Linking perceived overqualification to work withdrawal, employee silence, and pro-job unethical behavior in a Chinese context: the mediating roles of shame and anger

Abstract

Using appraisal theories of emotions as the theoretical basis, we delineate how perceived overqualification relates to work withdrawal, employee silence, and pro-job unethical behavior through the mediating effects of discrete emotions (i.e., shame and anger). We suggest that perceived overqualification is positively associated with shame and ultimately work withdrawal and employee silence, and it has a positive effect on pro-job unethical behavior through the mediating role of anger as well as the serial mediating effect of shame and anger. Data from a three-wave, time-lagged survey of 225 full-time employees in China, provides support for our theoretical model and hypotheses. Taken together, our results suggest that discrete emotional states can help to interpret how and why overqualified employees exhibit distinct action tendencies. We discuss the theoretical and practical implications of our findings.

Introduction

As COVID-19 exacerbates the continued downturn of the global economy and the lagging recovery of the slowing labor market (the UN’s World Economic Situation and Prospects, 2022), the widespread unemployment caused by layoffs often puts individuals into serious economic difficulties. Many job applicants have to compete for jobs that traditionally do not require so much education and skills with higher educational credentials, and accept jobs that require less education, experience, and abilities than they possess (Liu et al. 2015). Consequently, a growing number of employees experience overqualification, which denotes “the situation where individuals have qualifications such as education and skills that exceed job requirements” (Erdogan and Bauer 2009, p. 557). Overqualification is generally an assessment that comes from subjective social comparison and should be better measured by individual perceptions (Liu et al. 2015; Liu and Wang 2012). Furthermore, McKee-Ryan et al. (2009) indicated that perceived overqualification is highly correlated with objectively measured overqualification. Consistent with most of the previous studies (e.g., Erdogan and Bauer 2009; Liu et al. 2015; Maynard and Parfyonova 2013), we focus on employees’ perceived overqualification (i.e., a self-assessment) rather than objective overqualification, in the present study.

A considerable amount of research has established that perceived overqualification elicits a range of adverse individual behavioral outcomes, such as voluntary turnover (Erdogan and Bauer 2009; Maynard and Parfyonova 2013), cyberloafing (Cheng et al. 2020; Khan et al. 2022), and counterproductive work behavior (CWB, Liu et al. 2015; Luksyte et al. 2011). Researchers’ efforts have then shifted to understanding the psychological processes underlying these connections. To interpret why an overqualified employee exhibits a specific behavior, the existing literature examines mediators that are either tightly associated with perceptions of overqualification [e.g., person-job fit, Luksyte et al. (2011); relative deprivation, Erdogan et al. (2018) and Lee et al. (2021)], or developing as time goes on [e.g., self-esteem based on the organization, Liu et al. (2015); harmonious passion, Cheng et al. (2020)]. However, these mediators developed based on a single type of behavior fail to explain why and how overqualified employees may exhibit distinct behavioral tendencies. For example, they cannot explain why overqualified employees may exhibit both withdrawal behaviors (e.g., voluntary turnover and cyberloafing) and aggression behaviors (e.g., CWB) at work. In sum, focusing only on a specific behavior may lead to an incomplete or inaccurate understanding of the action tendencies of perceived overqualification.

We draw on the cognitive appraisal theories of emotions (Ellsworth and Scherer 2003) to understand individuals’ distinct emotional and behavioral reactions to perceived overqualification. Perceived overqualification is at least involved with two aspects in Ellsworth and Scherer’s (2003) theoretical framework—motivational bases and social dimensions. We argue that different emotional states (e.g., anger and shame) may be responsible for different actions related to these two aspects. In terms of motivational bases, overqualified employees often perceive their poor person-job fit as an unfavorable or even a frustrating work experience that thwarts their personal goals (Liu et al. 2015), which is likely to elicit anger (Berkowitz and Harmon-Jones 2004). Although a few studies have recognized the mediating role of anger toward employment situations (e.g., Liu and Wang 2012; Liu et al. 2015), such research only pays attention to anger that is directed toward the organization and regarded as retaliatory. According to the appraisal-tendency proposal (Lerner and Keltner 2000, 2001), overqualified employees may evolve into a hostile tendency toward other individuals in the organization. Although employees may believe that the organization or supervisor is responsible for the undesirable and unsatisfactory underemployment situation (Eisenberger et al. 2010), they normally are less inclined to take direct and obvious actions against the organization or supervisor who controls the resources they value (Yukl 2010). As a result, they may conduct deviant behaviors directed against their coworkers. Therefore, we posit that overqualified employees experiencing anger would engage in pro-job unethical behavior, which is defined as “actions that are intended to strengthen an employee’s bond with the job at the cost of violating standards of proper conduct” (Wang et al. 2019), directed against coworkers as an aggressive response. However, few studies in the literature have examined this relationship.

In terms of the social dimensions, other reactive emotions, such as shame, related to perceived overqualification have been largely ignored in the literature. Specifically, overqualified employees may perceive the undesirable underemployment situation as having negative implications for social dimensions, including their self-worth, self-image, and social identity. Such negative self-evaluations lead them to experience shame (Brown and Weiner 1984; Gilbert 1998; Gruenewald et al. 2004). Moreover, because overqualification indicates that the organization does not value overqualified employees, those employees may also frequently experience other emotions besides anger, such as shame. The withdrawal and avoidance tendencies of shame (Peng et al. 2019) motivate overqualified employees to exhibit work withdrawal and employee silence to escape such an unpleasant feeling.

In sum, drawing on the cognitive appraisal theories of emotions, this study proposes a model highlighting how discrete emotions represent an alternative mechanism that links perceived overqualification to withdrawal and pro-job unethical behavior at work (see Fig. 1). Specifically, shame with withdrawal and avoidance tendencies is associated with work withdrawal and employee silence, while anger as a hostile emotion with a “fight” or an aggressive tendency is associated with pro-job unethical behavior. We contribute to the overqualification literature in three main ways. First, by including both withdrawal and pro-job unethical behavior at work, our research moves the overqualification literature forward by providing a more integrated perspective of different behavioral tendencies rather than a specific behavior related to overqualification. So far, research in this field has mainly explored either withdrawal behaviors (e.g., Cheng et al. 2020; Erdogan and Bauer 2009; Maynard and Parfyonova 2013) or aggressive behaviors (e.g., Liu et al. 2015; Luksyte et al. 2011) as consequence variables. Overqualification scholars consequently lack a complete understanding of different negative behavioral reactions to overqualification, underlining the importance of new theoretical as well as empirical investigations. By introducing appraisal theories of emotions, our research aims to jointly explain overqualified employees’ different negative behavioral tendencies that are scattered in the previous literature.

Fig. 1.

Theoretical model of the current research

Second, we further contribute to the overqualification literature by demonstrating the novel mediating role of discrete emotions in linking overqualification to different behavioral tendencies. While scholars have investigated why overqualified employees engage in a specific behavior (e.g., Cheng et al. 2020; Luksyte et al. 2011), considerably less is known about why overqualified employees exhibit different behavioral tendencies. We posit that, overqualified employees may, over time, experience generalized emotional states that are each related to particular behavioral tendencies (Ellsworth and Scherer 2003). The current research is among the first to provide theoretical and empirical accounts of discrete emotions as the emotional mechanism that links overqualification to a range of dysfunctional behaviors.

Third, this research also complements and broadens the literature on overqualification and emotions by shedding light on the individual emotional change process. Previous studies on either overqualification or emotions have largely focused on the role of various emotions but are rarely involved in the individual emotional change process. We are among the first to pinpoint how the emotional processes of “withdrawal or avoidance” (i.e., shame) and “aggression or hostility” (i.e., anger) serially mediate the positive effect of perceived overqualification on aggression at work. Through exploring the positive serial mediating effect of shame and anger, our investigation paves a new avenue for examining more emotional change processes for overqualified employees.

Theoretical background and hypotheses

Cognitive appraisal theories of emotions

Based on a cognitive perspective, appraisal theories of emotions attempt to interpret determinants of individual emotional responses (Ellsworth and Scherer 2003; Lazarus 1991). Specifically, individual appraisal of an event or a situation, rather than the objective reality of the event or situation itself, determines whether and why individuals experience certain emotions (e.g., Tepper et al. 2018). We similarly indicate that the cognitive appraisal of overqualification and not the objective level of overqualification trigger individual emotional responses. Moreover, extant research in social psychology and organizational behavior has shown that perceptions influence how individuals feel and react to an event or a situation more so than the event or situation per se (e.g., Clarkson et al. 2010; Ng et al. 2019). Consequently, it is more appropriate to employ the perspective of cognitive appraisal to study perceived overqualification rather than objective overqualification.

Researchers usually apply appraisal theories of emotions to elaborate and interpret discrete emotions (e.g., Peng et al. 2019). Appraisal theories of emotions posit that those cognitive appraisals and interpretations of an event or a situation facilitate subsequent emotions which may be distinct (Ellsworth and Scherer 2003; Larsen and McGraw 2011). Drawing on Ellsworth and Scherer (2003), appraisal theories of emotions include four major dimensions: basic stimulus characteristics (e.g., novelty and pleasantness of the focal event or situation), social dimensions (e.g., justice norms and implications for social identity), motivational bases (i.e., the relevance and conduciveness of the event or situation to individuals’ need satisfaction and goal achievement), and power and coping (i.e., perceived controllability of the event or situation and individuals’ ability to cope with or accommodate to the event or situation). Extant research has found that appraisal patterns along these four dimensions tend to engender different or distinct emotional states (Lazarus 1991; Roseman 2013; Smith and Ellsworth 1985). This study adopts Ellsworth and Scherer’s (2003) framework and evaluates the emotional and behavioral outcomes of overqualification from the following two aspects: social dimensions and motivational bases.

Overqualification, as a type of person-job misfit, represents a sustained pattern of an underemployment situation (Erdogan and Bauer 2009; Lin et al. 2017) that is undesirable and frustrating. Hence, according to Ochsner (2000), compared to other relatively short work experiences, overqualified employees likely recapture shame and anger, and keep these undesirable feelings in long episodic memory. Drawing on the major appraisal dimensions outlined by Ellsworth and Scherer (2003), we expect shame and anger to be discrete emotions that are often experienced following perceived overqualification. Furthermore, shame and anger should lead to different motivations and behaviors in response to overqualification as an unpleasant and unfavorable underemployment situation.

Perceived overqualification and withdrawal at work: the mediating effect of shame

Perceived overqualification reflects the extent to which employees consider themselves as possessing more education, experience, or skills than the required job qualifications (Erdogan and Bauer 2009; Johnson and Johnson 1996; Johnson et al. 2002). We argue overqualified employees tend to experience shame toward the undesirable underemployment situation (i.e., overqualification). Shame is a basic human emotion that reflects feelings of low self-worth that derive from the devaluation of the social self (Gilbert 1998; Tangney 1991; Tangney et al. 2007). A negative judgment and evaluation of the self is at the core of shame. Shame normally includes feelings of embarrassment, disgrace, and humiliation (Tangney 1991).

In terms of social dimensions as one significant dimension of appraisal theories of emotions in Ellsworth and Scherer’s (2003) framework, we propose that overqualification has negative implications for social dimensions, including self-worth, self-image, and social identity. For involuntarily overqualified employees, overqualification, as a type of poor person-job fit (Cheng et al. 2020; Erdogan and Bauer 2009; Lin et al. 2017), implies a negative, derogatory comment or gesture about who they are, and what they can do (Bond and Venus 1991). Overqualified employees likely perceive and judge that they have fallen short of the established standards, including career success in the workplace (Erdogan et al. 2018). They may consider that their organization does not fully recognize their qualifications and value them, and thus they have not been justly assigned (Cheng et al. 2020). They further may perceive themselves to be inferior to those qualified others with the same qualifications as them. In addition, since individuals care about what they look like in front of others (Leary et al. 1995), overqualified employees cognitively appraise the loss of their dignity, authority, and power in the eyes of managers and other bystanders in the organization. As such, they may be negatively evaluated by themselves and others, posing an especially strong threat to self-worth, self-image, and social identity. Such negative self-evaluations and others’ evaluations (i.e., appraisal of social dimensions) promote feelings of shame (Brown and Weiner 1984; Gilbert 1998; Gruenewald et al. 2004).

Shame is characterized by withdrawal and avoidance tendencies (Peng et al. 2019). Overqualified employees experiencing shame tend to hide and withdraw from social interactions from which they likely anticipate disapproval by others (Peng et al. 2019; Tangney et al. 1996). They may therefore be motivated to escape such an unpleasant feeling (i.e., shame) by neglecting or exiting the undesirable employment situation, such as withdrawing from the situation or just keeping silent. We expect that shame with withdrawal and avoidance tendencies will positively relate to work withdrawal and employee silence.

Different from other forms of counterproductive work behaviors which are often depicted by a motivation to harm and disrupt the organization, work withdrawal consists of decreased job performance, lateness, absences, and turnover intentions, and is depicted by a motivation to avoid an unpleasant and unfavourable work situation (Koslowsky 2009; Spector et al. 2006; Yi and Wang 2015). Although undesired by the organization, work withdrawal may be regarded as a constructive coping style and a way to calm negative emotional states to return to the expected baseline state by employees (Grandey and Brauburger 2002). Overqualified employees experiencing shame may attempt to avoid the unpleasant and unfavorable underemployment situation or work experience, which has negative implications for their self-image, self-worth, and social identity. Work withdrawal could be viewed as a type of coping by overqualified employees experiencing shame in which they avoid a threat to their self-image, self-worth, and social identity in their work environment (Horan et al. 2019). Consequently, they may exhibit various forms of work withdrawal behaviors to remove themselves temporarily from the undesirable underemployment situation, ranging from being late for work, making excuses to get out of work, taking excessively long breaks, engaging in personal activities, to calling in sick when they are not sick (Hanisch and Hulin 1991; Horan et al. 2019). Rodriguez Mosquera et al. (2008) have shown that shame can predict withdrawal among participants from non-honour-oriented cultures.

As an important topic of organizational behavior research (Morrison and Milliken 2000; Van Dyne et al. 2003) and a passive counterproductive work behavior detrimental to organizations (Bolton et al. 2012; Van Dyne et al. 2003), employee silence refers to the intentional withholding of important information that might improve work procedures and outcomes by employees from others in their workgroup (Tangirala and Ramanujam 2008; Van Dyne et al. 2003). Based on employee motives, Van Dyne et al. (2003) differentiated three types of silence, namely acquiescent silence (i.e., disengaged behavior based on resignation), defensive silence (i.e., self-protective behavior based on fear), and pro-social silence (i.e., other-oriented behavior based on cooperation). Acquiescent silence has strongly passive elements and the other two forms of silence represent more proactive behaviors. In our study, we examined acquiescent silence as the outcome because it is highly relevant to employee shame that resulted from overqualification. Feelings of shame may reduce an overqualified employee’s belief that others will take his or her voice behaviors seriously (Peng et al. 2019). Lower efficacy in speaking up, therefore, motivates ashamed overqualified employees to withhold opinions and ideas that challenge the status quo from other workgroup members (Janssen and Gao 2015; Peng et al. 2019) to avoid embarrassing themselves (Brinsfield 2013; Livne-Ofer et al. 2019). Indeed, previous studies indicated that individual variables, including shame and limited self-efficacy, are significant players in engendering employee silence. Specifically, Noelle-Neumann (1974) proposed that feelings of self-doubt likely discourage individuals from sharing ideas that are inconsistent with public opinion. Edwards et al. (2009) further suggested that shame can drive employees to make decisions to keep silent. Livne-Ofer et al. (2019) have shown that shame as an inward-focused emotion mediates the effect of perceived exploitation on employee silence. Furthermore, because feelings of shame are an affective reaction toward the workplace as a whole, employee silence also is likely to be behavior-oriented in the workplace, more specifically, the working group, as a whole.

Taking them together, we propose the following hypothesis:

Hypothesis 1

Shame mediates the relationship between perceived overqualification and (a) work withdrawal and (b) employee silence.

Perceived overqualification and pro-job unethical behavior at work: the mediating effect of anger

In terms of motivational bases as one important dimension of appraisal theories of emotions in Ellsworth and Scherer’s (2003) framework, we further argue that overqualified employees also tend to experience anger toward undesirable and even frustrating employment situations and engage in pro-job unethical behavior. Pro-job unethical behavior refers to “actions that are intended to strengthen an employee’s bond with the job at the cost of violating standards of proper conduct” (Wang et al. 2019). To protect their self-job bond, employees may exhibit pro-job unethical behavior by engaging in unethical actions toward others (Wang et al. 2019), such as discrediting others’ performance or purposely excluding others in the workgroup (Wang et al. 2019).

As mentioned earlier, overqualification, as a type of person-job misfit (Cheng et al. 2020; Erdogan and Bauer 2009; Lin et al. 2017), describes an underemployment situation in which employees possess surplus qualifications relative to what a job requires (Erdogan and Bauer 2009; Johnson and Johnson 1996; Johnson et al. 2002), representing that individuals’ needs are not fulfilled by their organization (Liu et al. 2015). Overqualified employees likely feel that they are deprived of the job they are entitled to have (Erdogan et al. 2018; Hu et al. 2015) and the opportunities to fully play their talents and to show their aspirations (Liu et al. 2015). In other words, they tend to perceive their overqualification as an unfavorable and even a frustrating work experience that thwarts their personal goals, such as an interesting and qualification-matching job (Liu et al. 2015), recognition, appreciation, and personal growth in the organization. Anger appears when an individual experiences contempt, demeaning, harmful behaviors, or goal barriers (Berkowitz and Harmon-Jones 2004; Liu et al. 2015), and makes individuals more inclined to view, explain, and deal with events from a hostile perspective (Keltner et al. 1993; Lerner et al. 1998). Perceived obstruction of personal goals (i.e., appraisal of motivational bases) is likely to predispose overqualified employees to perceive, interpret, and act on their overqualification from a hostile perspective and thus engender anger (Berkowitz and Harmon-Jones 2004). Supporting these arguments, Gabriel et al. (2014) have found that mismatch triggers anger. Liu et al. (2015) have shown that perceived overqualification enhances the level of anger toward the employment situation. Consequently, we expect that overqualified employees tend to feel angry about their frustrating underemployment situation with poor person-job fit.

In particular, when overqualified employees attribute their frustrating underemployment situation to factors related to the organization or supervisor, they may consider that their expectations of fairness have been greatly violated. Overqualified employees likely believe that the organization or supervisor is usually responsible for the undesirable and unsatisfactory underemployment situation, thus they may hold hostile feelings toward the organization or supervisor embodying the organization (Eisenberger et al. 2010). The role of supervisors is critically important in triggering employees’ anger at work as they are gatekeepers between the organization and employees (Hammer et al. 2021). Domagalski and Steelman (2005) found that unjust treatment by supervisors was positively related to employee anger. Similarly, Fitness (2000) found that unjust treatment perpetrated by supervisors induces the highest level of employee anger. Harlos and Pinder (2000) also found a significant association between interactional injustice and employee anger. Liao et al.’s (2022) recent work, drawing on the affect theory of social exchange, showed that employees would experience increased anger when their resource exchange with leaders involves a negative imbalance.

Because anger is based on blaming others for the undesirable event, the behavior tendency related to anger is to “fight” or to take antagonistic behaviors against others are responsible for the undesirable event and the anger (Barclay et al. 2005; Keltner et al. 1993; Roseman et al. 1994). Consequently, anger tends to be not associated with withdrawal responses including running away and withdrawing to a safe place, but to be linked with the appraisal that others are accountable for negative events (Smith and Lazarus 1993) and “fight” or aggressive responses including yelling, hitting, and getting back at someone (Lim et al. 2016; Roseman et al. 1994). However, they normally are less inclined to take direct, obvious actions against the organization or supervisor, because the organization or supervisor controls the resources they value (Yukl 2010). Moreover, whereas overqualified employees may initially direct their hostile feelings toward the organization or their supervisor who acts as the agent of the organization (Liu et al. 2020), these feelings may, over time, generalize to antagonistic feelings toward other members of the organization. As a result, they may further be inclined to take direct, obvious deviant behaviors directed against their coworkers. We, therefore, expect that anger elicited by overqualification will be linked with aggressive responses, such as pro-job unethical behavior. Specifically, to express their anger about their misfit toward the organization or “get even” (Liu et al. 2015) as well as improve unpleasant underemployment situations, overqualified employees may ignore ethical concerns and be more prone to taking unethical actions and causing harm toward coworkers (Bazerman and Tenbrunsel 2011). In addition, they may display either verbal (e.g., a loud and impatient tone) or nonverbal (e.g., a frown in facial expression) signals, triggering adverse reactions in their coworkers, including withdrawal from those angry overqualified employees (Côté 2005; Peng et al. 2019). Coworkers’ adverse reactions would further verify those angry overqualified employees’ negative perceptions and explanations, encouraging pro-job unethical behavior directed toward coworkers (Peng et al. 2019).

Taking them together, we propose the following hypothesis:

Hypothesis 2

Anger mediates the relationship between perceived overqualification and pro-job unethical behavior.

Perceived overqualification and pro-job unethical behavior: the serial mediating effect of shame and anger

We further propose that shame likely triggers anger. With poor person-job fit exposure, ashamed employees who are overqualified may consider their self-image, self-worth, and social identity in their work environment as being damaged (Gausel and Leach 2011). Due to feeling good about themselves as one of the most important fundamental human motives, they tend to be overwhelmed by the sense of shame originating from overqualification, and thus are emotionally painful (Tangney and Dearing 2002). In turn, the painful feelings of shame likely elicit anger towards others. Specifically, as one of the most painful feelings for individuals (Tangney and Dearing 2002), shame likely promotes employees to adopt the strategy of “anger towards others” to deal with the social threat (i.e., the damaged self-image, self-worth, and social identity) and ease the painful feeling (Elison et al. 2014; Thomaes et al. 2011). Providing support for this argument, extant studies have shown that individuals high in trait shame (e.g., Harper and Arias 2004; Harper et al. 2005; Scott et al. 2015; Tangney et al. 1992) or exposed to a shameful event (Pivetti et al. 2016; Thomaes et al. 2011) are more likely to be angry.

Building upon the above reasoning and based on Hypotheses 1 and 2, we posit that shame and anger serially mediate the relationship between perceived overqualification and pro-job unethical behavior. Specifically, overqualified employees tend to perceive their self-worth, self-image, and social identity as having been compromised and thus experience shame. Shame as one of the most painful feelings for individuals (Tangney and Dearing 2002), in turn, is likely to trigger anger as the strategy to deal with the social threat and ease the painful feeling, thereby stimulating ignorance of ethical concerns and engagement in pro-job unethical behavior directed toward coworkers to express anger and get better treatment. In summary, we hypothesize:

Hypothesis 3

Perceived overqualification has a positive indirect effect on pro-job unethical behavior via a sequence of relationships in which perceived overqualification positively predicts shame, which in turn positively predicts anger, and subsequently positively predicts pro-job unethical behavior.

Methods

Sample and procedures

Following extant studies (e.g., Butts et al. 2015; Ng and Yam 2019; Panaccio and Vandenberghe 2012), we recruited study participants based on our personal and professional networks. The participants were full-time office employees in Southern China. They worked in various industries, including retail, manufacturing, informational technology, banking, education, and others. In the recruitment process, we ensured the anonymity, voluntariness, and confidentiality of the participation to participants, and asked them to complete a three-wave, time-lagged survey. Participants were only notified that the survey was associated with their job perceptions and behaviors. We did not disclose any specific goals of the survey to participants.

Data were collected through Web-based surveys conducted across three data collection periods at Time 1, Time 2 (2 weeks later after Time 1), and Time 3 (2 weeks after Time 2). A time lag of 2 weeks is a common practice in the field of organizational behavior research (e.g., Zhang et al. 2016; Yam et al. 2018), especially in emotion studies (e.g., Gabriel et al. 2020; Pundt and Venz 2017). The time taken for both generation and influence of emotions is relatively short. In our study, the mediators are two discrete emotions—shame and anger. Therefore, a time lag of 2 weeks is appropriate. At Time 1, we distributed the first survey to 400 employees; 383 responded, yielding a response rate of 95.75%. The Time 1 survey included demographic information and a measure of overqualification. At Time 2, 2 weeks after Time 1, we distributed the second survey to participants who engaged in the Time 1 survey, which was completed by 305 employees, yielding a response rate of 76.25%. The Time 2 survey included measures of shame and anger. At Time 3, 2 weeks after Time 2, 305 participants completed the Time 3 survey that included measures of work withdrawal, employee silence, and pro-job unethical behavior, yielding a response rate of 76.25%. After removing suspicious responses, the final matched data across the three timepoints included 225 employees, constituting an overall response rate of 56.25%.

One hundred and twenty-six (54.22%) participants were female. The average age was 30.75 years (SD = 8.91). Most were highly educated, with 78.22% holding college degrees. Participants were employed in a variety of industries, including retail (14.67%), manufacturing (13.33%), informational technology (10.67%), education (10.67%), banking (9.33%), and others (41.33%).

Measures

According to Brislin (1986)’s recommendation, all survey items were translated and back-translated from English to Chinese. Specifically, two of the authors proficient in both Chinese and English first translated all survey items from English to Chinese. Then, we invited a bilingual expert who did not know the research questions to translate the Chinese translations of these survey items back into English. Finally, the other two authors whose native language was Chinese, further revised the Chinese items to reduce those minor discrepancies.

Perceived overqualification (Time 1)

At Time 1, participants rated their perceived overqualification with a four-item scale developed by Johnson and Johnson (1996, 1997). Such a four-item scale has been widely applied in previous overqualification research (e.g., Erdogan and Bauer 2009; Hu et al. 2015; Johnson et al. 2002; Lee et al. 2021). Sample items include “Based on my skills, I am overqualified for the job I hold” and “My work experience is more than necessary to do my present job” (1 = strongly disagree to 5 = strongly agree; Cronbach’s alpha = 0.76).

Shame and anger (Time 2)

Participants rated the degree of each specific emotion they had experienced during the past 2 weeks at work in the Time 2 survey. Following Boudewyns et al. (2013), we used four items adapted from Tangney and Dearing (2002) to measure shame. Sample items include “I feel ashamed”, and “I feel embarrassed” (1 = strongly disagree to 5 = strongly agree; a = 0.93). We used a 3-item scale derived from Smith et al. (1993) to measure anger. Sample items include “I feel angry”, and “I feel enraged” (1 = strongly disagree to 5 = strongly agree; Cronbach’s alpha = 0.91).

Work withdrawal (Time 3)

Using a 4-item scale developed by Spector et al. (2006) and adapted by Chong et al. (2020), participants reported their work withdrawal at Time 3. These items include, “I Came to work late without permission”, and “I Stayed home from work and said I was sick when I was not” (1 = strongly disagree to 5 = strongly agree; Cronbach’s alpha = 0.93).

Employee silence (Time 3)

We measured employee silence using five items adapted by Tangirala and Ramanujam (2008) from Van Dyne et al.’s (2003) employee silence scale. Tangirala and Ramanujam’s (2008) scale, which was used in the context of patient safety in the hospital setting, has been widely used by later studies to assess employee silence in a general work context (e.g., Livne-Ofer et al. 2019; Xu et al. 2015). Participants reported the extent to which they withheld ideas, information, and opinions that could have improved the workplace. Sample items include “I chose to remain silent when I had concerns about the work in my workgroup” and “Although I had ideas for improving the work in my workgroup, I did not speak up” (1 = strongly disagree to 5 = strongly agree; Cronbach’s alpha = 0.88).

Pro-job unethical behavior (Time 3)

Using a 5-item scale adapted from Wang et al. (2019), participants reported their pro-job unethical behavior at Time 3. Sample items include “I failed to tell others in the workgroup that something they did was done wrong so they do not perform as well as me” and “I withheld information to put myself ahead of others in the workgroup” (1 = strongly disagree to 5 = strongly agree; Cronbach’s alpha = 0.93).

Controls

We measured employee gender (1 = male, 2 = female), age (measured in years), and educational level (measured in years) at Time 1. All of them were not related to the two discrete emotions as well as three behavior outcomes. We conducted all the analyses with and without these three demographic controls, and the two results were essentially identical. As Carlson and Wu (2012) recommended, we can report results without these controls but present their bivariate correlations with the study variables in Table 1. Considering that gender is an important factor in emotion studies (e.g., Davis et al.1992; Erickson and Ritter 2001; Scott and Barnes 2011), we reported regression results that included gender as a control variable.

Table 1.

Means, SDs, correlations, CRs, AVEs and Cronbach’s alphas

Variables	Mean	SD	CR	AVE	1	2	3	4	5	6	7	8	9
1. Gender (T1)	1.54	0.50	–	–	–
2. Age (T1)	30.75	8.91	–	–	− 0.09	–
3. Education (T1)	14.72	2.34	–	–	− 0.02	− 0.39***	–
4. Perceived overqualification (T1)	2.76	0.82	0.76	0.45	− 0.06	− 0.09	− 0.01	0.76
5. Shame (T2)	2.14	1.06	0.93	0.77	− 0.01	− 0.06	− 0.01	0.19**	0.93
6. Anger (T2)	2.17	1.02	0.91	0.67	− 0.06	0.00	− 0.10	0.19**	0.65***	0.87
7. Work withdrawal behavior (T3)	1.79	1.02	0.94	0.79	− 0.12	− 0.01	− 0.05	0.18**	0.28***	0.31***	0.93
8. Employee silence (T3)	2.61	0.95	0.89	0.62	− 0.06	− 0.04	0.06	0.13+	0.27***	0.32***	0.49***	0.88
9. Pro-job unethical behavior (T3)	2.00	0.98	0.94	0.74	− 0.02	− 0.03	− 0.08	0.14*	0.29***	0.30***	0.75***	0.49***	0.93

N = 225. Numbers 1–9 in the top row correspond to the variables in the respective sections of the table. Cronbach’s alphas are on the diagonal in bold and italic

T time, CR composite reliability, AVE average variance extracted

⁺p < 0.10, *p < 0.05, **p < 0.01, ***p < 0.001. Two-tailed tests

Analytic strategy

By using the Lavaan package in R software (https://www.R-project.org; Rosseel 2012), we first conducted a serial confirmatory factor analysis (CFA) to determine whether those measures in this study are acceptable or not. According to the recommendation of previous researchers (Browne and Cudeck 1992; Hu and Bentler 1999; McDonald and Ho 2002), we selected four common indices to evaluate the model fit: the Tucker–Lewis Index (TLI), Comparative Fit Index (CFI), Root Mean Squared Error of Approximation (RMSEA), and Standardized Root Mean Squared Residual (SRMR). Both Hypotheses 1 and 2 proposed that one mediator mediates the relationship between an independent variable and a dependent variable. Hypothesis 3 proposed that two mediators serially mediate the relationship between an independent variable and a dependent variable. To calculate these indirect effects, we used the PROCESS macro (model 4 and model 6) developed by Hayes (2022) in R software. Specifically, we used PROCESS to produce the estimates value of the indirect effects, along with a 95% bias-corrected confidence interval generated from 5000 bootstrap samples.

Results

Preliminary analysis

The means, standard deviations, and correlations among study variables are presented in Table 1. Cronbach’s alpha coefficients were all greater than 0.70, indicating that all scales showed adequate reliability (Nunnally 1978). As expected, overqualification was correlated with shame (r = 0.19, p < 0.01) and anger (r = 0.19, p < 0.001), and work withdrawal (r = 0.18, p < 0.01), employee silence (r = 0.13, p < 0.01), and pro-job unethical behavior (r = 0.14, p < 0.05). Shame was significantly correlated with work withdrawal (r = 0.31, p < 0.001) and employee silence (r = 0.32, p < 0.001). Anger was significantly correlated with pro-job unethical behavior (r = 0.29, p < 0.001).

We conducted confirmatory factor analyses (CFAs) to test the distinctiveness of the six variables (i.e., perceived overqualification, shame, anger, work withdrawal, employee silence, and pro-job unethical behavior). We first compared this six-factor model to all five-factor models in which two highly correlated factors were combined respectively. As Table 2 showed, the hypothesized six-factor model fit the data well and all indices met the acceptable criteria (χ² (260) = 667.23, CFI = 0.91, RMSEA = 0.08, SRMR = 0.06; Hu and Bentler 1999). Chi-square difference tests indicated that the six-factor model was significantly superior to all five-factor models (see Table 2). Besides, we followed the suggestion of Fornell and Larcker (1981) to further examine the discriminant validity. The square root of the average variance extracted was shown on the diagonal and was greater than the correlation coefficients (off-diagonal) for each construct in the relevant rows and columns (see Table 1). Thus, all six key variables could be treated as distinct constructs. Based on the CFA result of the six-factor model, all the indicators loaded significantly on their corresponding latent constructs, with standardized coefficients ranging from 0.51 to 0.94 (all p < 0.001) and surpassing the conventional cut-off value of 0.40 (Hinkin 1998). Besides, referring to Table 3, the composite reliability (CR) for all constructs were all above 0.70, and the average variance extracted (AVE) values were all above 0.50, except POQ, which was very close to the critical value (AVE = 0.45). Thus, according to the analysis above, the convergent validity of each scale was satisfactory.

Table 2.

Confirmatory factor analysis

Models	Factors	χ2 (df)	Δχ2 (Δdf)	CFI	TLI	SRMR	RMSEA
Baseline model	Six-factor model	735.22 (309)	–	0.91	0.90	0.05	0.08
Model 1	Seven-factor: common method factor added	731.60 (308)	3.62 (1)	0.91	0.90	0.06	0.08
Model 2	Five-factor: Anger and shame combined	843.22 (314)	107.99 (5)***	0.89	0.88	0.05	0.09
Model 3	Five-factor: WWD and ES combined	1169.52 (314)	434.30 (5)***	0.82	0.80	0.08	0.11
Model 4	Five-factor: ES and PUB combined	1141.84 (314)	406.62 (5)***	0.83	0.81	0.08	0.11
Model 5	Five-factor: WWD and PUB combined	965.34 (314)	230.12 (5)***	0.87	0.85	0.05	0.10
Model 6	Three-factor: Anger and shame combined; WWD, ES and PUB combined	1478.29 (321)	743.07 (12)***	0.76	0.74	0.08	0.13
Model 7	One-factor model	2910.76 (324)	2175.60 (15)***	0.47	0.42	0.17	0.19

N = 225. WWD work withdrawal, ES employee silence, PUB pro-job unethical behavior, CFI comparative fit index, TLI Tucker–Lewis index, RMSEA root mean squared error of approximation, SRMR standardized root mean squared residual

*p < 0.05; **p < 0.01; ***p < 0.001. Two-tailed tests

Table 3.

Results of regression analysis

Variables	Shame	Work withdrawal	Employee silence	Anger	Pro-job unethical behavior
Constant	1.45*** (0.32)	1.30*** (0.31)	2.07*** (0.30)	0.34 (0.18)	1.17*** (0.30)
Gender	− 0.01 (0.14)	− 0.28* (0.13)	− 0.15 (0.12)	− 0.01 (0.07)	− 0.07 (0.12)
Perceived overqualification	0.24** (0.08)	0.10 (0.08)	0.04 (0.08)	0.11* (0.05)	0.01 (0.08)
Shame		0.31*** (0.06)	0.31*** (0.06)	0.68*** (0.04)	0.16 (0.10)
Anger					0.26* (0.11)
R2	0.04	0.14	0.13	0.65	0.16
F	4.14*	11.48***	10.60***	134***	10.13***

N = 225. Unstandardized coefficients are reported with standard errors in the parentheses

*p < 0.05; **p < 0.01; ***p < 0.001. Two-tailed tests

Hypothesis testing

The results of the regression analysis are presented in Table 3. Hypothesis 1 predicted that shame mediates the relationship between perceived overqualification and work withdrawal (Hypothesis 1a) and employee silence (Hypothesis 1b). As shown in Table 3, after controlling for gender, perceived overqualification was positively related to shame (β = 0.24, p < 0.01), and shame was positively associated with work withdrawal (β = 0.31, p < 0.001) and employee silence (β = 0.31, p < 0.001). It can preliminary supported Hypothesis 1a and 1b.

The results demonstrated that the indirect effects of perceived overqualification on work withdrawal (indirect effect = 0.07; SE = 0.04; 95% CI [0.02–0.16]; direct effect = 0.09; SE = 0.08; 95% CI [− 0.06 to 0.25]) and perceived overqualification on employee silence (indirect effect = 0.07; SE = 0.03; 95% CI [0.02–0.14]; direct effect = 0.04; SE = 0.08; 95% CI [− 0.11 to 0.19]) through shame were both significantly different from zero (Preacher and Hayes 2008). Therefore, Hypothesis 1a and Hypothesis 1b were both supported.

As the same procedure above, we further examine Hypothesis 2, which predicted that anger mediates the relationship between perceived overqualification and pro-job unethical behavior. According to the results of regression analysis (see Table 3), perceived overqualification (β = 0.11, p < 0.05) and shame (β = 0.68, p < 0.001) both positively predicted anger, and anger was positively associated with pro-job unethical behavior (β = 0.26, p < 0.05). Further, the bootstrap analyses demonstrated the 95% confidence interval of indirect effects of perceived overqualification on pro-job unethical behavior through anger (indirect effect = 0.03; SE = 0.02; 95% CI [0.001–0.07]) did not contain zero. Thus, Hypothesis 2 was supported.

Hypothesis 3 predicted that perceived overqualification has a positive indirect effect on pro-job unethical behavior via a sequence of relationships in which perceived overqualification positively predicts shame, which in turn positively predicts anger, and subsequently positively predicts pro-job unethical behavior. Based on the results of testing Hypothesis 2, we further calculated the indirect effects of perceived overqualification on pro-job unethical behavior through shame and subsequently anger. The bootstrap analyses demonstrated the 95% confidence interval did not include zero (indirect effect = 0.04; SE = 0.02; 95% CI [0.01–0.10]). Thus, Hypothesis 3 was supported by data (Preacher and Hayes 2008).

Discussion

How and why do overqualified employees engage in withdrawal and aggression at work? Across a three-wave field study in China, we found support for our model of the effects of perceived overqualification on discrete emotions experienced at work, including shame and anger, and withdrawal and aggression at work. Specifically, perceived overqualification elicits shame, which in turn is positively associated with work withdrawal and employee silence. Perceived overqualification also triggers anger, which in turn is positively associated with pro-job unethical behavior. In addition, perceived overqualification has a positive indirect effect on pro-job unethical behavior through the serial mediating effect of shame and anger. Our findings provide several important theoretical implications for overqualification literature.

Theoretical implications

First, a noteworthy implication is that we empirically identify both withdrawal (i.e., work withdrawal and employee silence) and aggression at work (i.e., pro-job unethical behavior) as outcomes of perceived overqualification in the same theoretical framework. Previously established negative behavioral outcomes of perceived overqualification, including either withdrawal behaviors such as voluntary turnover (Erdogan and Bauer 2009; Maynard and Parfyonova 2013) and cyberloafing (Cheng et al. 2020; Khan et al. 2022), or aggressive behaviors such as counterproductive work behavior (Liu et al. 2015; Luksyte et al. 2011). However, the negative behavioral reactions to perceived overqualification are more nuanced than previously assumed. We serve work withdrawal and employee silence as withdrawal at work and regard pro-job unethical behavior as aggression at work. Our findings provide empirical support that overqualified employees tend to engage in both withdrawal (i.e., work withdrawal and employee silence) and aggression at work. This thus provides a complete understanding of the negative behavioral reactions to overqualification and highlights the need for a dialectical perspective on the negative consequences of overqualification in organizations.

Second, discrete emotions provide a unique lens to understand how and why perceived overqualification precipitates a range of undesirable behaviors. Previously established mediating variables, including person-job fit, psychological contract, and cynicism (Luksyte et al. 2011), job satisfaction (Maynard and Parfyonova 2013), relative deprivation (Erdogan et al. 2018; Lee et al. 2021), organization-based self-esteem (Liu et al. 2015), work-related positive affect and perceived job autonomy (Simon et al. 2019), perceived insider status (Li et al. 2019), harmonious passion (Cheng et al. 2020), and boredom sensations and being envied by coworkers (Howard et al. 2022), are helpful to understand the generally negative consequence of perceived overqualification. The equity theory suggests that employees are attentive to the equity of situations within the workplace (Tabibnia et al. 2008). Drawing on equity theory, some overqualification studies have examined anger as a mediator of perceived overqualification and counterproductive work behaviors (e.g., Liu et al. 2015). However, the equity theory fails to adequately explain individuals’ differential emotional and behavioral reactions to overqualification. We found that it is critical to consider not only anger but also the emotion of shame, and these two emotions may promote distinct outcomes that were previously not considered in the overqualification literature. The defining behavioral tendencies for shame and anger are to withdraw from the social situation and to engage in aggressive and retaliatory acts, respectively (Lazarus 1991; Roseman 2013). Shame motivates work withdrawal and employee silence, while anger drives pro-job unethical behavior. Drawing on the cognitive appraisal theories, we showed that discrete emotions represent unique processes through which perceived overqualification influences different behaviors.

Third, this research is among the first to integrate how the emotional processes of “withdrawal or avoidance” (i.e., shame) and “aggression or hostility” (i.e., anger) serially mediate the relationship between perceived overqualification and aggression at work. The overqualification literature highlights the critical role of emotional processes (Liu et al. 2015; Liu and Wang 2012). However, research focus on the emotional change process is scant. To fill this gap, our research theorizes how and why shame as the “withdrawal” emotion and one of the most painful feelings (Tangney and Dearing 2002) experienced by overqualified employees triggers anger as the “aggressive or hostile” emotion that is proximal to pro-job unethical behavior. Studying the emotional processes of “withdrawal or avoidance” and “aggression or hostility” simultaneously allows us to identify a serial mediating pathway through which perceived overqualification relates to pro-job unethical behavior. Our findings provide empirical support that shame as the “withdrawal or avoidance” emotion is painful; anger as the “aggression or hostility” emotion serves as a more proximal predictor for pro-job unethical behavior and relates the “withdrawal or avoidance” emotion to pro-job unethical behavior. Specifically, the positive serial mediating effect of shame and anger suggests that overqualified employees are prone to experience shame due to their damaged self-worth, self-image, and social identity. Such a most painful feeling for individuals (i.e., shame) (Tangney and Dearing 2002) triggers anger, an emotional state with the strategy to deal with the social threat and ease the painful feeling of shame, thus leading to a high level of pro-job unethical behavior. This finding extends the overqualification literature by revealing overqualified employees’ emotional change process. Taken together, this serial mediating pathway provides a more comprehensive understanding of how the emotional processes of “withdrawal or avoidance” and “aggression or hostility” work together to link overqualification to aggression at work.

Practical implications

Our findings reveal that employees’ perceived overqualification could result in distinct negative emotions (i.e., shame and anger) and behaviors. However, if the organization just rejects overqualified applicants or replaces overqualified employees in an oversimplified and crude way based on such findings, the organization will also lose the high job performance that those overqualified employees may achieve (Liu et al. 2015). Hence, we are not here to suggest that organizations should reject overqualified applicants or replace overqualified employees with just-qualified ones. Instead, by revealing their unpleasant psychological process, we argue that, when managed properly, overqualified employees could be beneficial to the organization (Erdogan and Bauer 2009; Liu et al. 2015). Hence, we recommend that organizations implement formal policies and procedures to reduce or eliminate overqualified employees’ withdrawal and pro-job unethical behaviors at work. According to our study, perceived overqualification induces negative emotions and behaviors due to two reasons, one is from the motivational bases, and the other is from the social dimension aspect. Hence, effective recommendations can be generated based on these two aspects.

First, in terms of the motivation basses, we found the mediating role of anger in the overqualification–aggression linkage. Organizations and managers thus should assist and support those overqualified employees to achieve their personal goals and thus minimize their anger as a hostile emotion. Specifically, managers should take the initiative in communicating the poor person-job fit with their overqualified employees (Liu et al. 2015), make them believe that they are the assets of the organization, and find ways including mentoring, participative decision-making, or job crafting (Lin et al. 2017) to improve their unpleasant underemployment situation. Finally, they do not need to ignore ethical concerns and take aggressive actions at work in order to express their anger and change the status quo.

Second, in terms of the social dimension aspect, we found the mediating role of shame in the overqualification-withdrawal linkage. We recommend that organizations and managers help overqualified employees manage shame in adaptive ways (Bonner et al. 2017), improve their self-worth, self-image, and social identity, and thus minimize their shame as a painful feeling; Both mentoring and participative decision-making can improve their self-worth, self-image, and social identity as well as be helpful to achieve their personal goals. González-Gómez and Richter (2015) suggested managers to invest in training or intervention programs regarding the regulation of shame, such as suppression of emotional displays to promote social relationships in the workplace. As a result, these employees will no longer tend to be ashamed and eventually withdraw from work.

Furthermore, organizations and managers should provide emotional regulation training for employees to regulate negative emotions and generate positive emotions to help them recover from negative emotions in stressful situations (Tugade and Fredrickson 2004; Liu et al. 2015). For example, in emotional training, overqualified employees can be suggested to vent and regulate negative emotions by participating in mindfulness meditation, physical exercise, going out with friends, performing personal hobbies, and other activities.

Limitations and future research

First, our measure of perceived overqualification is self-reported, which might suffer from common method bias. Although previous studies also measured overqualification through employees’ self-reports (e.g., Cheng et al. 2020; Erdogan et al. 2018; Lee et al. 2021), we have to note that self-reported overqualification may differ from objective overqualification as well as other-reported overqualification. To test the common method bias, we conducted Harman’s one-factor test by the Psych package of R, which loaded all six key variables into an exploratory factor analysis. The unrotated factor solution showed that the first factor explained 38.70% of the total variance, close to the critical criterion of 40%. To further check whether common method bias disrupted the interpretation, we utilized the ULMC approach recommended by Podsakoff et al. (2003). Based on the six-factor model, we added a latent method factor to build a seven-factor model. As shown in Table 3, there was no significant difference between the hypothesized six-factor model and the seven-factor model (Δχ² = 3.62, df = 1, p > 0.05). Therefore, common method bias was not a big problem in our study (Richardson et al. 2009). In addition, our study applied a convenience sampling of employees from various organizations and multiple industries, leading to a biased sample with a nonprobability sampling and limited generalization. Hence, we suggest future research to examine both other- and self-reported overqualification of employees from a single organization.

Second, although we have tested the mediating roles of shame and anger, perceived overqualification also may relate to other emotional states, including fear and sadness. Without the help and support of organizations and managers, overqualified employees may consider that their control over their underemployment is limited, which leads to their fear due to the lack of control (Smith and Ellsworth 1985). Moreover, to a certain extent, the poor person-job fit implies that the career is not successful, which may create feelings of sadness. Consequently, fear, sadness, and other emotions are worthy of attention in future studies of overqualification.

Third, we have not examined any boundary condition of the link between perceived overqualification and withdrawal and aggression at work. Future research can examine both individual (e.g., trait shame, trait anger) and contextual (e.g., mentoring, participative decision-making) variables as moderators. For example, as the individual characteristics, trait shame might magnify the effects of perceived overqualification on shame as an emotional state and withdrawal at work, and trait anger might amplify the effects of perceived overqualification on anger as an emotional state and aggression at work. As the contextual factors, mentoring and participative decision-making might improve employees’ self-worth, self-image, and social identity and be beneficial for them achieving their personal goals, and hence be less likely to create feelings of shame and anger, as well as dysfunctional behaviors.

Fourth, the generalizability of our results may be limited because we only collected data in China and did not measure any cultural variables. Chinese culture emphasizes collectivism and group relationships and regulates individual behavior through feelings of shame, which is an important self-conscious emotion (Fung and Chen 2001; Li et al. 2004), whereas Western culture places greater emphasis on individualism, independence, and individual goals (Javidan et al. 2006). It can be inferred that, compared to those from a Western cultural context, overqualified employees in Chinese organizations are more likely to experience shame, and less likely to experience anger. Even in anger, overqualified employees in China are less likely to conduct supervisor-target aggressive behaviors due to the high power-distance cultural context. Instead, they would engage in coworker-target pro-job unethical behaviors. We, therefore, encourage future research to replicate our model findings with a sample of employees from other cultures, and measure cultural variables to prevent criticism regarding the potential lack of generalizability to other populations.

Fifth, we have not differentiated between voluntary and involuntary overqualification. A basic assumption of the present study is that it is involuntary and undesirable for employees to be overqualified. Indeed, it is proper for the majority of employees with a high need for competence and growth (Liu et al. 2015; Maynard and Parfyonova 2013). However, overqualification may also be a desirable attribute for certain employees, such as older employees who aim to maintain their lifestyle and social relations, and younger employees who aim to obtain work experience and training (Liu et al. 2015; Liu and Wang 2012). As an employment situation chosen by employees themselves, voluntary overqualification is less likely to result in feelings of shame and anger. Consequently, we recommend future research to further examine both voluntary and involuntary overqualification, and investigate the potentially different effects they may have.

Conclusion

In the present study, we built upon appraisal theories of emotions and found that perceived overqualification relates to distinct action tendencies through the emotional mechanism. It leads to increased work withdrawal and employee silence through the mediating role of shame, as well as increased pro-job unethical behavior through the mediating role of anger and the serial mediating effect of shame and anger. We hope the introduction of appraisal theories of emotions and the present work can spark future research on overqualification.