see Table 1). Higher total scores indicated a stronger tendency to select or engage in power harassment behaviors. For the purposes of this study—namely, theoretical validation of the proposed framework and the provision of policy and practical implications—the total score across all 12 items was used as the index of power harassment tendencies, rather than analyzing the six subscales separately.

Based on the theoretical position that conceptualizes power harassment as a higher-order construct representing defensive-aggressive reactions arising within hierarchical workplace relationships, the present study employed a total composite score rather than scores for each subscale. Although the six categories in the Japanese context are legally defined classifications, this study positioned them as diverse behavioral manifestations of a common underlying construct—workplace aggression expressed against a background of power asymmetry. Because the primary objective of this study was to examine the psychological mechanisms linking ego threat, narcissistic tendencies, and the selection of power-harassing behaviors, the use of an aggregated index was considered theoretically appropriate. Furthermore, integrating the six categories made it possible to capture individuals’ general propensity to engage in power-harassment responses rather than risk associated with any specific behavioral form. This approach is also consistent with Robinson and Bennett (1995) and Spector et al. (2006), who treat workplace aggression as a multidimensional yet conceptually integrated construct.

It should be noted that the development of a new psychometric scale per se was not the primary objective of this study; therefore, the development process was intentionally limited to the essential procedures required for generating a valid measure.

2.3.5. Ego Threat Scale

The State Self-Esteem Scale developed by Heatherton and Polivy (1991) was used in this study. Following Tabata and Ikegami (2011), eight items from the performance and social subscales were selected, while items irrelevant to the study’s purpose (e.g., “I am concerned about my weight”) were excluded. This scale measures an individual’s self-evaluation at the present moment (Abe and Konno 2007) and has been used in prior studies as a measure of ego threat (e.g., Watanabe and Karasawa 2012).

Consistent with these prior studies, participants were instructed as follows to restrict their responses to their current emotional state: “The following questions are designed to assess what you are thinking right now. Please respond based on how you feel at this moment, rather than how you usually feel”.

Items were rated on a five-point Likert scale ranging from 1 (does not apply) to 5 (applies). Higher total scores indicate greater perceived ego threat.

2.3.6. Supervisor Support Scale

Supervisor support was measured using the Social Support Scale developed by Komaki (1994). This scale consists of 14 items and includes two subscales: emotional support and instrumental support. In this survey, the total score across all items was used for analysis. Participants rated the support they received from their immediate supervisors on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Higher total scores indicate greater perceived supervisor support.

2.3.7. Organizational Support Scale

Perceived organizational support was assessed using a Japanese translation of the Organizational Support Scale developed by Eisenberger et al. (1986). This scale measures employees’ beliefs regarding the extent to which their organization values their contributions and cares about their well-being. The scale has a one-factor structure, with several items sharing similar content. Therefore, considering the purpose of the present study and the response burden on participants, three items with high factor loadings were selected: “The company (organization) takes pride in my accomplishments at work” (loading: 0.76), “The company (organization) really cares about my well-being” (0.83), and “The company (organization) values my contributions to its well-being” (0.71). Participants rated on a seven-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). Higher total scores indicate greater perceived organizational support.

2.4. Statistical Analysis

First, factor analyses using principal axis factoring with Promax rotation were conducted for all scales, and items deemed inappropriate were removed. Next, basic descriptive statistics were calculated. Subsequently, correlation coefficients were calculated, followed by hierarchical multiple regression analyses to examine (a) how the effects of ego threat and supervisor/organizational support on power harassment vary depending on levels of self-esteem and narcissistic type, and (b) how the two forms of support moderate these relationships.

Power harassment was entered as the dependent variable, and gender, age, occupation, ego threat, and the interaction terms of ego threat with supervisor and organizational support (ego threat × supervisor support, ego threat × organizational support, and ego threat × supervisor support × organizational support) were entered as independent variables.

Specifically, in Step 1, gender, age, and occupation were entered as independent variables. In Step 2, ego threat, supervisor support, and organizational support were added. In Step 3, the interaction terms between ego threat and the support variables were included. To avoid multicollinearity, the interaction terms were centered following Aiken and West (1991). The significance level was set at 5%, and SPSS ver. 28 and AMOS ver. 28 were used for the statistical analyses.

3. Results

3.1. Descriptive Statistics

Company A administered the questionnaire survey to a total of 1621 individuals and obtained 600 valid responses (valid response rate: 37.0%). The sample consisted of 300 men and 300 women (see

Table 2. Participant Attributes.

		n	%
Sex	Male	300	50.0
	Female	300	50.0
	Total	600	100.0
Occupation	Executives/Manager	10	1.7
	Company Employee	386	64.3
	Public Employee/Government Workers	29	4.8
	Self-employed/Freelance Worker	42	7.0
	Part-time/Temporary Worker	133	22.2
	Total	600	100.0

). The mean age was 44.8 ± 13.4 years (men: 44.8 ± 13.5 years; women: 44.8 ± 13.4 years). Regarding occupational categories, 10 respondents (1.7%) were corporate or organizational executives/managers, 386 (64.3%) were company employees, 29 (4.8%) were public-sector employees or government workers, 42 (7.0%) were self-employed or freelance workers, and 133 (22.2%) were part-time or temporary workers (see Table 2).

Regarding self-esteem groups and narcissistic types, 302 were categorized into high self-esteem group (Self-Assertion type: 94, Need for Attention and Praise type: 76, Sense of Superiority and Competence type: 132) and 298 into low self-esteem group (Self-Assertion type: 106, Need for Attention and Praise type: 124, Sense of Superiority and Competence type: 68) (see Figure 2).

Note that as a result of the factor analysis, item Q8 of the Self-Esteem Scale and item Q5 of the State Self-Esteem Scale were excluded from analyses due to their low communality values (0.020 and 0.046, respectively).

3.2. Factor Analysis of Power Harassment Scale

An exploratory factor analysis (EFA) using the principal factor method with promax rotation was conducted on the 12 items measuring power harassment. The results indicated that a one-factor solution explained 81.7% of the total variance, and Cronbach’s alpha demonstrated excellent internal consistency (0.979) (

Table 3. Pattern Matrices for the Power Harassment Scale (n = 600).

1-Factor Model			2-Factor Model
Item	Factor	Communality	Item	Factor 1	Factor 2	Communality
6	0.964	0.930	11	0.953	0.023	0.944
11	0.964	0.929	8	0.951	0.025	0.943
8	0.963	0.928	7	0.931	0.023	0.902
9	0.951	0.905	9	0.922	0.044	0.916
7	0.942	0.888	6	0.916	0.065	0.938
5	0.940	0.884	5	0.858	0.103	0.886
4	0.938	0.880	10	0.837	0.110	0.858
10	0.925	0.856	12	0.825	0.098	0.819
12	0.903	0.816	4	0.812	0.153	0.879
1	0.840	0.706	2	−0.073	0.832	0.601
3	0.820	0.672	3	0.209	0.714	0.790
2	0.641	0.410	1	0.279	0.656	0.797
α	0.979	–	α	0.879	0.987	–

Principal axis factoring and promax rotation were applied.

). A subsequent confirmatory factor analysis (CFA) showed that although the CFI and TLI indicated good fit (0.979 and 0.972, respectively), the RMSEA exceeded the conventional cutoff value of 0.08 (0.092), suggesting that overall model fit was not fully satisfactory (see

Table 4. Results of a Confirmatory Factor Analysis for the Power Harassment Scale (n = 600).

Model	χ2	df	CFI	TLI	RMSEA	AIC
1-Factor	307.371 *	51	0.979	0.972	0.092	361.371
2-Factor	206.304 *	50	0.987	0.983	0.072	262.304

* p < 0.001, CFI: Comparative Fit Index, TLI: Tucker–Lewis Index, RMSEA: Root Mean Square Error of Approximation, AIC: Akaike Information Criterion.

and Figure 3).

Accordingly, the model was revised to a two-factor structure and reexamined using CFA. The two-factor solution explained 85.6% of the total variance, and the alpha coefficients for the two factors indicated good reliability (0.879 and 0.987) (Table 3). The CFA results showed good fit indices (CFI = 0.987, TLI = 0.983), and the RMSEA fell within the acceptable range (0.072) (see Table 4 and Figure 3). In addition, the AIC decreased substantially compared to the one-factor model (361.371 → 262.304), indicating that the two-factor model provided a better and more parsimonious fit to the data. CFAs were also conducted for three- to six-factor models; however, these solutions either converged to the two-factor model or exhibited cross-loadings (see Appendix A).

Based on these results, the two-factor structure was judged to be appropriate. Factor 2 consisted of items reflecting physical aggression (Items 1, 2) and “yelling reprimands” (Item 3) and was interpreted as representing “severe aggression.” In contrast, Factor 1 was interpreted as representing relatively “milder aggression.” However, because the present study focused on the overall intensity of power harassment, analyses were conducted using the total score.

3.3. Hierarchical Multiple Regression Analysis

The means, standard deviations, and correlation coefficients of the variables included in the hierarchical multiple regression analysis are presented in

Table 5. Descriptive Statistics and Correlations of Variables Used in the Regression Analysis (High Self-Esteem Group, n = 302).

Variable	M	SD	2	3	4
1 Power Harassment	2.00	1.79	0.436 **	−0.130 *	0.025
2 Ego Threat	2.46	0.79	-	−0.150 **	−0.130 *
3 Supervisor Support	3.40	0.80			0.524 **
4 Organizational Support	4.09	1.26			-

* p < 0.05, ** p < 0.01.

and

Table 6. Descriptive Statistics and Correlations of Variables Used in the Regression Analysis (Low Self-Esteem Group, n = 298).

Variable	M	SD	2	3	4
1 Power Harassment	2.27	1.79	−0.152 **	0.091	0.196 **
2 Ego Threat	3.42	0.70	-	−0.192 **	−0.184 **
3 Supervisor Support	2.85	0.92			0.558 **
4 Organizational Support	3.09	1.32			-

** p < 0.01.

.

The Step 3 regression results showed that, among participants with high self-esteem, the regression model was significant for Self-Assertion type (R² = 0.273, p < 0.01), Need for Attention and Praise type (R² = 0.317, p < 0.01), and Sense of Superiority and Competence type (R² = 0.353, p < 0.001) (

Table 7. Main Effects of Ego Threat and Support on Power Harassment, and the Moderating Effect of Support (High Self-Esteem Group, n = 302).

Independent Variable	Self-Assertion Type (n = 94)			Need for Attention and Praise Type (n = 76)			Sense of Superiority and Competence Type (n = 132)
	Step 1	Step 2	Step 3	Step 1 *	Step 2 *	Step 3	Step 1	Step 2	Step 3
Sex (1-Female, 2-Male)	0.073	0.044 ***	0.064 ***	0.168 *	0.092 ***	0.079 ***	0.264 ***	0.197 ***	0.174 ***
Age	−0.216	−0.178 ***	−0.192 ***	−0.133 *	−0.100 ***	−0.116 ***	−0.133 ***	−0.051 ***	−0.045 ***
Executives/Manager +	0.008	0.026 ***	−0.003 ***	-	-	-	0.246 ***	0.228 ***	0.233 ***
Company Employee +	0.138	0.085 ***	0.049 ***	−0.075 *	−0.031 ***	−0.031 ***	−0.042 ***	−0.045 ***	−0.072 ***
Public Employee /Government Workers +	0.013	0.009 ***	0.013 ***	-	-	-	−0.067 ***	−0.057 ***	−0.065 ***
Self-employed/Freelance Worker +	−0.008	−0.032 ***	−0.034 ***	−0.159 *	−0.126 ***	−0.098 ***	−0.116 ***	−0.096 ***	−0.069 ***
Ego Threat		0.372 ***	0.407 ***		0.412 ***	0.415 ***		0.342 ***	0.204 ***
Supervisor Support		−0.202 ***	−0.241 ***		−0.297 ***	−0.384 ***		−0.035 ***	−0.054 ***
Organizational Support		0.221 ***	0.222 ***		0.202 ***	0.258 ***		0.149 ***	0.142 ***
Ego Threat × Supervisor Support			0.119 ***			−0.185 ***			−0.045 ***
Ego Threat × Organizational Support			−0.162 ***			0.096 ***			0.116 ***
Ego Threat × Supervisor Support × Organizational Support			−0.055 ***			−0.039 ***			0.251 ***
R2	0.095	0.257 ***	0.273 ***	0.058 *	0.295 ***	0.317 ***	0.168 ***	0.288 ***	0.353 ***
ΔR2	0.095	0.163 ***	0.015 ***	0.058 *	0.238 ***	0.022 ***	0.168 ***	0.120 ***	0.065 ***

* p < 0.05, *** p < 0.001. ⁺ Part-time/Temporary Worker was set as the reference. - No respondents.

). Ego threat showed a positive main effect on power harassment for Self-Assertion type (β = 0.407, p < 0.001), Need for Attention and Praise type (β = 0.415, p < 0.01), and Sense of Superiority and Competence type (β = 0.204, p < 0.05). Moreover, in the Need for Attention and Praise type, supervisor support showed a negative main effect (β = −0.384, p < 0.01), whereas organizational support showed a positive main effect (β = 0.258, p < 0.05). Furthermore, for the Sense of Superiority and Competence type, the three-way interaction among ego threat, supervisor support, and organizational support had a positive moderating effect (β = 0.251, p < 0.05).

In contrast, among participants with low self-esteem, only the regression model for the Need for Attention and Praise type was significant (R² = 0.263, p < 0.001) (

Table 8. Main Effects of Ego Threat and Support on Power Harassment, and the Moderating Effect of Support (Low Self-Esteem Group, n = 298).

Independent Variable	Self-Assertion Type (n = 106)			Need for Attention and Praise Type (n = 124)			Sense of Superiority and Competence Type (n = 68)
	Step 1	Step 2	Step 3	Step 1 *	Step 2 *	Step 3	Step 1	Step 2	Step 3
Sex (1-Female, 2-Male)	0.080 *	0.030 **	0.025 **	0.230 **	0.237 **	0.279 ***	0.177 *	0.198 *	0.194 *
Age	−0.019 *	0.041 **	0.036 **	−0.070 **	−0.112 **	−0.096 ***	−0.006 *	−0.041 *	−0.052 *
Executives/Manager +	0.022 *	0.067 **	0.051 **	0.182 **	0.167 **	0.156 ***	-	-	-
Company Employee +	0.056 *	−0.010 **	−0.007 **	0.168 **	0.160 **	0.133 ***	0.103 *	0.078 *	0.076 *
Public Employee /Government Workers +	0.136 *	0.154 **	0.171 **	0.182 **	0.156 **	0.154 ***	−0.123 *	−0.182 *	−0.188 *
Self-employed/Freelance Worker +	0.021 *	−0.024 **	−0.029 **	−0.024 **	−0.035 **	−0.050 ***	−0.030 *	−0.044 *	−0.034 *
Ego Threat		−0.055 **	−0.036 **		−0.252 **	−0.314 ***		−0.084 *	−0.150 *
Supervisor Support		0.004 **	−0.004 **		0.001 **	0.030 ***		−0.227 *	−0.268 *
Organizational Support		0.386 **	0.380 **		0.033 **	0.117 ***		0.109 *	0.022 *
Ego Threat × Supervisor Support			−0.029 **			−0.053 ***			0.292 *
Ego Threat × Organizational Support			0.108 **			0.095 ***			0.074 *
Ego Threat × Supervisor Support × Organizational Support			−0.043 **			0.267 ***			0.049 *
R2	0.027 *	0.171 **	0.179 **	0.140 **	0.206 **	0.263 ***	0.061 *	0.107 *	0.199 *
ΔR2	0.027 *	0.144 **	0.008 **	0.140 **	0.066 **	0.057 ***	0.061 *	0.045 *	0.092 *

* p < 0.05, ** p < 0.01, *** p < 0.001. ⁺ Part-time/Temporary Worker was set as the reference. - No respondents.

). In this type, ego threat showed a negative main effect on power harassment (β = −0.314, p < 0.01), and the three-way interaction among ego threat, supervisor support, and organizational support again indicated a positive moderating effect (β = 0.267, p < 0.05).

Note that because all VIF values were below 3.0, multicollinearity was not considered a concern.

4. Discussion

As a result of the hierarchical multiple regression analysis, the Step 3 regression model was significant for all narcissistic types in the high self-esteem group, whereas it was significant only for the Need for Attention and Praise type in the low self-esteem group. Accordingly, the discussion primarily focuses on (a) differences among narcissistic types within the high self-esteem group, for which the model was significant across all types, and (b) differences between high and low self-esteem within the Need for Attention and Praise type, for which the regression model was significant in both groups.

4.1. Differences by Narcissistic Type Within the High Self-Esteem Group

In the high self-esteem group, significant main effects of supervisor support and organizational support were observed only in the Need for Attention and Praise type; however, the two forms of support exhibited opposite influences on power harassment. Specifically, when individuals perceived support from their supervisors, they were less likely to report engaging in power harassing behaviors, suggesting a potential mitigating association. In contrast, when organizational support was introduced, the tendency to choose power harassment increased.

This seemingly inconsistent pattern can be explained by psychological factors such as the stability of self-evaluation and the degree to which individuals depend on external evaluation. Although individuals with the Need for Attention and Praise type strongly desire recognition and attention from others, when their self-esteem is high, their self-evaluations are relatively stable. As a result, they experience less anxiety or threat in response to external evaluation. In this state, support from a supervisor—who is close and likely to have an established trust relationship—is interpreted as acknowledgment or constructive evaluation rather than a threat to self-worth. Consequently, such support is received as reasonable advice or informational assistance and is unlikely to invoke defensive or aggressive behavior, thus power harassment is suppressed.

From a cultural perspective, the findings of the present study may reflect hierarchical relational dynamics that are not limited to Japan. In organizational environments characterized by clear status differences and strong concerns for relational face, support from authority figures may be interpreted not merely as benevolent assistance but as implicit evaluation or an indication of incompetence. Within such contexts, receiving support itself may threaten one’s social standing or relational face, thereby activating defensive self-regulatory processes. Particularly among individuals who are highly sensitive to external evaluation, support may function less as emotional or instrumental assistance and more as a status-relevant signal. Accordingly, the present findings can be positioned as suggesting a broader cultural mechanism in which hierarchical structures shape the psychological meaning and behavioral consequences of support.

In contrast, organizational support is often provided in a formalized and standardized manner, with weaker interpersonal relationships. Consequently, its intentions are less transparent. For high self-esteem individuals with the Need for Attention and Praise type, such formalized support is more easily perceived as a signal of being monitored or evaluated, and thus as a potential threat to self-worth. According to the Threatened Egotism Model, when potential threats to one’s self-image or self-worth arise, defensive aggression is more likely to be expressed as power-harassing behavior. Given their heightened sensitivity to others’ evaluations, formal organizational support may function as pressure on their self-worth for this type, increasing the likelihood of selecting aggressive responses as a defensive reaction.

Regarding the moderation effects, only the Sense of Superiority and Competence type showed a significant three-way interaction among ego threat, supervisor support, and organizational support. Under ego-threatening conditions, the simultaneous perception of both forms of support increased the likelihood of selecting power harassment. This tendency may be due to this type’s heightened concern for maintaining self-efficacy and status, stemming from their sense of superiority and competence orientation. Concurrent external interventions may be perceived as “intrusions on autonomy” or “pressures on status.” Although supervisor support and organizational support individually exerted little influence, their simultaneous presence can be interpreted as signals that one’s judgment is being overridden or that one’s status is being undermined, thereby triggering defensive aggression.

Taken together, individualized and dialogic support provided by proximal supporters—such as supervisors—appears effective in suppressing power harassment for individuals characterized by the Need for Attention and Praise type. In contrast, under ego-threatening conditions, providing supervisor support and organizational support simultaneously may have adverse effects for individuals characterized by the Sense of Superiority and Competence type. These findings suggest that power harassment-prevention strategies and support systems must consider not only the quantity of support but also its quality, the relationship between the supporter and the recipient, the recipient’s narcissistic type, and the timing of support delivery (simultaneous vs. independent).

4.2. Differences by Self-Esteem Level Within the Need for Attention and Praise Type

4.2.1. Main Effect of Ego Threat

Among individuals classified as the Need for Attention and Praise type, the impact of ego threat on power harassment differed markedly by self-esteem level—positively in the high self-esteem group and negatively in the low self-esteem group. For those with high self-esteem, elevated self-evaluations combined with heightened sensitivity to others’ approval make defensive motives more likely to be activated when they faced with ego threat. Such motives may manifest as aggressive behaviors, including power harassment. This tendency is consistent with this type’s psychological profile: they rely on external validation as a key source of self-worth and are therefore particularly inclined to perceive negative feedback or unmet expectations as threats to the self.

However, these behavioral tendencies should be interpreted with caution, as culturally embedded norms regarding hierarchy and face may influence how participants evaluate and report their own behavioral responses.

Cultural characteristics of Japanese workplaces may further intensify this reaction. Norms surrounding hierarchical senior–junior relationships, group harmony (wa), the distinction between public façade (tatemae) and private feelings (honne), and the protection of social face (Ohno 2005; Markus and Kitayama 1991) can make negative evaluations—such as public reprimands that damage one’s social face—especially salient as an ego threat. Such humiliating events may amplify defensive motivations aimed at restoring one’s status or reputation, ultimately increasing the likelihood of aggressive responses.

In contrast, individuals in the low–self-esteem group, whose baseline self-evaluations are less favorable, tend to adopt avoidance-oriented strategies rather than aggressive ones when encountering ego-threatening situations. Consequently, power harassment is relatively suppressed among them. This avoidance tendency may be further reinforced in cultural contexts that prioritize harmony and conformity (Markus and Kitayama 1991). Overall, these findings indicate that although ego threat serves as a central psychological antecedent of power harassment, the direction and magnitude of its influence depend critically on the individual’s level of self-esteem.

4.2.2. Main Effects of Supervisor Support and Organizational Support

The present findings highlight that supervisor support and organizational support function differently in shaping aggressive behaviors, depending on individuals’ self-esteem levels and narcissistic tendencies. Among high self-esteem employees with strong needs for attention and praise, supervisor support appeared to buffer against power harassment. This suggests that when individuals possess both a strong desire for approval and sufficient self-regard, they can interpret supervisors’ guidance as informational or emotionally supportive rather than evaluative or threatening. This interpretation aligns with cultural norms in East Asian societies, including Japan, where relational and emotionally oriented leadership is often viewed positively and understood within high-context communication as “protection” or “consideration” (Farh and Cheng 2000; Cheng et al. 2004; Doi 1971). Such forms of support may therefore help maintain face while alleviating declines in self-evaluation.

In contrast, neither supervisor support nor organizational support showed significant effects among individuals with low self-esteem. Threatened egotism theory (Bushman and Baumeister 1998) suggests that those with fragile self-worth are prone to interpret others’ support as criticism, monitoring, or evidence of their own inadequacy. When support is interpreted in this way, psychological acceptance does not occur, and the suppressive effects of support on aggressive behavior cannot materialize.

Notably, organizational support was related to a higher likelihood of harassment among high self-esteem employees with a strong need for approval. This counterintuitive effect underscores that high self-esteem does not necessarily imply stable or secure self-evaluation. For individuals whose self-worth hinges on external recognition, organizational support—particularly in the form of evaluations, personnel systems, or formal communication—may be interpreted as monitoring or judgment rather than assistance. In the Japanese cultural context, where hierarchical structures and uncertainty avoidance are comparatively strong, and where stability in role and status is closely tied to individuals’ social standing (Hofstede 2001; House et al. 2004), formal evaluations that may alter one’s role or status are easily interpreted not as routine managerial procedures but as signals of declining status or loss of face. Under such conditions, organizational support may amplify ego threat, activate defensive aggression, and increase the risk of harassment.

Taken together, these findings suggest that the psychological meaning of “support” is not uniform. Rather, it is filtered through individuals’ self-esteem and approval-related tendencies as well as the cultural context in which interactions occur. In Japan, where relational support from supervisors is more readily interpreted as benevolent, but organizational interventions tend to carry negative evaluative implications, distinguishing between these two types of support becomes a critical moderator of employees’ emotional and behavioral responses. These results underscore the need for culturally sensitive and psychologically attuned organizational support practices that account for individual differences in self-worth regulation.

4.2.3. Differences in the Moderating Effects of Support by Self-Esteem

In the high self-esteem group, the two-way interactions of “ego threat × supervisor support” and “ego threat × organizational support” were not significant. However, the three-way interaction of “ego threat × supervisor support × organizational support” showed a significant positive moderating effect on the relationship between ego threat and power harassment for individuals of the Sense of Superiority and Competence type. This suggests that layered support may be perceived as “excessive interference” or “a threat to self-efficacy,” thereby triggering defensive aggression aimed at maintaining a sense of superiority. Because high self-esteem individuals often highly value autonomy and self-efficacy, they may interpret external support as questioning their competence, increasing the risk of counterproductive outcomes.

The direction of defensive responses also varies by narcissistic type. For the Sense of Superiority and Competence type, with external evaluation as the reference, individuals tend to adopt internal defensive strategies such as avoidance or self-regulation, which can manifest as harassment under similar supportive conditions.

In the low self-esteem group, the same three-way interaction exhibited a significant positive effect on power harassment for the Need for Attention and Praise type. Low self-esteem individuals have unstable self-evaluations and are less able to perceive support as benevolent; instead, they often interpret it as “being evaluated” or “monitored.” Consequently, support may unintentionally amplify ego threat, thereby increasing the likelihood of harassment. When support is provided simultaneously by both supervisors and the organization, evaluation-related anxiety may compound, further elevating the risk of harassment behaviors.

4.2.4. Conceptual Clarification of Power Harassment

In the present study, power harassment was conceptualized as a higher-order expression of defensive aggression arising within hierarchical workplace relationships. Accordingly, the analysis focused not on specific behavioral forms but on a general tendency toward power-harassing behaviors as a whole. Although the six categories defined in the Japanese administrative classification capture distinct observable behaviors, they may reflect a common underlying psychological process related to ego threat and the regulation of self-worth. Therefore, the aggregated index employed in this study should be interpreted as indicating a generalized behavioral tendency rather than risk associated with any specific form of behavior.

4.2.5. Summary of Hypothesis Testing

The hypotheses formulated in the present study (H1–H4) were generally supported. Specifically, among the high self-esteem group, ego threat was positively associated with power harassment, whereas among the low self-esteem group, it was negatively associated with power harassment (partially supporting H1). Supervisor support functioned suppressively among individuals with high self-esteem and a strong need for attention and praise, whereas organizational support was found to function in a facilitative manner or to exert little effect, depending on individuals’ psychological characteristics and narcissistic type (supporting H2a/H2b). Furthermore, the interaction effects between support and ego threat varied according to individuals’ narcissistic traits and levels of self-esteem (partially supporting H3a/H3b and supporting H4), confirming that the psychological meaning and interpretation of support depend on individual differences. Taken together, these findings demonstrate the validity of the theoretical framework underlying the hypotheses and underscore the necessity of individualized approaches in designing support-based interventions.

4.2.6. Implications, Limitations, and Directions for Future Research

The findings of this study suggest that for certain narcissistic types and the level of the self-esteem, support from supervisors and organizations does not invariably function as a suppressive factor against power harassment. Under specific conditions, support may instead amplify ego threat and consequently increase the likelihood of engaging in power harassment. Particularly in contexts where multiple layers of support are provided simultaneously, there is a risk that such support may have unintended adverse effects depending on individuals’ narcissistic orientation and level of self-esteem. These results offer important insights for reconsidering how support is structured within organizations. They highlight the necessity of considering not only the content of support but also its timing, delivery method, and the psychological characteristics of recipients.

These findings suggest that flexible and individualized support, taking into account employees’ psychological dispositions, may be beneficial for workplace assistance, although causal effects cannot be confirmed. For individuals with strong needs for attention and praise and low self-esteem, support should be offered in ways that do not appear imposed—such as by offering choices or introducing interventions gradually. For those with strong sense of superiority and competence and high self-esteem, support should be framed as informational or auxiliary rather than evaluative, avoiding implications that their competence is being questioned. By tailoring the method, timing, and provider of support to individual characteristics, organizations can simultaneously prevent power harassment and promote psychological adjustment.

From a practical standpoint, (a) supervisor support should prioritize personalized feedback and consensual, employee-centered guidance rather than public instruction or education; (b) organizational support should emphasize consultation, mediation by neutral third parties, and the assurance of psychological safety rather than surveillance or control; and (c) visible evaluation systems such as awards or rankings may inadvertently motivate power harassment among employees with strong needs for praise, and therefore should be implemented cautiously in conjunction with behavioral norms and indicators of mutual respect (Hofstede 2001; House et al. 2004). Furthermore, by focusing on the interaction between individual factors (narcissistic type and self-esteem level) and organizational forms of support, this study contributes a novel perspective to the understanding of mechanisms underlying power harassment in workplaces.

In discussing the limitations, this study may be subject to social desirability bias and common method bias associated with the self-report measurement of aggressive behavior. In other words, participants may have underreported their own aggressive tendencies or reinterpreted their behaviors in more socially acceptable ways. In the present study, several procedural steps were taken to mitigate these risks. Specifically, data were collected anonymously through an independent research company, and participants were clearly informed that the researchers would not have access to personally identifiable information and that responses would be analyzed only at the aggregate level. Nevertheless, to examine the potential influence of common method bias more rigorously, future research should incorporate multiple sources of information (e.g., peer ratings, supervisor evaluations, and behavioral indicators) and adopt experimental or longitudinal research designs.

In addition to measurement-related issues, the present study also has several more general limitations. First, its cross-sectional design limits causal inference; longitudinal studies are needed to confirm temporal order. Second, the reliance on self-report data introduces the possibility of biases due to social desirability or memory distortions. Third, the sample consisted entirely of Japanese workers, limiting the generalizability of findings to other cultural contexts. Moreover, because this study focused on narcissistic types and self-esteem, it did not address other individual factors such as personality traits, stress-coping capacity, or emotion-regulation skills. Future research should employ longitudinal and intervention-based designs to accumulate empirical evidence that can inform the development of effective workplace support systems.

5. Conclusions

This study demonstrated that supervisor and organizational support can influence the manifestation of power harassment. In particular, the findings indicate that when perceptions of ego threat increase, the impact of support varies depending on its form and provider. For example, while support from supervisors tended to suppress aggressive behavior, organizational support could instead potentially facilitate power harassment. These results suggest that the influence of support is shaped not only by its content but also by the recipient’s psychological characteristics.

Self-esteem and narcissistic tendencies function as key moderators shaping how support affects power harassment. Workers with high self-esteem were more likely to interpret organizational support as a threat to their competence, which in turn increased the likelihood of aggressive behavior via heightened ego threat. In contrast, workers with low self-esteem were generally more receptive to support but could still perceive it as excessive pressure or expectations, resulting in counterproductive reactions. Additionally, the effects differed by narcissistic type: individuals characterized by a strong need for attention and praise were particularly sensitive to external approval, and their behavior was readily influenced by perceived external evaluation.

This article is based on, and has been revised and expanded from, a poster presentation delivered at the 16th Biennial Conference of the Asian Association of Social Psychology, held in July 2025.