“Who can I blame and what can I do?”: making sense of psychological contract evaluations between Belgian and Chinese employees

This study explores cultural differences in the sensemaking process of employees following psychological contract evaluations. Data from 20 Belgian and 21 Chinese employees were collected using qualitative methods. An analysis of 94 critical incidents relayed by the employees reveals the attributional, emotional, and behavioral reactions that are triggered by psychological contract under-, exact-, and over-fulfillment. Our findings suggest that supervisors were seen as directly responsible for most of the employees’ psychological contract evaluations. Emotional responses are more complex in the process of attributing responsibility. Behavioral actions are subsequently used to deal with three types of psychological contract evaluations. A number of subtle differences are found between the Belgian and Chinese employees. The results highlight the unfolding and dynamic nature of the psychological contract in cultural comparisons.


Introduction
The psychological contract (PC) describes a person's beliefs regarding the terms of a reciprocal exchange agreement between that person and another party (Rousseau 1989). It encompasses the individual's perceptions of the mutual obligations between both parties (Rousseau 1989). A PC breach refers to an employee's cognition that an organization has failed to deliver on one or more of its obligations to the individual commensurate with his or her contributions to the organization (Morrison and Robinson 1997). When the organization is seen to meet all its obligations, the PC is said to be fulfilled (Conway and Briner 2009). Lambert et al.'s (2003) expanding the approach of PC breach further distinguishes between two types of breach. By considering every discrepancy between what is delivered and what is perceived to be obligated as a breach, they distinguish over-fulfillment (i.e., receiving more than what had been obligated to them) from under-fulfillment (i.e., receiving less than what had been obligated to them). In recent decades, scholars have started to explore the temporal processes of PC evaluations (e.g., Conway and Briner 2002), and empirical studies have demonstrated the generative mechanisms that drive changes in PC evaluations (Bankins 2015). These studies have been predominantly quantitative and focused on changes in employee attitudes and behaviors due to fluctuations in breach and fulfillment perceptions (e.g., Conway and Briner 2002). Simultaneously, a smaller body of qualitative studies have identified the construction of responses to PC evaluations, highlighting the sensemaking processes triggered by PC breach and illustrating that employees do not passively reciprocate negative events such as breach (Bankins 2015;Chaudhry et al. 2009;Parzefall and Coyle-Shapiro 2011;van der Schaft et al. 2020;Woodrow and Guest 2020). PC theory states that breach triggers a sensemaking process (Morrison and Robinson 1997), and previous research has explored how employees make sense of such events (Bankins 2015;Parzefall and Coyle-Shapiro 2011). However, sensemaking can be conceived as a way to study how people construct a stable social world through everyday utterances and actions (Garfinkel 2016). Past sensemaking studies have focused on breach as under-fulfillment, whereas the PC literature acknowledges that breach represents any discrepancy between delivered and obligated inducements and that it therefore comprises both under-and over-fulfillment (Lambert et al. 2003). There is evidently a lack of a more thorough understanding of how employees make sense of under-fulfillment, over-fulfillment, and exact-fulfillment events.
Moreover, PC evaluations trigger sensemaking processes that may be influenced not only by situational variables but also by individual differences, such as one's personality and cultural background (Shih and Chuang 2013). To date, the role of culture in these sensemaking processes is not well understood; cultural differences in employees' evaluations on their PC have received scarce empirical attention (Schalk and Roe 2007). Most studies on PC have relied on Western samples, with only limited attention devoted to other parts of the world, such as Eastern countries (e.g., Hui et al. 2004). Nonetheless, there are strong reasons to believe that cultural differences play an important role in how employees perceive, interpret, and react to PC breach and fulfillment (Thomas et al. 2003). Coyle-Shapiro et al. (2019) call for PC researchers to consider contextual factors, such as cultural differences, in PC evaluations. We address this gap in the literature by comparing sensemaking processes triggered by PC breach and fulfillment perceptions between Chinese and Belgian employees.
Given the above, this study makes three main contributions to the existing literature. First, research on sensemaking processes triggered by PC evaluations has thus far relied on a traditional perspective regarding PC breach. Put differently, studies have focused on breach as under-fulfillment, overlooking the sensemaking process that is triggered by breach as overfulfillment. We address this issue by relying on an expanded view for PC evaluations (Lambert et al. 2003). Moreover, we treat PC processes as dynamic phenomena that can alter over time. Second, there is a dearth of research on sensemaking theory from an Eastern cultural context. By comparing sensemaking processes in a Western (Belgium) context and an Eastern (China) context, we enrich our understanding of the boundary conditions created by cultural differences on PC and sensemaking process. Finally, the attributional, emotional, and behavioral reactions to workplace events remain unexplored (Bankins 2015;Kuppens et al. 2017). We attempt to further explore these reactions through the lens of sensemaking process.

Sensemaking theory and psychological contract evaluations
Sensemaking is a generic term that refers to the process of interpretation whereby individuals and groups interpret and reflect on phenomena, which normally occurs when employees confront real events or issues that are somehow confusing or surprising (Weick 1995). These events or issues trigger efforts to retrospectively make sense of what occurred (Weick 1995). Therefore, such events may first interrupt an individual's ongoing activities, such as when employees experience a breach of the PC (Rousseau et al. 2018). Employees start to elicit specific process, including interpreting the event and acting based on this interpretation, to restore their PCs, in which they might either renegotiate, repair the PC, or have it deteriorate to an impaired state (Rousseau et al. 2018;Tomprou et al. 2015). In this process, contextual factors are crucial, which means that people in different contexts (e.g., different cultures) would be affected by the prevalent norms and expectations of their context (Weick 1995). Overall, PC breach, in the form of both under-and over-fulfillment, forms an unexpected event that interrupts the exchange relationship and is therefore believed to trigger a sensemaking process (Morrison and Robinson 1997).
According to attribution theory (Weiner 1985), individuals judge or attribute events to reduce uncertainty, increase control, and predict future events. Attribution refers to individuals making sense of their own behaviors or those of others, and this is relevant to PC studies because it helps explain negative reactions to PC breach (Weiner 1985). If employees perceive their employer has intentionally broken promises, they typically react more strongly. However, if employees believe that the PC breach was outside of the employer's control, they typically are more understanding and forgiving and, hence, react less strongly or not at all (Lester et al. 2002). For example, when an employer promises a new employee a promotion within 2 years but does not deliver on such a promise, this may be interpreted as reneging by employees (Robinson and Morrison 2000). In such a situation, organizations clearly know their obligations but fail to follow through because they are unable or unwilling to do so (Morrison and Robinson 1997). From the employee's perspective, the PC breach might be seen as deliberate and intentional (Morrison and Robinson 1997). However, sometimes the employer may have made some vague promise-for example, that the employee would be promoted rapidly-which is interpreted by the employee that promotion would follow within 2 years. When this promotion does not occur, the employee may perceive a PC breach (Robinson and Morrison 2000). Such a breach results from incongruence, since the employee and the employer possess different perceptions about the promise that was made (Turnley and Feldman 1999). In this case, incongruence might be attributed as being unintentional. Finally, PC breach can be attributed to external factors outside the organization's control (Cassar and Briner 2011), which could also be attributed as unintentional behaviors, such as employers being unable to fulfill obligations due to the changing conditions in the business environment. In general, attributing blame to intentional or unintentional factors would generally lead to different attitudinal and behavioral outcomes (Costa and Neves 2017).
Individuals try to attribute the event during the first stage of the sensemaking process, with this interpretation guiding individuals' corresponding actions. In such cases, a distinction of the attribution process is often made between "situational" and "dispositional" attributions (Weiner 1985). In the former case, people attribute the Du and Vantilborgh Frontiers of Business Research in China (2021) 15:10 event to transient, situational factors, whereas in the latter case, they attribute the event to stable, person-specific factors, such as personality or values. Attribution theory also assumes that cognitions initiate an emotional process that refines and differentiates individuals' experiences (Weiner 1985). In its first stage, this emotional process will lead to differences in experienced core affect, which can be described based on two dimensions: valence and arousal (Barrett 1998). Valence is positive or negative affectivity, that is, pleasantness versus unpleasantness. Meanwhile, arousal describes bodily activations, that is, calm or relaxed versus excited or aroused (Barrett 1998). In the second stage, the attribution of events such as PC breach leads to specific discrete emotions being experienced, such as anger or fear. As such, the attributional model goes beyond the broad descriptions of emotional reactions to PC breach, which typically remains limited to either feelings of violation or job satisfaction and dissatisfaction (Lambert et al. 2003;Zhao et al. 2007).
During the secondary appraisal process, individuals determine the coping strategies to enact (Lazarus and Folkman 1985). With regard to coping behaviors for every event, Skinner et al. (2003) distinguish three categories: proximity seeking (e.g., support seeking), mastery (e.g., negotiation), and accommodation (e.g., adjusting one's preferences to current constraints).

Role of culture in psychological contract
Culture comprises the values, attitudes, beliefs, and behavior systems shared by members of a social group that are passed on to each generation (Hofstede 1984). Most, even qualitative, research on the PC has been conducted with individuals from cultures that score high on individualism (for an exception, see Du and Vantilborgh 2020). Collectivistic societies generally emphasize in-group characteristics: Priority is given to group goals instead of individual goals; on how individual actions could impact the group; on advocating for sharing resources within the group; on members fighting for group values; and on social behaviors being influenced strongly by social rules (Singelis et al. 1995). In contrast, individualist societies tend to attach less importance to ingroups and focus instead on individual freedom.
It is safe to assume that culture influences cognitive schemas such as PC (Thomas et al. 2003). Thomas et al.'s (2003) cognitive and motivational theory explains how the PC is perceived and understood in case of PC breach. The cognitive component of their model aids in interpreting and perceiving employment relationships from different cultural contexts, while the motivational component explains the desires and preferences of individuals under culturally different self-concepts. For example, many scholars claim that employees facing PC breach in individualistic cultures focus on competition and fighting for their own benefits, whereas employees in collectivistic cultures focus on harmony and submission to group benefits (Huang 2016).
However, not every discrepancy between obligated and delivered inducements is considered to be a breach. Only when the discrepancy exceeds a threshold and thus surpasses the "zone of acceptance" will employees interpret the discrepancy as a breach (Schalk and Roe 2007). Collectivists might have a higher threshold for the perception of PC breach than individualists (Thomas et al. 2003)-including under-and overfulfillment-because, for example, individualists look for the explicit benefits for themselves, whereas collectivists are more concerned about the relationship between both parties and therefore tolerate discrepancies to maintain harmony (Thomas et al. 2003).
Overall, how one makes sense of a PC event, in part, depends on cultural differences. In the first attribution stage, collectivists may be likely to attribute PC events (under-, exact-, and over-fulfillment) to factors outside an organization's control, whereas individualists are expected to do the opposite. For example, individualists may blame the failure to gain a promotion on the supervisor's or organization's inadequate evaluation process. In contrast, collectivists may attribute this to the impossibility of evaluating all information and the inevitable subjectivity in personnel decisions (Thomas et al. 2003). Indeed, previous studies have found that Asians are more likely to make a situational attribution than Westerners (Choi et al. 1998). Turning to the emotional appraisal, studies have found that negative emotions are more commonly experienced in Western than Eastern contexts, whereas Eastern individuals (e.g., Japanese) experience more socially engaged emotions, such as friendly feelings (Mesquita et al. 2016). Moreover, high-arousal positive emotions are generally preferred by individuals from Western cultures due to core cultural values again emphasizing the influence of the environment and personal achievement, while Eastern individuals generally prefer low-arousal positive emotions, such as feeling relaxed and peaceful due to people adjusting themselves to the environment (Kuppens et al. 2017). Finally, since individuals in Eastern cultures may not be prone to show emotional reactions, we expect their behavioral reactions to be milder than Western individuals. Research has found that in collectivist cultures, people tend to have a shared understanding concerning the behaviors that are expected or considered appropriate in a given situation, whereas individualists in Western cultures express dissonant opinions more easily (Triandis 1994).
In general, most research regarding individualism and collectivism in PC stems from quantitative studies. Since there are not many qualitative studies of cultural comparisons in PC evaluations, we consider our study to be the start of a line of research that may offer a more nuanced perspective on cultural values that impact PC evaluations. In this study, we compare Belgian and Chinese employees as Chinese society has strong collectivist values with hierarchical principles (Hui et al. 2004), while Belgian society generally emphasizes individualist values with egalitarianism (Du and Vantilborgh 2020), suggesting that they could be regarded as representatives of individualistic and collectivistic societies, respectively. However, we acknowledge that there may be regional difference within both countries; hence, we focus on employees from the Flemish region in Belgium-which generally emphasizes equality and independence (Billiet et al. 2006)-and employees from the northern region of China-which is considered by many to be less influenced by Western culture (Fan 2000).
The sensemaking approach as presented in this study provides a methodological grounding, in which the application of cognitive processes assists in identifying the patterns/themes in data that can make sense of the processes of PC evaluations in two different cultural contexts-namely, Belgium and China. Based on the literature, we formulate the following research questions: RQ1: How do employees in Belgian or Chinese cultural context make sense of PC under-fulfillment, exact-fulfillment, and over-fulfillment? Method Sample Forty-one employees (n Chinese = 21, n Belgian = 20) were interviewed. After 17 interviews with the Chinese and 15 interviews with the Belgian interviewees, data saturation was reached, which indicated that new contents or categories stopped emerging. All interviewees were employed full-time except for two Belgian employees, who worked parttime. We used a purposeful sampling strategy to ensure employees came from various industries (e.g., administration, engineering and manufacturing, finance, education, and support services) and held diverse functions in these industries. Forty-four percent of the respondents were male, and respondents had a mean age of 34.05 years (SD = 7.78), with 9.33 years (SD = 7.97) of work experience on average. All participants obtained at least a university degree.

Procedure
We used the critical incident technique to collect data (Flanagan 1954). This technique offers a tool to explore significant events, such as PC breach, and it allows researchers to further explore how employees experience, make sense of, and react to these events (Parzefall and Coyle-Shapiro 2011). This study primarily sought to explore interviewees' sensemaking processes and reactions to PC under-, over-, and exact-fulfillment, and based on this, to explore differences between Belgian and Chinese employees. Interviews were semi-structured and conversational. For each interview, respondents were asked three questions to elicit critical incidents. Adopting the view that PC evaluations encompass situations that disrupt the ongoing reciprocal exchange relationship by either exceeding, delivering, or falling short of obligations (Lambert et al. 2003;Rousseau et al. 2018), we asked interviewees to describe a situation in which they believed their employer failed to fulfil/ exactly fulfil/over-fulfil one or more obligations to them. Interviewees were encouraged to provide sufficient details on each incident to increase the accuracy and quality of the information pertaining to the reported incidents.
Belgian participants were interviewed in English (fluency in English was included as an inclusion criterion when we recruited participants), while Chinese participants were interviewed in Chinese, the results were transcribed and subsequently translated into English. All interviews were conducted by the first author of this paper. Each respondent was asked to describe three incidents related to employer obligations. Interviews lasted approximately 1 hour to 1 hour and a half (M = 77.05 min), except for one interview, which lasted 30 min. A total of 94 critical incidents (n Belgium = 54, n China = 40) were relayed to us, an average of 2.7 incidents per Belgian interviewee and 1.9 per Chinese interviewee.

Data analysis
We analyzed our interview data in three stages:

Du and Vantilborgh Frontiers of Business Research in China
(2021) 15:10 Page 6 of 22 Stage 1: Sensemaking process In this stage, we listed all critical incidents and divided them into three groups: under-, exactly-, and over-fulfilled. Each incident was interpreted on its sensemaking properties (Weick 1995). First, PC evaluations may constitute a threat, status-quo, or aid to employees' identity (Conway and Briner 2002), and making sense of an employer's behavior could help protect the employee's self-identity (Parzefall and Coyle-Shapiro 2011). Second, PC evaluations are viewed as an equilibrium or interruption taking place in the ongoing exchange relationship (Parzefall and Coyle-Shapiro 2011;Rousseau et al. 2018), subsequently influencing employees' future behaviors. Next, we coded how interviewees perceived the responsibility and intentionality of each critical incident, thus characterizing the attribution of PC evaluation events. This led to the creation of attribution themes and subthemes, such as one theme representing the perceived responsibility of the event being either assigned to the direct supervisor or to the organization. The themes that emerged were interpreted using attribution theory, checking for theoretical plausibility. Two subthemes emerged for responsibility-supervisor versus organization-while two subthemes emerged for intentionality-intentional versus unintentional. Supervisor responsibility means that employees attribute an event to their direct supervisor, whereas organization responsibility means there is no specific person, but rather a contextual organizational factor that is attributed as the cause of events (e.g., the rules and culture of the organization). Whereas supervisor responsibility can be seen as a dispositional attribution, organization responsibility can be described as a situational attribution. An incident was coded as intentional behavior when interviewees used a "reason explanation," such as the beliefs, desires, or values of people (Knobe and Malle 2002). When interviewees referred to causal history explanations (e.g., emotions, bodily states, personality traits, experience, or the rules and responsibilities of the parties), we categorized the action as unintentional.

Stage 2: Affective reactions
In the second stage, we checked for discernible patterns of affective reactions in the sensemaking process triggered by PC evaluations. We classified affective reactions to events into five theoretically plausible themes: no arousal, low arousal-positive, low arousal-negative, high arousal-positive, and high arousal-negative. No arousal means interviewees reported no emotional reaction to the pertinent event. The four other themes align with the core affect model as they capture distinct valence-arousal combinations (Kuppens et al. 2017): positive-valence high-arousal (PH), positivevalence low-arousal (PL), negative-valence high-arousal (NH), and negative-valence low-arousal (NL). For instance, driven, enthusiastic, and excited may be considered as PH; grateful, appreciate, and respectful may belong to PL; angry, hostile, and disgusted may be part of NH; and disappointed, depressed, hurt, sad, and bored may be attributed to NL.

Stage 3: Behavioral reactions
In the final stage, we explored behavioral patterns following PC evaluations. The constant comparative method based on the typology of Skinner et al. (2003) was used to create descriptive categories based on the collected data. Four higher-order themes emerged and were theoretically meaningful: accommodation, negotiation, distraction, and antagonism. Lower-order themes described specific behaviors within each higherorder theme. Overall, seven strategies of coping actions regarding PC evaluations (i.e., lower-order themes) were identified and cover a spectrum of different coping behaviors: acceptance and focusing on the future belong to accommodation; seeking compromise belongs to negotiation; finding a new solution, complaining to others, and obeying belong to distraction; and reviling belongs to antagonism.

Attributions of responsibility
According to Morrison and Robinson's (1997) study, the attribution of responsibility is at the heart of the sensemaking process. The employees we interviewed attempted to determine who was responsible for the incidents they relayed. Out of the 94 incidents we identified, supervisors (both immediate and subordinate supervisors) emerged as the responsible party for the under-to over-fulfillment of the PC in 61 incidents, which is consistent with prior research (Parzefall and Coyle-Shapiro 2011). In contrast, 34 incidents were attributed to situational organizational factors (one incident was attributed to both the supervisor and organization). For example: "It is normal that we give advice [to the clients] because my [new] direct boss doesn't know the clients as well as we know them. In this case, my boss changed the advice without asking me." (Belgian interview #12) Regarding exact-fulfillment incidents, for instance (supervisor attribution): "My grandfather was sick. I want to go home to visit him, but before I spoke my willing to my leader, he told me in advance that the work was completed and that I could go back home. It just fulfilled my own wishes." (Chinese interview #2) As for over-fulfilled incidents, for example (supervisor attribution): "One of my female bosses comforted me that I should not worry about work at this time (take a leave for personal reasons) and she will instruct my colleagues to take over my work." (Belgian interview #12) As Table 1 shows, we found no significant differences in attribution of over-fulfilled, exact-fulfilled, and under-fulfilled events to supervisors or organizations between Belgian and Chinese interviewees. However, attribution clearly differed between overfulfilled, exactly-fulfilled, and under-fulfilled events, χ 2 (2) = 13.70, p = 0.001, with underfulfillment and exact-fulfillment being more likely attributed to the supervisor, and over-fulfillment being more likely attributed to the organization.
Turning to the intentionality themes, we analyzed the reasons provided to us and causal history explanations to categorize incidents as either intentional or unintentional. As explained above, "intentional" implies that the behavior is purposeful and occurs only in a specific event. For example, one interviewee complained about a lower performance mark due to the direct supervisor not giving a promised mark, which caused the interviewee's salary to decrease that year: "One year, in the performance scoring, the leader did not give me the score I had expected, because this score is directly linked to the performance award. He did not give me the expected scores, and this finally led to a decrease in the performance award of that year … We had communicated before, or we could say the score he had promised me was relatively higher than the final score. He might have made promises to a lot of people that he would give very good scores. However, regarding the final score, he has to weigh the entire department, the importance of each position to him, and give a corresponding score." (Chinese interview #3) For the interviewee, this incident was only a one-time event, and the explanation of the incident suggests that the interviewee attributed it to the intentional behavior of the supervisor. In contrast, when people rely on causal history explanations, they typically indicate that the relevant behaviors were out of a person's control and involuntary. Parzefall and Coyle-Shapiro (2011) elaborate that such explanations suggest that the other party "is just like that" and cannot change his or her own behavior. Such unintentional behaviors can have a long history, rather than being a one-time event. Breach events that are caused by unintentional behaviors might be solved by, for example, replacing the supervisor with someone else (Parzefall and Coyle-Shapiro 2011: a trait explanation) or by the employee leaving the organization: "The previous job [as an accounting staff] was really boring. They [the organization] controlled stuff, checked stuff, so they did control of the controls … So, I didn't feel very productive, and that's why I quit." (Belgian interview #5) The above quote shows how the PC under-fulfillment event was not attributed to a specific individual. Instead, the reason for the lack of autonomy and the high amount of control appears to be the culture of the organization, which appears to have been stable and unchangeable. As Table 1 shows, we found no significant differences in intentionality between the Chinese and Belgian interviewees when we consider our three evaluations. Moreover, there were no differences between under-, exact-, and over-fulfillment in terms of attributed intentionality, χ 2 (2) = 1.43, p = 0.49.

Affective reactions
Based on the critical incidents that were imparted to us, we identified five affective reactions to PC under-, exact-and over-fulfillment:

No arousal
Three evaluations elicited no arousal affective reactions, which is somewhat surprising as past research typically associates strong emotional responses with under-and overfulfillment (Conway and Briner 2009 In contrast, one interviewee experienced no arousal when faced with an overfulfillment event: "From the perspective of employees, this situation [over-fulfillment] basically does not exist. If I have to say, maybe there is one thing. The company has a training course for all employees. I participated in some of the lectures and activities. These were about how to deal with your marriage or life balance … I just attended without a special feeling." (Chinese interview #5) Compared with the Chinese interviewees, it is noteworthy that the Belgian interviewees only mentioned no arousal reactions in cases of exact-fulfillment, whereas they always experienced low-or high-arousal emotions following under-or over-fulfillment. As the quotes above show, some Chinese interviewees also experienced no arousal emotions in cases of under-(n = 2) and over-fulfillment (n = 1). These interviewees were possibly experiencing an emotionally numb (Totterdell and Holman 2003), which is associated with surface acting in the process model of emotions (Hochschild 1983) in which employees regulate their emotional expressions to fulfill the duties of their jobs, while hiding their emotions. According to the narrative by these Chinese participants, no arousal emotions were experienced following a long period of breach and they no longer cared about these events, such that they may have become emotionally numb to these circumstances.
When exact-fulfillment incidents mentioned by Belgian (n = 2) and Chinese (n = 5) interviewees elicited no arousal emotions, they described the fulfillment of these obligations as the normal state of affairs-their supervisor or organization was expected to fulfill these obligations and, hence, it came as no surprise and did not affect them.

Low arousal
We distinguished PL from NL. PL reactions were reported even when employers failed to fulfill obligations. One Belgian interviewee recalled a mixture of PL and NL emotions, namely feeling respectful and disappointed, after experiencing PC under-fulfillment. We labeled this type of emotional reaction as co-occurrence, meaning that two opposite emotions appear simultaneously in one incident. For example: "The director asked me to replace my colleague [because he was ill at that time] partly for some projects he was doing. So then I said okay, well I'm already doing two jobs within the organization … I felt good because people trust you and that's the most important feeling, but at the same time, I felt the negative part of the feeling that I cannot do everything, I have two jobs, you asked me to do a third one, actually it is not possible, and I think the employer has an obligation to protect employees as well. I felt a bit annoyed." (Belgian interview #8) Nevertheless, NL emerged as the main affective reaction to under-fulfilled incidents, which is consistent with previous findings that breach triggers negative affective reactions (Zhao et al. 2007). Likewise, PL emerged as the main affective reaction to exactfulfillment. A more complicated picture emerged when interviewees reported affective reactions to over-fulfillment. We again found evidence for the co-occurrence of positively and negatively valenced low-arousal emotions. Evidently, over-fulfillment may not always trigger a positive response, which is consistent with Montes and Irving's (2008) finding that over-fulfillment can also result in negative affective reactions. Interestingly, co-occurrence of positive and negative emotions was again only reported by Belgian respondents.
"There was one staff of the provincial government, and she is retired. They need another person to go to that meeting [the provincial government meeting] … I don't want to go there! I have my own work. [I am learning things step by step.] But maybe it's not negative. I think I need to grow." (Belgian interview #10) For some respondents, over-fulfillment incidents triggered only negative emotions. For three Chinese interviewees, over-fulfillment elicited low-arousal negative emotions such as discontentment and dissatisfaction.
"I was very dissatisfied with the company's scheduled Saturday to let everyone listen to this lecture … I think this is beyond their responsibility." (Chinese interview #12) Du  However, most Belgian interviewees (n Belgium = 14 out of 17) linked over-fulfillment incidents to low-arousal positive emotions such as feeling motivated, happy, warm, excited, and appreciated. Furthermore, due to Chinese respondents describing less critical incidents of over-fulfilled events, compared to under-and exact-fulfilled events, only a couple of Chinese respondents (n China = 2 out of 8) expressed PL in the aftermath of an over-fulfilled PC, for example, "I was pregnant like three months after I started [my new job], which made me feel really uncomfortable, but he [my supervisor] was happy, he wished me the best. It's not a really nice information for your employer because he's going to be without you for several months. However, the way he reacted was just so happy for me and understanding, and this was exactly how to fulfill one of the obligations … He really understood my situation, which was very important to me." (Belgian interview #18)

High-arousal
Finally, we turn to high-arousal reactions, again distinguishing positive from negative valence. Interviewees reported high-arousal affective reactions only related to incidents of under-and over-fulfillment, and never related to exact-fulfillment. Examples of positive high-arousal emotions included excitement and feeling motivated, while examples of negative high-arousal emotions included hostility toward the employer, anxiety, and anger (Kuppens et al. 2017). Our interviews clearly show that participants first experienced low-arousal negative emotions, which then turned into high-arousal negative emotions. This hints at the existence of a sequential process in which the intensity of experienced emotions gradually increases. When negative high-arousal emotions were triggered, respondents became overwhelmingly hostile and cynical toward the leader or the organization. All the negative high-arousal emotions were triggered within PC under-fulfillment and displayed a sequential process. "After I resigned, I received my salary for that month. However, I found that I was short of some money. I went to find my direct leader. He did not reply to me at all. He treated me as follows: 'you had resigned; I will not reply to you regarding anything.' I think this was quite a fraud company, he is totally like a jerk!" (Chinese interview #8) Positive high-arousal emotions were only mentioned related to over-fulfillment, but were rarely expressed by both Belgian (n = 1) and Chinese (n = 2) interviewees.
"I am teaching two courses right now. From next year I have to teach three courses, so I would get an additional course, which is okay, but I am not too happy about it … But a couple of weeks ago, I heard that I don't have to teach the additional course next year … I was more excited that I wasn't thinking about it. I was very surprised." (Belgian interview #1) Comparing the emotional reactions of Belgian and Chinese interviewees, we find no significant differences between both groups (see Table 2). However, looking at the relationship between PC evaluation type and emotional reaction, we find evidence of systematic differences, χ 2 (8) = 93.83, p < 0.001. In particular, no arousal reactions were mostly elicited by exact-fulfilled obligations. Low arousal positive emotions were also mostly associated with exact-fulfilled obligations and to a lesser extent with over-fulfilled obligations. Higharousal positive emotions only followed over-fulfilled obligations, while high-arousal negative emotions only accompanied under-fulfilled obligations.

Behavioral reactions
The coping behaviors that emerged covered four themes, ranging from positive to negative reactions: accommodation, negotiation, distraction, and antagonism.

Accommodation
In these cases, interviewees adapted their perceptions by reassessing PC evaluation incidents (see, Bankins 2015) or by accepting the status quo. They also adapted by focusing on their future employment relationship. Overall, this reaction appeared to be a discretionary attempt by the interviewees to deal with the events in a positive way. Accommodation was an especially relevant coping behavior in case of exact-fulfillment, as all Belgian and Chinese interviewees reported this behavior in all exact-fulfillment incidents.
"I was able to go for a business trip … This was a very happy time in my life. The leader affirmed [I can go for that trip], but I especially thank that leader at the time … I felt very lucky." (Chinese interview #20) Additionally, accommodation also appeared sporadically in under-fulfillment incidents (6.1%) and commonly in over-fulfillment incidents (65.2%). For example: "Because I felt that it was not really well communicated among the team that it was decided I should take up the extra tasks … I think people understood that, but in a context in which some people were already experiencing burnouts. I think it is the employer that has the obligation to take care of the people … It was more like annoyed, I think in general the feeling was okay, I felt respect because he [my supervisor] just asked me, because someone needed to do it [since] this was an important job. I think I'm also a person who likes helping others." (Belgian interview #8)

Negotiation
This type of reaction seemed to be of a neutral nature, as interviewees reported using communication strategies to seek compromises between both parties. We found examples of negotiation reactions in ten Belgian and four Chinese under-fulfillment incidents. However, when negotiation efforts were ignored or rejected by the employer, interviewees reported switching to different coping behaviors. For example: "I didn't tell anyone, because it was really shocking for me. Eventually, I [got up the nerves and] spoke to my supervisor, and I said, 'don't you think it was a bit hard for me [to be humiliated/criticized in public]?' Then my supervisor said, 'no, for people like you, we must do this, that's the only way you can understand.' and I'm thinking, 'how the hell does my supervisor know me? ' … I was very, very angry with my supervisor … I will never forgive [him]. I stopped talking to anyone for several months." (Belgian interview #9) In addition, two Belgian respondents used a negotiation strategy to deal with overfulfillment (see the quote below). Negotiation reactions were not mentioned in reference to exact-fulfillment for both groups, or to over-fulfillment for Chinese interviewees.
"I think maybe I am not experienced enough to go there [the new job assignment].
This was difficult to explain to him [my supervisor] that one will grow, and maybe he's right, you can learn a lot of things, but at the moment, I think I need some experience to go to that meeting. [After I told the supervisor that I did not want to go.] He insisted." (Belgian interview #10)

Distraction
While we also view seeking distraction as a neutral coping behavior, there are quite some differences with negotiation behavior. When employees cope by negotiating, they communicate directly with their employer to discuss issues. The outcome of the negotiation may be a compromise, or it could result in the employee or the employer rejecting the other party's offer. Therefore, negotiation may lead to seeking compromise and accommodation. However, interviewees who reported distraction engaged in an alternative activity to cope with an event, such as getting together with friends (Skinner et al. 2003). This coping behavior tends to avoid facing employers directly and instead attempts to redirect attention. The alternative activities could be anything as long as they distract individuals from their current PC evaluation and, thus, help them to maintain the employment relationship.
Respondents tried to redirect their attention while venting their emotions to friends and colleagues or while just continuing to do their job. Through distraction, they tried to avoid thinking about the PC events. Distraction reactions mainly emerged in relation to under-fulfillment incidents. In addition, it was reported by Chinese respondents in relation to some over-fulfillment incidents. For example: "I had a bad feeling about this [event], and I felt that I was not able to talk about this to him [my supervisor] … There was another boss, and she asked something else about another work of mine, and then … I don't know why I can talk about this [event] to her, and she said to me 'you should go to him and talk to him.' but I told her 'I don't think I'm going to do this because I have this feeling that he is not open for discussion.' and then I guess, she talked about this to him." (Belgian interview #12) Interestingly, when comparing Belgian to Chinese interviewees, distraction reactions reported by Chinese interviewees focused more on redirecting attention while obeying their supervisors (comply with/fulfill supervisor's order), whereas Belgian interviewees sought distraction by complaining to others or by finding a new solution (comply with/fulfill own will), for instance: "When I was an accounting supervisor, our supervisor was strict with us. We started work at 9:00 a.m., but he asked us to arrive at 8:00 a.m. to join a morning meeting, and we got off work at 6:00 p.m., but he asked to stay until 7:00 p.m., which was invisible to make our working hours longer by two hours … I don't dare to talk to him. He will be angry. He will convince me until it makes me feel that I am wrong, because he is the type of person … There is no need to argue. It doesn't matter, I just followed his order." (Chinese interview #6)

Antagonism
Antagonism is the most negative of the emerging themes; respondents became hostile toward their employer, blaming them for the events, and finally exiting the organization or experiencing burnout. Antagonistic reactions only appeared in cases of under-fulfillment incidents. However, all examples of antagonism were preceded by negotiation behaviors, which means that antagonism is not a primary response to under-fulfillment incidents; it only emerges after a process of negotiation remains unsolved and exceeds the tolerance limit of employees.
"My previous boss told me I would have a good position. But I have to wait … but in the end, it was not true … After that the situation was clear [the company assigned me a job], they told me 'thank you very much, but we have no job for you.' They lied to me … They have a lot of company branches, it's all rules … and it was not a good experience for me, because there is no managing … The direct boss told me … Everything was decided by my boss … and I had to leave." (Belgian interview #15) We find no evidence for significant differences in behaviors between Belgian and Chinese interviewees in response to under-and exact-fulfillment (see Table 3); a significant difference for over-fulfillment can be observed. In reaction to over-fulfillment, Belgian interviewees reported using accommodation and negotiation behaviors, whereas Chinese interviewees reported accommodation and distraction behaviors. Jointly considering Belgian and Chinese interviewees, we find evidence for a relationship between PC evaluation type and behavioral response, χ 2 (6) = 80.68, p < 0.001. Exact-fulfillment of obligations only elicited accommodation behaviors, while over-fulfilled obligations elicited mostly accommodation and-to a lesser extent-distraction and negotiation behaviors. Under-fulfillment elicited all four behavioral reactions, with a predominance of distraction and negotiation behaviors.
Finally, to visualize the entire sensemaking process triggered by under-, exact-, and over-fulfillment, we created Sankey flow diagrams (see Fig. 1) for the Belgian (top figure) and Chinese (bottom figure) interviewees' reported incidents. As can be seen, some subtle differences can be observed. First, no arousal appeared to play a larger role in the reported sensemaking processes of the Chinese interviewees, and this was typically associated with attributing the event to unintentional acts by the organization. Second, Belgian interviewees typically reported distraction in response to low-arousal negative emotions, whereas Chinese interviewees also reported this behavior in response to high-arousal negative emotions.

Discussion
Asking people at one point in time to consciously reflect on what was promised to them and what they receive may be a good way to capture stable, inter-individual differences in PC fulfillment and breach (Hofmans 2017). Hence, the aim of this study was to gain a better understanding of the sensemaking process that is triggered by PC under-, exact-, and over-fulfillment, and we used a critical incident technique to explore this. Moreover, we contribute to the literature by exploring cultural differences in these sensemaking processes, comparing Belgian and Chinese employees' accounts of PC evaluations, thereby also filling this gap in previous research (e.g., Woodrow and Guest 2020).

Implications for theory
Comparing under-, exact-, and over-fulfillment Our findings highlight the importance of distinguishing negative from positive discrepancies when studying PC breaches (Lambert et al. 2003;Rousseau et al. 2018). The first key element of the sensemaking process concerns how employees attribute a breach or fulfillment of their PC. In our study, under-fulfillment and exact-fulfillment are more likely attributed to the supervisor, whereas over-fulfillment is more likely attributed to the organization. This highlights the critical role that supervisors play in managing the PC of employees. The finding that over-fulfillment is more likely attributable to the organization is somewhat surprising in this regard but could be due to the nature of the obligations that are reported as being over-fulfilled under the control of other organizational actors, such as the human resource department. It could also reflect a tendency for employees to seek a specific person who is responsible for negative discrepancies, whereas no such need may arise in case of positive discrepancies. Interestingly, the type of PC evaluation of the participants is unrelated to intentionality, suggesting that whether the other party breaches or fulfills the PC is perceived as being due to situational factors. However, our findings do suggest that attributed intentionality plays a role in the emotional and behavioral reactions to PC evaluations. The second key factor in the sensemaking process is the emotional response to the PC evaluation. Our findings clearly show that the full spectrum of emotions should be considered when studying PC evaluations, rather than focusing solely on feelings of violation that are limited to high-arousal negative emotions (Morrison and Robinson 1997). PC evaluations can elicit both positive and negative emotions, with varying degrees of arousal. High-arousal negative emotions typically follow perceptions of underfulfillment, in particular, when these are attributed to intentional behavior by the supervisor (see Fig. 1). In line with the literature, we find that not every breach elicits an emotional response, as under-and over-fulfillment are sometimes associated with no arousal emotions (Morrison and Robinson 1997). Importantly, our qualitative analysis reveals that emotional responses are sometimes complex, with some interviewees reporting the co-occurrence of positive and negative emotions. This finding reveals the complexity of some PC breaches, which cannot always be categorized clearly as being a negative or a positive situation for employees. Theoretically, this may suggest that some breaches can elicit mixed reactions because they can simultaneously help with reaching one goal while thwarting the attainment of another goal (Rousseau et al. 2018). In addition, we also find support for a sequential process in experienced emotions, with some interviewees reporting low-arousal emotions at first, which later aggravates into high-arousal emotions. This finding underscores the dynamic nature of PC processes (Rousseau et al. 2018). In particular, it appears that especially the arousal dimension of affect fluctuates over time, depending on the response of the organization.
The sensemaking process eventually elicits a behavioral response, and we find support for various behaviors that employees use to cope with PC evaluations. Exactlyfulfilled obligations are typically followed with accommodation behavior, which aligns with the notion that fulfilled obligations are effortlessly processed and accommodated into the PC during the maintenance phase (Rousseau et al. 2018). Over-fulfilled obligations are also often associated with accommodation behavior, suggesting that employees may integrate this new information into their PC and potentially even adjust their perception of what the organization owes them (Rousseau et al. 2018). Underfulfillment elicits a broad range of behaviors. Importantly, we find evidence for sequences in behavior, as interviewees report starting with negotiation behaviors, butdepending on the outcome-subsequently switch to accommodation, distraction, or antagonism behavior. Antagonism is clearly associated with high-arousal negative emotions, which suggests that this is the end stage of an escalated conflict .

Comparing Belgian and Chinese employees
In general, we find that Belgian and Chinese interviewees' sensemaking processes triggered by under-, exact-, and over-fulfillment do not differ substantially. However, when zooming in, a number of subtle differences stand out. First, we can observe a difference in behavior in reaction to over-fulfillment. In particular, Belgian interviewees are more likely to use negotiation in response, whereas Chinese interviewees use distraction behaviors. This distinct pattern of behaviors could be explained by cultural differences, as the egalitarian nature of Belgian culture may lead to over-fulfillment being interpreted as a signal that one is valued and, hence, could negotiate for a better deal (Du and Vantilborgh 2020). In contrast, in a hierarchical collectivistic culture, employees may try to avoid over-fulfillment, as they do not want to jeopardize harmony and in-group relations, especially when receiving more inducements compared to others (Peretz and Fried 2012). Second, no arousal reactions appear to play a larger role for Chinese employees. At the same time, no arousal emotions seem to have different meanings under two cultural contexts. We find no arousal emotions are more like emotional numbing for Chinese participants, which are represented by numbness, which is similar to emotional estrangement (Totterdell and Holman 2003). However, Belgian participants who experience no arousal reactions feel that the event is not salient enough to merit an emotional reaction. This suggests that the emotional response to PC breach may be influenced by culture, and in line with that, employees in collectivistic cultures may be aware that their employer under-fulfilled certain obligations but not display any emotional response (Lim 2016) while obeying their superiors. Over time, they may show more emotional numbness toward work and may not have much emotional arousal to the perceived discrepancies. Similarly, this may also be one of the reasons that they report less over-fulfilled events. In addition, it is important to point out that cultural differences may also manifest in how interviewees react during interviews, as the Chinese interviewees may have been less likely to admit feeling low or high-arousal emotions in certain situations. This is consistent with the cultural norm in many Asian societies that discourages experiencing or expressing high-arousal emotions (Lim 2016).
Finally, we also find that both Belgian and Chinese interviewees use distraction as a behavioral response to PC evaluations, but they use distinct distraction strategies. Whereas Belgian interviewees mainly seek distraction by complaining to others, Chinese interviewees try to redirect their attention while obeying their supervisors. This difference may emerge because in collectivist societies, of which China is one, preserving harmony in the in-group is valued meaning that Chinese employees may accept a negative event without contending this situation with their colleagues or supervisor (Huang 2016). In contrast, individualistic societies tend to promote the expression of the individual's emotions, resulting in a higher likelihood of employees venting negative emotions. However, this could potentially increase the likelihood that negative emotions spread across colleagues in individualistic societies (Achnak et al. 2018).

Implications for managerial practice
Our findings have a number of practical implications. First, our results show that PC evaluations can trigger a sequence of emotions and behaviors that gradually deteriorates the exchange relationship. Managers need to be wary of early warning signals so that they can prevent an escalation of conflict. Clearly communicating with employees about mutual obligations and offering appropriate social accounts for events to help employees correctly attribute PC breaches appears to be an important task for managers. Second, we illustrate that cultural differences exist in the sensemaking process of employees following PC evaluations. We find that individualistic interviewees sometimes report complex mixtures of positive and negative emotions in response to PC evaluations, meaning that it may be less clear to managers whether they experience a positive or a negative event. Moreover, Chinese employees tend to experience less over-fulfilled obligations and lower arousal emotions, which might indicate they are accustomed to suppressing themselves to a certain extent, and may have a lower likelihood of actually noticing over-fulfilled obligations. Moreover, we note that Chinese employees were likely to redirect their attention while still obeying their supervisor. This implies that it may be more difficult for managers of Chinese employees to actually notice that the employee perceives a PC breach, as they are less likely to complain about the situation.
Overall, our findings strengthen the notion that management tactics for both individual-level changes and group-level interventions are important. Managers should be trained to carefully manage employees' PC, such as by learning to effectively communicate with employees and through cultural awareness and sensitivity to better understand the values and characteristics of employees.

Limitations and suggestions for future research
Our study has a number of limitations that should be taken into account. First, it is important to highlight that we cannot generalize our findings to all Belgian and Chinese employees. However, obtaining generalizable findings is not the aim of qualitative research (Graneheim and Lundman 2004). Compared to other qualitative studies on PC processes, our sample is relatively large (Bankins 2015;Parzefall and Coyle-Shapiro 2011), and we attempt to purposefully sample employees with diverse ages, work experience, and employment context. We also sample employees from specific regions of each country (i.e., the north of China and the Flemish Region of Belgium) to reduce cultural differences within the sample for each country. While our qualitative results showcase some of the subtle differences between Belgian and Chinese employees' PCs, we recommend future studies to use quantitative approaches with representative samples to statistically test for cultural differences in PC processes. Second, sensemaking is retrospective, which means individuals may not tell a fully accurate account of events (Weick 1995). Interviewees are required to recall at least three events within a short amount of time. As a result, our findings may reflect the most salient or recent events that interviewees have experienced. We try to counter this by giving interviewees enough time to recall as much details as possible. An alternative approach could be to use experience sampling methods to immediately capture breach perceptions as they take place over time.