The effects of negative mentoring experiences on mentor creativity: The roles of mentor ego depletion and traditionality

The literature often focuses on the positive effects of mentoring, especially for protégés. When mentoring experiences are negative, the assumption is that these negative effects are less detrimental to the mentor than to the protégé, owing to the mentor's greater relative power in the relationship. This study uses ego depletion theory to examine the link between negative mentoring experiences (as rated by protégés) and mentor creativity and focuses on the mediating and moderating roles of mentor ego depletion and mentor traditionality on this link. The results are based on data from 227 protégés, 187 mentors, and 187 supervisors of mentors in Chinese organizations. The findings support a mediating effect of ego depletion on the negative relationship between negative mentoring experiences and mentor creativity. In addition, the study finds that traditionality attenuates both the positive relationship between negative mentoring experiences and ego depletion and the indirect effect of negative mentoring experiences on mentors' creativity through ego depletion. The implications for management theory and practice are discussed.

. Evidence from various studies indicates that a mentor's creativity is positively related to their protégé's creative development (Simonton, 1975). Thus, organizations are likely to select creative employees to become mentors (Allen, Smith, et al., 2009;Allen, Finkelstein, & Poteet, 2009) and to rely on such mentors to benefit junior employees (Gentry, Weber, & Sadri, 2008).
This reliance is somewhat risky, because many mentoring programs are less effective than expected. The mentoring literature has reported on the negative mentoring experiences of both protégés and mentors (Eby, Butts, Lockwood, & Simon, 2004;Kram, 1985;Scandura, 1998). Negative mentoring experiences can have harmful effects on both protégés and mentors and are often described as discordant, disappointing, or dissatisfying by one or both parties of the mentoring relationship (Eby & McManus, 2004).
Researchers have provided evidence that the negative mentoring experiences of protégés' are significant predictors of their stress (Kim & Choi, 2011), depression , psychological withdrawal, intention to leave the mentoring relationship (Eby, Butts, Durley, & Ragins, 2010), relationship quality with the mentor (Burk & Eby, 2010), turnover intention, and job satisfaction (Eby & Allen, 2002). In addition, mentors' negative mentoring experiences are significant predictors of protégés' intention to leave the mentoring relationship and protégés' reports of relational quality and mentoring received (Eby, Durley, Evans, & Ragins, 2008). Such experiences are also predictors of mentors' burnout (Burk & Eby, 2010;Eby et al., 2008), relationship quality with the protégé , unwillingness to mentor (Burk & Eby, 2010), intention to remain in the mentoring relationship , and work-family conflict (Yi, Kwan, Hu, & Chen, 2017). Nevertheless, to the best of our knowledge, no research has been conducted on the detrimental effects of protégés' negative mentoring experiences on their mentors' creativity.
This paucity of research probably results from the common assumption that mentors hold the power in mentoring relationships owing to their relatively high status and the valued resources at their disposal, and that they are less likely to be considered as victims of negative mentoring relationships . This assumption is somewhat prejudiced, because victimization can be reciprocal in a relational context (Lian, Ferris, Morrison, & Brown, 2014). Given that the mentor's creativity is one of the critical predictors of a protégé's creativity (Koseoglu et al., 2017;Simonton, 1975), and that enhancing employee creativity is one of the main purposes for establishing a mentoring program (Liu et al., 2015), the first aim of this study is to explore the relation between protégés' negative mentoring experiences and their mentors' creativity.
Researchers in the area of mentoring have long argued that the nexus between negative mentoring experiences and mentor consequences should be investigated in its entirety . Such investigation should include examining the mediators of those consequences, because only such an approach can untangle the complexities of mentoring-related effects (Yi et al., 2017). To achieve this approach, we use ego depletion to explain the mechanism underlying the relationship between negative mentoring experiences and the mentor's creativity. Ego depletion is "the condition that arises when the self's resources have been expended and the self is temporarily operating at less than full power" (Baumeister, 2002, p. 133). The concept of ego depletion is generally used to explain how negative experiences in one task can result in negative outcomes in another task (Hagger, Wood, Stiff, & Chatzisarantis, 2010;Inzlicht & Kang, 2010;Lin, Ma, & Johnson, 2016;Mackey, Huang, & He, 2020;Methot, Lepine, Podsakoff, & Christian, 2016;Yam, Fehr, Keng-Highberger, Klotz, & Reynolds, 2016). Ego depletion theory suggests that every act of regulation or volition may consume the self's limited selfregulatory resources and deplete the self, leaving them with fewer volitional resources to perform subsequent tasks (Baumeister, 2002).
Accordingly, we propose that mentor ego depletion is a mediating mechanism that explains why and how a protégé's negative mentoring experiences may depress a mentor's creativity. In other words, we suggest that protégés who have negative mentoring experiences are more likely to express their negative moods (e.g.,  and attitudes (e.g., Eby & Allen, 2002) and to exhibit negative behavior (e.g., Topa & Perez-Larrazabal, 2016), and that these reactions can cause conflict, stress, and strain in a mentoring relationship (Lambert et al., 2004). The mentor may spend significant self-regulatory resources to cope with these negative experiences, which can result in ego depletion. In that case, ego-depleted mentors may lack the motivation and ability to regulate themselves, disengage themselves from creative processes, and therefore display a reduced level of creativity.
Although we contend that negative mentoring experiences have detrimental effects, we also argue that not all mentors are affected by inharmonious relationships to the same extent. Scholars in the field of mentoring have acknowledged that personal values play a critical role in mentoring relationships (Ragins, 1995;Zhou, Lapointe, & Zhou, 2019). Hagger (2015) noted that personal values may affect an individual's perceptions and coping responses to workplace stressors. Based on these arguments, we propose that mentor traditionality, which typically involves the endorsement of hierarchical relationships and obedience to authority, may alleviate the detrimental effects of protégés' negative mentoring experiences. We suggest that traditionality may help mentors to feel less concerned about their protégés' negative mentoring experiences or their negative moods and attitudes. Consequently, traditional mentors are less likely to feel depleted by such experiences.
This study makes four major contributions to the literature on mentoring, creativity, ego depletion, and traditionality. First, we investigate the relationship between the protégé's negative mentoring experiences and their mentor's creativity. This investigation answers the call by  for more research on the relation between negative mentoring experiences and work behavior-related outcomes that they were unable to examine in their studies. For researchers and practitioners, it also highlights that protégés' negative mentoring experiences are more toxic and harmful to mentors than previously believed. Second, we use ego depletion theory to link the effects of negative mentoring experience with mentors' levels of creativity. This study indicates that mentors may deplete their self-regulatory resources to cope with protégés who have negative mentoring experiences, and that this depletion of essential resources can reduce the mentor's creativity. In this way, our study offers a novel perspective on how a protégé's negative mentoring experiences can affect the mentor's outcomes and provides steps to mitigate the drawbacks for mentors who experience negative mentorships. Third, we contribute to the creativity literature by identifying additional antecedents of employee creativity (Shalley & Gilson, 2004). Considering the important roles of mentoring and self-regulatory resources in creative processes, this study illuminates a new way through which organizations can affect employees' creativity. Finally, we examine the moderating effects of mentor traditionality, thereby enhancing our understanding of the role traditionality plays in social dynamics at work, especially for high-status people.

| The depleting effects of negative mentoring experiences
Negative mentoring experiences refer to both specific incidents and the characteristics and behaviors of mentors and protégés that reduce mentoring effectiveness . Examples include neglect, exploitation, sabotage, bad attitudes, and mentorprotégé conflicts, as well as mismatch within the dyad Eby et al., 2008). Past studies have shown that mentors' negative mentoring experiences influence the outcomes of both the mentors themselves and their protégés (Eby et al., 2008).
Several studies have also shown that protégés' negative mentoring experiences can have harmful effects on the protégés themselves (Burk & Eby, 2010;. However, few studies have examined the association between protégés' negative mentoring experiences and mentor outcomes. As mentioned previously, one possible reason for this oversight is that mentor-protégé relationships involve power differences . The literature has mainly supported the perspective that protégés are a disadvantaged group and are more likely than mentors to become victims in mentoring relationships. Eby et al. (2010) explained this assumption as follows: Because of their greater relative status and power, mentors are in a position to do considerable damage to protégés, both personally and professionally … In contrast, although bad experiences with protégés may be unpleasant and cause some distress for mentors, the net effect of bad experiences may not be as great since protégés do not control resources or rewards that are valued by mentors (e.g., work assignments, pay, and promotions) (p. 86).
However, from the perspective of ego depletion theory, studies have indicated that relational interactions, especially dysfunctional interactions, create resource loss, regardless of power and status (Eby, 2007;Lian et al., 2014;Mackey et al., 2020). Although protégés may perceive their mentoring experience as negative while their mentors may not, the protégés' perceptions of negative mentoring experiences can signal dysfunctionality in the mentoring relationship (Eby et al., 2008).
We suggest that a mentor's ego depletion is a proximal outcome variable resulting from their protégé's negative mentoring experiences. The concept of ego depletion can explain why people measure costs and benefits according to reciprocity and can show how the results of trade-offs influence an individual's attitudes and behaviors (Lian, Yam, Ferris, & Brown, 2017;McAllister, Mackey, & Perrewé, 2018). As discussed previously, ego depletion theory proposes that every volitional act (e.g., making choices, regulating emotions, and initiating or inhibiting behavior) consumes an individual's limited resources for self-regulation (Baumeister, 2002). A person's self-regulatory resources cannot recover quickly after they are exhausted (Halbesleben, 2006). Furthermore, repeated exhaustion results in a chronic lack of resources (Baumeister, 2002;Liu, Kwan, & Zhang, 2020;Yam et al., 2016), which is termed ego depletion (Baumeister, Bratslavsky, Muraven, & Tice, 1998).
Maintaining mentoring relationships with protégés in the midst of negative mentoring experiences can easily become resource-depleting, because such protégés are more likely to express negative moods (e.g., , attitudes (e.g., Eby & Allen, 2002), and behavior (e.g., Topa & Perez-Larrazabal, 2016) during their interactions with their mentors. On perceiving their protégés' negative reactions, mentors must invest their self-regulatory resources to restrain themselves from striking back, and instead seek to respond in constructive ways.

Mentor Ego Depletion
Mentor Creativity

Protégé Negative
Mentoring Experiences Mentor Traditionality + -F I G U R E 1 The conceptual model of this research For example, mentors need to use their self-regulatory resources to interpret, seek reasons, and find solutions for protégés' negative reactions, and they are expected to regulate their emotions and adjust their own behavior (Compas, Connor-Smith, Saltzman, Thomsen, & Wadsworth, 2001). All of these activities demand the expenditure of self-regulatory resources (Inzlicht & Kang, 2010). Over time, such expenditure may have long-lasting effects on mentors' daily work, which can lead to ego depletion (Mackey et al., 2020;Yam et al., 2016).  classified the issues contributing to a protégés' negative mentoring experiences into five main types: (a) mismatches within the dyad, (b) distancing behavior, (c) manipulative behavior, (d) lack of mentor expertise, and (e) general dysfunctionality. These five issues provide a detailed framework for understanding how protégés' negative mentoring experiences can negatively affect mentors. The classification of issues suggests three main reasons for a negative effect. First, a dyadic mismatch may result in relational conflicts and communication problems (Scandura, 1998), which impose additional demands on a mentor's finite self-regulatory resources. A mentor must first overcome his or her differences with the protégé and then offer the protégé adequate developmental support. Second, protégés may express a series of negative emotional, attitudinal, and behavioral responses when they sense distance or manipulation from their mentors. For example, they may demonstrate their anger and discontent (Eby & Allen, 2002), indicate that the relationship is unfair , decline to incorporate their mentors' instructions into their work , or engage in counterproductive or vengeful behavior (Topa & Perez-Larrazabal, 2016). Faced with such circumstances, mentors may feel anxious and apprehensive. They may worry about being reprimanded for failing to meet the organization's mentoring norms. They may also feel a need to resolve their protégés' negative reactions and to regulate their own behavior. All of these issues can tax mentors' internal resources. Third, if protégés repeatedly express dissatisfaction with their mentors' lack of expertise or personal problems, the mentors' resources are likely to be further depleted. The mentors are then obliged to perform their work duties while regulating their behavior, improving their expertise, and coping with their personal problems to meet their protégés' expectations.
In summary, negative mentoring experiences tend to deplete a mentor's personal resources. This consumption of limited resources eventually leads to a condition called ego depletion. Thus, the following hypothesis is proposed.
Hypothesis 1. Protégés' negative mentoring experiences are positively related to their mentors' ego depletion.

| Implications for mentors' creativity
Research has shown that ego depletion affects employee behavior (Byrne et al., 2014;Lin et al., 2016;Mackey et al., 2020). As we describe next, one particular way that the depletion caused by negative mentoring experiences may affect mentors is in their ability to be creative. Amabile's (1996) componential model of creativity suggests that creativity is generated in several crucial stages. These stages include discovering a problem or an opportunity, collecting and organizing the necessary resources or information, and then generating, filtering, and modifying ideas for application. These steps require a significant investment of self-regulatory resources (Shalley, Zhou, & Oldham, 2004). For example, in the initial step of the creative process (discovering problems or opportunities), employees must have enough energy to be curious and passionate about their work. In the second stage, employees must acquire or reactivate a store of knowledge that is relevant to a problem or opportunity. This stage usually involves a considerable amount of learning (Amabile, 1983). The effort to accumulate knowledge and perform learning tasks may consume large amounts of self-regulatory resources. Emotional and cognitive resources also figure prominently in the third stage (generating and modifying ideas) (Amabile, 1983). Creativity typically involves a long journey of trial and error (George, 2007). Each attempt at a new approach or solution requires investment of resources .
Given the requirements of resources for on-the-job creativity, ego depletion (or a state of diminished personal resources) is likely to reduce an employee's capacity for further resource expenditure (Baumeister, Vohs, & Tice, 2007). The depletion of personal resources may influence a mentor's creativity in two ways: by reducing his/her ability or by reducing his/her willingness (Han, Harms, & Bai, 2017). First, mentors who lack self-regulatory resources can seldom focus on contributing beyond their basic duties (Baumeister et al., 1998). They become less able to find creative solutions for work-related problems.
Likewise, searching for information and experimenting with new ideas are resource-draining tasks. Ego-depleted mentors seldom have the extra resources to pursue such work. Hence, ego depletion impairs a mentor's ability to engage in the creative process, which negatively affects their creativity.
Second, Baumeister, Muraven, and Tice (2000) suggested that a lack of resources activates an individual's instinct to conserve remaining resources and undermines their willingness to perform well in subsequent tasks, especially when the employee believes that those tasks are not formally required. Compared with other forms of work performance (e.g., task performance), an employee's creativity is more susceptible to negative conditions because it requires a high level of investment and is associated with a high degree of uncertainty (Zhang, Liao, Li, & Colbert, 2020). As creativity is an optional endeavor that is not formally required by organizations, employees are free to reduce the time and effort they spend on creativity . Therefore, ego-depleted mentors commonly conserve their resources by declining to exert themselves in creative tasks (Byrne et al., 2014). Once their resource conservation instincts have been activated, mentors are often unwilling to undertake any extra tasks that may be related to the creative process (Baumeister et al., 2007), such as identifying problems or searching for solutions.
Therefore, we suggest that ego-depletion leads mentors to disengage from creative work, due to their instinct for protecting their reduced self-regulatory resources. Previous research has indicated that resource-depleted employees are less likely to be creative (Butler, Little, Walter, & Phillips, 2015;Liu et al., 2020;Price & Yates, 2015).
In summarizing these points, we propose the second hypothesis.

Hypothesis 2. Mentors' ego depletion is negatively related to their creativity.
Given the above-described arguments, we suggest that a mentor's ego depletion is negatively related to his/her creativity, and that this state of ego depletion is often caused by negative mentoring experiences. We consider an integrated framework and argue that mentors must expend additional resources on self-regulation when their mentoring relationships turn negative if they are to deal with the conflicts and incidents involved. A mentors' ego depletion ultimately thwarts his or her ability and willingness to engage in creative work.
Ego depletion, therefore, is expected to mediate the relationship between negative mentoring experiences and mentor creativity.
Accordingly, we propose the third hypothesis.

Hypothesis 3. Ego depletion mediates the negative relationship between protégés' negative mentoring experiences and mentors'
creativity.

| The moderating role of traditionality
Traditionality is a personal or social tendency to be profoundly influenced by traditional culture (Farh, Earley, & Lin, 1997). Such culture typically ranks ruler over subject, father over child, and senior over junior (Hui, Lee, & Rousseau, 2004). Although these relationships are characterized by hierarchical differences, parent-child relationships are more intimate, owing to consanguinity and the sense of family responsibility (Jankowiak, 1992). Under the influence of traditional culture, the mentor-protégé relationship is similar to the relationship between a teacher and a student or between a parent and a child (Zhou et al., 2019). Traditionalists typically embrace the belief that the mentor-protégé relationship should be both hierarchical and intimate (Bozionelos & Wang, 2006;Zhou et al., 2019). On the one hand, traditional mentors consider authoritative behavior by the superior in a relationship to be reasonable (Spreitzer, Perttula, & Xin, 2005). Traditional mentors are often formal when interacting with those they consider inferior (Farh, Hackett, & Liang, 2007). On the other hand, highly traditional mentors typically regard their protégés as their "children," treat them with benevolence, and trust them inwardly (Wang & Kim, 2013;Zhou et al., 2019).
Accordingly, we propose that a mentor's level of traditionality may reduce the degree to which their protégés' negative mentoring experiences can cause ego depletion in two ways. First, previous studies have shown that when protégés have negative mentoring experiences, they may express negative moods (e.g., , attitudes (e.g., Eby & Allen, 2002), or behaviors (e.g., Topa & Perez-Larrazabal, 2016) during their interactions with their mentors. However, protégés with highly traditional mentors are less likely to show their dissatisfaction. Traditional mentors display authoritative behavior in their interactions with their protégés (Spreitzer et al., 2005), and their protégés seldom dare to express negative moods or behavior in return.
Second, mentors with traditional values are inclined to treat their protégés as their children (Zhou et al., 2019). Therefore, the mentorprotégé relationship is more intimate than the relationship between a supervisor and a subordinate. Traditional mentors are typically generous and patient when their protégés display negative moods, attitudes, or behaviors (Zhou et al., 2019). They do not regard their protégés' negative reactions as an affront. Thus, traditional mentors are less likely to experience stress or to expend resources on coping when their protégés express dissatisfaction. In short, highly traditional mentors are not troubled by protégés' negative mentoring experiences or reactions and are less likely to experience ego depletion when coping with such issues.
Although traditionality is more common in Asian countries, research has shown that it plays an important role in other regions (Spreitzer et al., 2005). In Western organizations, conventional mentors, who tend to be outwardly authoritative with their protégés but inwardly generous and trusting, may also experience less resource depletion when dealing with their protégés' negative mentoring experiences. Thus, we propose a fourth hypothesis.

Hypothesis 4. A mentor's level of traditionality moderates the relation-
ship between a protégé's negative mentoring experiences and the mentor's ego depletion, such that this relationship is weaker when the level of traditionality is higher.
As previously proposed, protégés' negative mentoring experiences commonly lead to their mentors' ego depletion, which impairs the mentors' creativity. Mentor traditionality moderates the relationship between the protégés' negative mentoring experiences and the mentors' ego depletion. In other words, mentor ego depletion plays a mediating role, which is moderated by the mentor's traditionality. The self-regulatory resources of mentors with high levels of traditionality are less likely to be affected by negative mentoring relations. Therefore, the negativity created in such a relationship has less effect on a traditional mentor's creativity. We propose a fifth hypothesis as follows.

Hypothesis 5. Mentor traditionality moderates the indirect relationships
between negative mentoring experiences and mentors' creativity through ego depletion, such that the indirect effect is weaker when mentor traditionality is high.

| METHODS
To test our theoretical model (Figure 1), we collected field survey data from protégés, mentors, and supervisors of mentors at three points in time separated by 1-month intervals. Specifically, protégés evaluated their negative mentoring experiences at Time 1, mentors evaluated ego depletion and traditionality at Time 2, and mentors' supervisors evaluated mentors' creativity at Time 3. This approach is consistent with the contention that multiple-source and multiple-time data can alleviate concerns about common method bias (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003).

| Samples and procedures
Our data were collected from three construction industry organiza- Some of our participants worked on construction projects in the countryside, which made it difficult to distribute paper-based questionnaires face to face. Hence, we chose to use electronic questionnaires for these participants. The designated numbers on the paper-based questionnaires were hidden for the sake of anonymity. With assistance from the HR managers, the questionnaires with hidden numbers were assigned to the corresponding participants. After the questionnaires were completed, the participants were asked to seal them in blank envelopes and return them to us directly.
We used the SO JUMP program to collect the data from the electronic questionnaires. SO JUMP is a convenient survey tool that is commonly used in China. This program can generate a series of random passwords for the respondents to online questionnaires. In this study, we assigned each participant a unique password that they would use to access their questionnaires. Each password was unique and could only be used once. Once the questionnaires were completed, we identified and matched the participants by the passwords assigned to them.
The data collection period lasted 3 months and was conducted in three phases. In Phase 1 (T1), we administered questionnaires to all 413 protégés, asking them to report on any negative mentoring experiences. We received 302 responses. One month after Phase 1, we conducted phase 2 (T2), in which all 341 mentors were asked to report on their levels of traditionality and ego depletion and to pro-

| Measures
The original scales were developed in English. We used a backtranslation process (Brislin, 1986) to ensure their applicability to a Chinese sample. First, the items were translated into a Chinese version by a management professor and two doctoral candidates in management studies. Then, another management professor and two other doctoral candidates independently translated the Chinese version back into English. Finally, one bilingual scholar compared the three versions to determine whether they were comparable.
To further ensure that our survey tools were applicable to the Chinese context, we provided the Chinese version to 10 MBA and EMBA students at a local Chinese university. None of these students identified any measurement items that were inapplicable to the Chinese cultural context or to the organizations concerned.

| Negative mentoring experiences
Negative mentoring experiences were reported by the protégés at T1.
The measures of negative mentoring experiences were developed by  and considered five dimensions of dysfunctionality: (a) mismatch within the dyad (nine items, such as "The personal values of my mentor are different from my own"); (b) distancing behavior (seven items, such as "My mentor is reluctant to talk about things that are important to me"); (c) manipulative behavior (11 items, such as "I am intimidated by my mentor"); (d) lack of mentor expertise (seven items, such as "My mentor lacks expertise in areas that are important for the type of work that he/she does"); and (e) general dysfunctionality (eight items, such as "My mentor has a bad attitude").
The items were rated from 1 (strongly disagree) to 7 (strongly agree).
The Cronbach's alpha for this scale was .89.

| Ego depletion
At T2, ego depletion was reported by the mentors using a five-item scale (Twenge, Muraven, & Tice, 2004;Yam et al., 2016). The mentors were specifically instructed to recall and report on their feelings during the past month. The items included the following: "During the past month, it took a lot of effort to concentrate on something," "During the past month, I felt drained," and "During the past month, I felt like my willpower was gone." These items were rated from 1 (strongly disagree) to 7 (strongly agree). The Cronbach's alpha for this scale was .95.

| Traditionality
Traditionality was also reported by the mentors at T2. The scale used was originally developed by Yang, Yu, and Yeh (1989). We applied the five-item version of this scale, which was developed by Farh et al. (1997) and included the following: "The best way to avoid mistakes is to follow the instructions of senior persons," and "Those who are respected by parents should be respected by their children." The items were rated from 1 (strongly disagree) to 7 (strongly agree). The Cronbach's alpha for this scale was .84.

| Creativity
At T3, the mentors' supervisors evaluated the mentors' creativity. Creativity was measured using a 13-item scale developed by Zhou and George (2001) and modified for use in the Chinese context by Zhang and Bartol (2010). The items included "He/she suggests new ways to achieve goals or objectives." The items were rated from 1 (strongly disagree) to 7 (strongly agree). The Cronbach's alpha for this scale was .98.

| Control variables
We controlled for the mentor's age (years), gender (1 = male, 2 = female), educational level (1 = junior high school or below to 5 = Master's degree or above), position (1 = frontline employees to 5 = CEO), and tenure (years). We included these controls because other researchers have considered these types of information to be significantly related to employee creativity (Gong, Huang, & Farh, 2009;Oldham & Cummings, 1996;Zhang & Bartol, 2010). For example, research has suggested that gender differences in creativity may be explained by differences between men and women in both biological and environmental factors (Abra & Valentine-French, 1991). In addition, position influences creativity through its effect on network centrality (Ibarra, 1993). We also created and controlled for two organization dummy variables, because our data were collected from three different organizations. Some mentors in our sample had more than one protégé. In these cases, we asked each protégé to rate his/her negative mentoring experiences and took the average. Thus, we controlled for the number of protégés to exclude the possible influence of this factor. 1 4 | RESULTS

| Descriptive statistics
The descriptive statistics and correlations of the variables in this study are shown in Table 1.

| Hypotheses testing
We used hierarchical multiple regression and bootstrapping analyses to test our hypotheses. Hypothesis 1 proposed a positive relationship between negative mentoring experiences and mentors' ego depletion. The results of Model 2a in Table 2 show that negative mentoring experiences were positively and significantly related to ego depletion (β = .19, SE = .06, p < .01), which supported Hypothesis 1. Hypothesis 2 proposed a negative relationship between mentors' ego depletion and creativity. The results of Model 6a in Table 2 show that ego depletion was negatively related to creativity (β = À.35, SE = .09, p < .001), which supported Hypothesis 2. Our results without controls were also reported to demonstrate the robustness of our model. By comparing the results of Models 2a and 2b and the results of Models 6a and 6b in Table 2, we observed that the findings did not change significantly with or without the control variables.
We used a path analytic approach with Mplus 8 to test the mediating effect of mentors' ego depletion that was proposed in Hypothesis 3 (Muthen & Muthen, 2012). Specifically, the bootstrapping process was conducted with 20,000 iterations to compute a confidence interval around the indirect effect. The results indicated that the indirect effect of negative mentoring experiences on mentors' creativity through ego depletion was significant (β = À.05, SE = .03, 95% CI [À.12, À.02], excluding zero). Thus, Hypothesis 3 was supported.
Hypothesis 4 proposed that mentor traditionality moderates the relationship between negative mentoring experiences and mentors' ego depletion. The results of Model 3a in Table 2 show that the interaction of negative mentoring experiences and traditionality was significantly and negatively related to ego depletion (β = À.12, SE = .05, p < .01). We conducted simple slope tests (Aiken & West, 1991) to further confirm this moderating effect. As shown in Figure 2, at a low level of traditionality (1 SD below the mean), negative mentoring experiences were positively related to ego depletion (β = .38, SE = .12, p < .01). At a high level of traditionality (1 SD above the mean), the relationship between negative mentoring experiences and ego depletion was nonsignificant (β = .07, SE = .07, n.s.). Therefore, Hypothesis 4 was supported.
Hypothesis 5 proposed that mentor traditionality would moderate the indirect effect of negative mentoring experiences and creativity through ego depletion (a moderated mediation model). We tested

| Supplemental analyses (relative importance analyses of the dimensions of the Protégés' negative mentoring experiences)
This study adopted an additive approach to measure the protégés' negative mentoring experiences by assuming that all five forms of negative mentoring experiences had identical effects on mentors' ego depletion and creativity. This assumption presented one possibility for how protégés' negative mentoring experiences could influence mentors, but past research has provided evidence that different forms of negative mentoring experiences may vary in frequency and severity (Eby, 2007). Thus, the five types of negative mentoring experiences could heterogeneously affect mentors' outcomes.
One potential contribution of this study is its focus on those dimensions of mentoring experience that are particularly likely to cause ego depletion among mentors (in conjunction with traditionality). To test the relative weight effects of the five dimensions of negative mentoring experiences on mentors' ego depletion and creativity, we used the epsilon statistic, which is a common Note: N = 187. *p < .05; **p < .01. Two-tailed tests. method used to determine the relative contribution made by each predictor when accounting for the variance of dependent variables, especially when the predictors are highly correlated (Babalola et al., 2021;Johnson, 2000). In this method, the percentages attributable to each variable (%R 2 ) were percentages of the variance explained in the outcome by the set. Predictors with higher relative weights represented greater increases in the R 2 values that were explained by such predictors.
The results of Models 1 and 3 in Table 3  have a protégé who felt that the mentor lacked expertise than to have a protégé who felt that the mentor wielded power in a tyrannical manner, or a protégé who felt that the mentor intentionally excluded him or her.
The above-described findings were not completely consistent with the findings of the regression analysis in Table 3. Although lack T A B L E 2 Results of hierarchical regression analysis with ego depletion and creativity of expertise had the highest rescaled relative weight with ego depletion (27.29%), the regression coefficient of lack of expertise was lower than that of manipulative behavior or general dysfunctionality. Such inconsistency demonstrated that regression coefficients may provide misleading estimates of relative importance when there is more than one predictor, especially when the predictors are moderately or highly correlated with each other (Dalal, Baysinger, Brummel, & LeBreton, 2012;LeBreton, Ployhart, & Ladd, 2004). Thus, the results of the relative-weight analysis offered a more accurate estimation of relative importance than the results of the regression analysis.

| DISCUSSION
Although research has provided evidence that negative mentoring experiences harm both protégés and mentors Eby et al., 2008;Eby & Allen, 2002;Kim & Choi, 2011;Yi et al., 2017), little research has been done on how or when protégés' negative mentoring experiences affect mentors' outcomes, especially in terms of the mentors' creativity. Based on ego depletion theory, this study proposed a first-stage moderated mediation model to explain how and when protégés' negative mentoring experiences impair mentors' creativity. Our findings indicate that protégés' negative mentoring experiences are positively related to mentors' ego depletion, and the state of ego depletion is negatively related to mentors' creativity. Furthermore, mentors' ego depletion mediates the relationship between protégés' negative mentoring experiences and mentor creativity. Considering the culture-specific aspects of mentoring research, we find that the effects of negative mentoring experiences on mentors' ego depletion are weaker for mentors with elevated levels of traditionality.
We discuss the theoretical and practical implications below.

| Theoretical implications
This study makes several theoretical contributions to research on mentoring and creativity. First, it contributes to the mentoring literature by establishing a theoretical model that accounts for the effects of negative mentoring experiences on mentors' activities associated with creativity and shows that mentoring is very important to organizational effectiveness. To enhance their employees' creativity, organizations establish and maintain mentoring relationships by selecting high performers as mentors (Liu et al., 2015). However, this approach may backfire, as mentoring relationships can be detrimental to mentors' creativity if negative mentoring experiences occur. Unlike previous studies that have focused on protégés' outcomes, this study shifts the focus to mentors' creativity. In doing so, we challenge the assumption that mentors are less likely to be affected by negative experiences than their protégés owing to their higher levels of rank and power (e.g., Eby et al., 2010). The findings of this study highlight the harmful effects of negative mentoring experiences on all parties involved and demonstrate the importance of solving this problem. In addition, we indicate that mentors' creativity is a critical predictor of protégés' creativity (Simonton, 1975). As aggregated employee creativity has a positive relationship with firm innovation  creativity and the first to analyze the mechanism underlying this link.
Our findings direct researchers to further explore the mediating role of mentors' ego depletion in future studies on negative mentoring experiences. We find that ego depletion reduces a mentor's willingness and ability to participate in other job tasks. Future research could also investigate the connections between negative mentoring experiences and other types of in-role (e.g., work performance), and extrarole (e.g., OCB and proactive behavior) outcomes that arise owing to resource depletion.
Although experimental studies have proposed that the effects of ego depletion are overestimated owing to publication bias (e.g., Xu et al., 2014), we suggest that these studies' null results were likely caused by their use of dual-task experiments rather than surveys for data collection. Specifically, Baumeister et al. (2000) indicated that individuals who report ego depletion are not fully depleted of their resources. They can perform well in other tasks if necessary. However, if they do not value those tasks, they tend to conserve their diminished resources. Participants in dual-task experiments were asked to fulfill certain tasks and were likely to regard those tasks as valuable. Therefore, they did not conserve their limited resources and did their best to finish the tasks. In our study, creative behavior was not considered compulsory . Thus, the mentors were less likely to devote additional effort for creativity when their resources were diminished.
Third, this study supplements the literature on creativity by exploring the relationship between ego depletion and creativity and extends our understanding of the critical role that self-regulatory resources play in creative processes. Previous studies have regarded creativity as an extra-role behavior and have acknowledged that employees have the freedom to adjust their engagement in creativity (Amabile, 1996). Employees are willing to invest resources on creative processes only when they have sufficient resources or when they feel it is necessary to do so. As creative behavior entails investment of resources (Kiazad, Seibert, & Kraimer, 2014) and a high level of risk (Dewett, 2006)

| Practical implications
This study's findings provide three practical implications for organizations. First, although mentoring is considered an effective method for improving employee development (Eby et al., 2013), the detrimental effects of formal mentoring programs, particularly on mentors (Ragins & Cotton, 1999), should also be taken seriously (Chen, Wen, & Hu, 2017). Our findings show that negative evaluations of mentor-protégé relationships by protégés can negatively affect the creativity of mentors. Thus, we suggest that organizations should consider dissolving negative mentoring relationships.
Organizations should provide both mentors and protégés with opportunities to gain insights about each other and encourage them to choose each other based on mutual attraction. Giving the participants a choice in forming these relationships would help to avoid or minimize mentor-protégé mismatches .
In addition, the results of our supplementary analyses suggest that among the five dimensions of negative mentoring experiences, criticism regarding lack of expertise had the greatest effect on mentors. Accordingly, training is needed to improve mentors' expertise and ensure that both mentors and protégés have appropriate expectations of each other (Hu, Wang, Kwan, & Yi, 2021).
Our findings also indicate that ego depletion is a barrier to creativity. Thus, organizations should take proper measures to protect or replenish their mentors' personal resources as a means to enhance their employees' creativity. For example, organizations should provide mentors with support and frequent communication to help them handle negative moods or misconceptions (Lin et al., 2016). Likewise, selfaffirmation training may offer an effective way to enhance mentors' abilities to exercise self-control and to quickly recover from ego depletion (Yam et al., 2016

| Limitations and suggestions for future research
Despite the theoretical and practical contributions mentioned above, this study has several limitations. First, negative mentoring experiences were rated only by the protégés. Although this approach can help to alleviate concerns about common method bias (Podsakoff et al., 2003), we acknowledge that ratings may vary between protégés and mentors. Future studies should obtain assessments from both mentors and protégés to increase the accuracy and reliability of the measures used (Schriesheim, Neider, & Scandura, 1998). Second, the study did not adequately address the concern that ego depletion may lead to negative mentoring experiences. Our time-lagged research design did not provide solid evidence of the causality involved (Law, Wong, Yan, & Huang, 2016 (Farh et al., 2007), could function as a buffer for the destructive effects of negative mentoring experiences. We encourage future studies to investigate the roles of other cultural variables in moderating the destructive effects of negative mentoring.
This study provides a springboard for future research on negative mentoring. Other scholars have concluded that mentoring relationships are dynamic, as the relational processes and consequences related to mentoring tend to change over time (Allen, Eby, Chao, & Bauer, 2017). Future research should test the underlying mechanisms proposed in this study to better capture these dynamic processes across time, preferably using a longitudinal, dyadic research design. Various scholars have also called for future research to examine the crossover effects of changes in emotions, attitudes, and behaviors between partners in close relationships (Lyons & Sayer, 2005). A previous longitudinal study (Chun, Sosik, & Yun, 2012) indicated that such crossover effects influence multiple long-term outcomes (e.g., affective well-being and organizational commitment) of both mentors and protégés. By considering different stages of mentorship, future research could examine changes in experiences, reactions, and feelings in the context of negative mentoring. More specifically, a longitudinal, dyadic research design could help us to understand the dynamics involved in the relationships between protégés' negative mentoring experiences and negative reactions and mentors' coping strategies, thereby outlining the dynamic processes leading to mentors' ego depletion and poor performance.

| Conclusion
Although the literature has examined the consequences of negative mentoring experiences, most studies have focused on the detrimental self-effects of mentors' or protégés' negative mentoring experiences or the effects of mentors' negative mentoring experiences on protégés.
This may be due to the usual conceptualization of mentors as "perpetrators" in a mentorship owing to their greater relative status and power. However, the results of our study reveal that protégés' negative mentoring experiences also have effects on mentors' outcomes. On the basis of ego depletion theory, we developed a moderated mediation model, and our tests of that model revealed several important patterns.
The findings showed that protégés' negative mentoring experiences had a negative effect on mentors' creativity, and that mentors' ego depletion was a crucial mediator of this relationship. Furthermore, mentor traditionality played a pivotal moderating role in this relationship. In summary, this study offers a springboard for future research and practice related to negative mentoring experiences.

ACKNOWLEDGMENTS
This study was funded by the National Natural Science Foundation of  .

ENDNOTES
1 Although we tested all hypotheses with controls for a more conservative test of our predictions, we also tested all hypotheses without controls. The results indicated that the pattern and significance of our findings for all of the hypotheses were equivalent with and without controls.
2 Although most of our data were drawn from different sources, we reported the CFA results of the study's four main variables here to provide readers with insight into the discriminant validity of the key variables. The CFA results suggested that the hypothesized four-factor model (χ 2 = 950.30, df = 344, p < .001, CFI = .92, TLI = .92, RMSEA = .09, SRMR = .05) fit the data better than the single-factor model (χ 2 = 3,596.12, df = 350, p < .001, CFI = .59, TLI = .56, RMSEA = .20; SRMR = .18) and that this four-factor model was better than all of the other alternative models, including the four types of three-factor models.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.