SALESPERSON DEVIANT BEHAVIOR IN INDONESIAN RESTAURANT SERVICE ATTENDANTS

Workplace deviance has emerged into a well-researched construct within these past 10 years. The purpose of this study is to explain the relationship between abusive supervision and salesperson deviant behavior, moderated by negative reciprocity beliefs (NRB). The respondents were 150 full time service attendants, in the sense that their job description includes dealing with customers and selling menu to the customers. This produces a certain deviant behavior related to the customers. The data was processed using Hierarchical Moderated Regression Analysis using SPSS 23. We found that abusive supervision has direct relationship to salesperson deviance, both positively and significantly. NRB is also revealed to have a moderating effect in the relationship between abusive supervision and salesperson deviance, in which the relationship is stronger when the salesperson endorses NRB rather than not. In conclusion, restaurants may have to pay serious attention to working condition. Not only abusive supervision will affect the restaurant working condition, but also the service perceived by the customers which in the end will lead to financial problems. These service attendants may project their deviant behavior stronger after receiving abusive treatment, when they endorse NRB.


INTRODUCTION
Workplace deviance has emerged into a well-researched construct within these past 10 years (Mitchell and Ambrose, 2007). Some behaviors within workplace deviance include sabotage, theft, and withdrawal (Bibi et al., 2013;Harper, 1990). It can be directed to co-workers, organizations, supervisors, and customers (Bennett and Robinson, 2000;Darrat et al., 2010).
The definition of workplace deviance itself varies, as the terms for describing this behavior also vary. Some researchers such as Cohen et al. 2013), Harper, (1990, and Bibi et al. (2013) labeled it as counterproductive work behavior, while other researchers labeled it as workplace incivility (Lee and Jensen, 2014;Matthews and Ritter, 2016). However, almost all of the previous researchers agree that workplace deviance is voluntary and costly to organizations.
Although workplace deviance as a construct has been examined rather thoroughly, one part of the construct, which is salesperson deviance has not. Only in this past decade that researchers started to pay attention to this construct (Darrat et al., 2010;Hochstein et al., 2015;Jelinek and Ahearne, 2006). Jelinek and Ahearne (2006) stated that workplace deviance should be brought into sales and marketing literature since it has its own characteristics which are quite distinct from other type of works. A salesperson often interacts directly with customer as well as other departments within an organization (Foster and Cadogan, 2000;Hakansson et al., 2004;Hohenschwert, 2012;Mullins et al., 2014). Often, a salesperson has to compete with co-workers to get a customer or a client (Bateman, 2009). Moreover, a salesperson often has to cope with commission-basis pay instead of fixed pay (Chung, 2015). Thus, a salesperson is more likely to produce different kinds of deviance, such as supervisor-directed, co-workers-directed, organizational-directed, and especially customer-directed deviance (Darrat et al., 2010).
Based on the annotation presented, we would like to explore the relationship between abusive supervision and workplace deviant behavior with negative reciprocity beliefs (NRBs) as a moderator. In particular, we would like to further delve into the salesperson deviant behavior. The research framework is Figure 1.
It is expected that this research will add interesting findings in the workplace deviant behavior literature, as we will focus on salespeople, who are front-liners of a company. Furthermore, the result will present more knowledge regarding the three constructs seen from the perspective of Indonesian salespeople, which possess high power distance index (Wulani et al., 2014).

Abusive Supervision and Salesperson Deviance
Past researchers have found strong relationship between abusive supervision and workplace deviance (Mackey et al., 2015;Mitchell and Ambrose, 2007). Hua et al. (2009) explained that abusive supervision victims are more likely to engage in workplace deviant behavior which is directed to their supervisor. Mackey et al. (2015) stated that the victims will show deviant behavior directed to their co-workers. Abusive supervision itself can be in a form of bullying or overusing the supervisors' authority, although this behavior excludes physical contacts (Tepper, 2000).
Although we could not find a study which examined the direct relationship between abusive supervision and salesperson deviant behavior directed to customers, it is explainable logically. People who are abused by their supervisors tend to seek revenge as a form of power restoration mechanism in a workplace (Hua et al., 2009). This includes the salespeople. However, it has to be noted that these salespeople often do not possess adequate power to make any difference regarding the action of revenge towards their supervisors. This is due to the nature of the job itself, in which salespeople are usually working based on contract, and their performance is evaluated based on target (Franco-Santos and Bourne, 2008;Kesari, 2014). This implies that salespeople do not possess any significant power within an organization. Thus, the action of revenge, which has to be directed elsewhere to maintain their power, is directed to the customers who also do not possess any power within the organization according to the salespeople's perception.
Hypothesis 1a: Abusive supervision will be positively related to salesperson deviance. Akremi et al. (2010) suggested that in explaining deviance, some personality traits should be considered. It is actually logical, since deviance is an outcome of an individual's personality. Mitchell and Ambrose (2007) suggested that NRB is an important construct that affects workplace deviance, particularly interpersonal deviance and organizational deviance. Their findings are reinforced by Akremi et al. (2010) suggesting that future workplace deviance researchers should not disregard NRB as a construct defining the behavior. According to Eisenberger et al. (2004), NRB can be defined as the vengeful nature of people being mistreated, or the belief that revenge is normal if they experience negative treatment from another human being.

NRBs as a Moderator in the Relationship between Abusive Supervision and Salesperson Deviance
The salespeople experiencing abusive supervision may project their retaliation in the form of deviant behavior. In Indonesian context, this might not be true. Since the power distance is higher than most western countries (e.g. USA), they may process the abusive supervision as a usual phenomenon because of the thought that supervisors that possess higher position than the salespeople's hierarchical position are allowed to be abusive (Khatri, 2011). However, those who possess NRBs will most likely have the tendency to retaliate against the abusive supervisors (Cropanzano and Mitchell, 2005) because they think that it is normal to have such retaliation behavior (Mitchell and Ambrose, 2007).
Hypothesis 1b: NRBs will moderate the relationship between abusive supervision and salesperson deviance, in which the relationship will be stronger when the salespeople possess high NRBs.

Sampling Procedure and Data Collection
The sample collected for the research purpose is 150 full time service attendants in 10 restaurants in Surabaya, both independent restaurants and hotel-based restaurants. The data was collected with the assistance of 10 surveyors. The human resource or restaurant managers were contacted a month before the data collection period to ensure that the research purpose and method is communicated well to the service attendants. To reduce common method bias problem, we informed both managers and the servers that the participants would be anonymous, as Podsakoff et al. (2003) suggested. To ensure the anonymity of the respondents, one of the researchers showed up during the data collection period while the respondents filled the questionnaires. Later, those will be put inside an envelope and directly given to us.
Out of the 150 questionnaires collected, 22 respondents filled a question which was instructed to be left blank. This indicated that they did not read the instructions carefully and thus, the data was not used as there is a high chance that the respondents filled the questionnaires carelessly. Furthermore, 17 respondents stated that they were still under probation and cannot be considered as full time service attendants at the moment. The resulting sample for this research is 111 respondents, yielding 74% response rate.
We selected only front line service attendants in the sense that these workers can be categorized as salespeople. As the main job for the service attendants is selling the restaurant menus, and further up-sell specific menus instructed by their supervisors, the job fits the salespeople criteria nicely. Furthermore, they are in direct contact with the customers, which provides them the opportunity to display deviant behavior towards customers.

Measures
We measure abusive supervision using 5-item scale developed by Wulani et al. (2014). This scale is a more compact version to the previous abusive supervision scale developed by Tepper (2000). We use that scale because of the fact that this scale is developed to explain abusive supervision in Indonesian setting. The respondents will be asked on their perception of abusive supervision experience from 1 = strongly disagree, to 7 = strongly agree. Some of the questions are "My supervisor says something rude to me" and "My supervisor orders me to do a task that does not make sense." The scale used in measuring salesperson deviance is based on the 4-item scale developed by Jelinek and Ahearne (2006) which is labeled as frontline deviance. We decided to use this scale considering that the questionnaire items are deviant behavior directed to both customers and organization. However, we adjust the wordings of the items to fit the scale into our objects which are restaurants.
Respondents will be asked on how often they conduct these deviant behaviors from 1 = never to 7 = always. The prefix of the question will be "In the past 6 months, have you ever...". Some of the questions are "Complained to friends and family about this restaurant" and "Used deceptive tactics while selling or up-selling the menu.
NRBs will be measured using the scale developed by Eisenberger et al. (2004). Respondents will be asked on how they would perceive their agreement upon the statements in the questionnaire from 1 = strongly disagree to 7 = strongly agree. Some of the questions are "If someone dislikes me, I should dislike them" and "If someone treats me badly, I feel I should treat them even worse."

Validity and Reliability Tests
Prior to conducting the hypothesis test, we conducted the validity and reliability tests to ensure that the tools and the purpose of the measurement tools are fit for this research (Trochim, 2006).
In doing so, we used the Pearson correlation test and computing Cronbach's alpha using SPSS 23 to do the validity and reliability tests respectively. The results of the tests are as follows: Table 1 shows that all the indicators fit their construct well, yielding significant correlation values at the 0.01 level (2-tailed).
The result of the computation of Cronbach's alpha is shown in Table 2. The alpha values for abusive supervision, salesperson deviance, and NRBs are 0.793, 0.84, and 0.764 respectively. All these values are >0.6, stating that the model has good internal consistency reliability (Flynn et al., 1994;Trochim, 2006) and therefore, it is efficient to use the scales in measuring these constructs.

Results of the Hypotheses Test
After conducting the validity and reliability tests, we conducted multi-collinearity diagnostics in order to make sure that the constructs used in this research are not correlated. We used SPSS 23 to compute the VIF values, which are 1.016, 1.142, and 1.076 for abusive supervision, NRB, and its moderating effects respectively. All these values are below 10.00 and therefore it can be concluded that the model does not have multi-collinearity issues (Farrar and Glauber, 2006;Wang et al., 1990).
Inconsistent to previous researches (Bennett and Robinson, 2000;Mitchell and Ambrose, 2007), we found that the mean values for abusive supervision and deviance measure are quite high (Table 3), countering the statement that these constructs are low base-rate phenomenon. The mean values indicate that the data has a normal distribution. Thus, we conducted no additional centering to conduct the model measurement and hypothesis test.
In order to assess the hypotheses, we used hierarchical moderated regression analysis (HMRA). The results of the regression model are  Table 4, showing that abusive supervision has a positive and significant relationship with salesperson deviance. Thus, our results support the first hypothesis. In addition, we found that NRBs as a predictor does not have such relationship to salesperson deviance.
The second hypothesis proposes that NRBs will be a moderating variable of the relationship between abusive supervision and salesperson deviance. The results shown also support this, presenting that abusive supervision and negative reciprocity yield positive and significant relationship to salesperson deviance. To further illustrate our findings, we provide the interaction graphing presented in Figure 2.

The interaction of abusive supervision and
NRBs is based on the values representing +1 and −1 standard deviation from the mean as it is suggested by Mitchell and Ambrose (2007). It is shown that the relationship between abusive supervision and salesperson deviance is stronger when the NRB is high, represented by the steeper slope between the two lines representing the higher and lower values of NRB.

DISCUSSION AND CONCLUSION
In their previous research, Hua et al. (2009) Zellars et al. (2002) explained that abusive supervision discourages employee to display positive behavior and therefore it is reasonable to conclude that salesperson as an employee is also affected by the abusive behavior of this salesperson's supervisors. Furthermore, our findings support Mitchell and Ambrose (2007)'s findings that NRB strengthens the relationship between abusive supervision and workplace deviance behavior in the case of salesperson working in a restaurant, or the service attendants. Another way to state our findings is that a service attendant who possesses high NRBs will tend to display deviant behavior when this attendant receives abusive treatment from the supervisors. Khatri (2011) stated that in a high power distance country such as Indonesia, supervisors have absolute control over their subordinates. This also applies to the case of restaurants. In Indonesia, restaurant supervisors usually have similar job description to their subordinates, with the addition of creating the monthly roster and the daily restaurant report. However, the authority given to these supervisors, especially in the operational decision making process in the absence of their superiors is quite high. While they do not have the authority of producing warning letters, they have high influence towards the managers in issuing such punishment. This is often followed by abusive behavior, especially because their perceived power is considerably high. Paired with the fact that they are in constant contact with their subordinates, the possibility of doing the abusive behavior is higher in the restaurant industry.
As it is suggested by Hua et al. (2009), the subordinates, in this case the restaurant front liners who are the service attendants, will conduct a certain behavior which is aimed to restore their lack of power. In this case, the behavior has deviant nature, and it is most likely directed to the supervisor displaying a high level of abusive behavior (Mitchell and Ambrose, 2007), as it was stated in the descriptive analysis. However, Khatri (2011) explained that in a country with a high power distance level, this usually does not happen, as subordinates do not want to oppose their supervisors. This is due to the fear of being scored low in the performance appraisal (Liew et al., 2011), which later leads to the possibility of punishments, or even lay offs. Thus, the restaurant front-liners need to channel the deviant behavior somewhere else, which are the restaurant itself as an organization, or often their customers.
Supervisors are seen as the model worker in the restaurants. When they display abusive or deviant behavior, the subordinates    will also think that it is normal to do so, as it is suggested by the social learning theory (Bandura, 1973). Tepper (2007) added that abusive behavior of supervisor is a normal occurrence in a country high in power distance level, including Indonesia (Hofstede and Fink, 2007). Thus, the service associates who are exposed to this kind of behavior daily will likely to project deviant behavior themselves, whether it is directed to their supervisors, organizations, or customers.
Our results presented in Table 4 stated that NRB strengthens the relationship between abusive supervision and salesperson deviance. This contradicts Mitchell and Ambrose (2007)'s findings in the sense that the deviant behavior construct we use is different than theirs. While their findings stated that negative reciprocity is only influence the relationship between abusive supervision and the deviant behavior directed to supervisors and not the other type of deviant behaviors, we found that in our case, it does have such influence. The scale we use to measure the deviant behavior actually encompasses the organization-directed deviance and customerdirected deviance. However, for future research, their approach should be considered, especially in creating the comparison between the moderating effect of NRBs in the relationship between abusive supervision and different deviant behaviors.
As it was previously stated, restaurant service attendants will react to the display of abusive supervision. This is strengthened when these attendants possess the NRBs, as they perceive that it is normal to act vengeful when they receive negative treatments. This includes the abusive supervision. However, we also found that this behavior does not have significant relationship to salesperson deviant behavior directly. This is also logical, as NRB has to be triggered before the service attendants display deviant behavior.
To conclude, we found that abusive supervision is an antecedent of salesperson deviant behavior. Mitchell and Ambrose (2007) stated that abusive supervision can stimulate the environment of deviant behavior. Moreover, these service attendants may project their deviant behavior stronger after receiving abusive treatment, when they endorse NRBs. It has to be noted, however, that the deviant behavior is not projected towards the supervisors. Instead, the service attendants display such behavior towards the restaurants they are working for, and their customers. Of course, this brings negative effect to the restaurants as a part of service-based industry that relies on people to provide excellent service.
Restaurants may have to pay serious attention to this phenomenon. Not only abusive supervision will affect the restaurant working condition, but also the service perceived by the customers which in the end will lead to financial problems (Robinson and Greenberg, 1998). Moreover, it can also affect the psychological condition of the subordinates. Although Hua et al. (2009) suggested that the deviant behavior caused by employees may be subjected to their personalities, it is absolutely costly, both time-wise and financially, to assess these personalities one by one. Thus, it has to be noted that various attempts should be administered to reduce abusive supervision. It can be through the application of psychological tests during supervisor recruitment, character-based training, or teamwork activities.
This research has its own limitations. First of all, the data we use is cross-sectional in nature. Therefore, determining the causal effect between the constructs is not possible. The survey itself was conducted in a single time period, exposing this research to common method bias problem. Although, we followed Podsakoff et al. (2003)'s suggestion in ensuring the anonymity of the respondents to reduce common method bias, it is suggested for the future researchers to conduct the data collection in a longitudinal design. Finally, although we translated Tepper (2000)'s definition of abusive supervision into our scale, in which we measure abusive supervision through the subordinates' perceptions, we did not do so regarding the salesperson deviant behavior construct. Thus, it may present our result with the perception bias. Future research should seek to measure salesperson deviant behavior from the perspectives of the supervisors instead of using self-report questionnaires based on the service attendants' perception.