Silence Is not Golden: The Effects of Prohibitive Voice on Customer Citizenship Behaviors

This work analyzes the effects that service employee prohibitive voice behavior has on customer perceptions of the procedural justice received, and the consequent customer citizenship behaviors (CCB). Prohibitive voice expresses the service workers‟ concern about existing or upcoming practices, events, or acts that may harm their service outcome. While this voice has been recognized as constructive and helpful to organizations, its roles regarding customer perceptions and CCB remain unexplored. Beginning with a brief review of the conceptual background of prohibitive voice, procedural justice and CCB, this paper proposes a model to test these relationships empirically. This study contributes to service research by applying the concept of “prohibitive voice” to service encounters. Implications for managing service encounters are also discussed herein.


Introduction
A service encounter or moment of truth occurs whenever a customer interacts directly with any contact person (Crosby, Evans & Cowles, 1990).Recognition of the importance of how service workers" behavior affects customers" perception and behavior has inspired marketers to focus on the maintenance of personal interaction at service encounters (Crosby, et al., 1990;Crosby & Stephens, 1987;Patterson, Razzaque & Terry, 2014).In addition, the relationships between employee behaviors and consumer behaviors are well recognized in service research (Beatson, Lings & Gudergan 2008).For example, studies show that employee behaviors influence customers" evaluation of service quality (Collier & Bienstock, 2006;Rafaeli, Ziklik & Doucet, 2008).Beatson, et al. (2008) find a positive link between employee behaviors (service orientation, pro-social service behaviors) and customer relationship quality.As Crosby, et al. (1990) advise, successful exchange episodes can eventually lead to an enduring service worker-customer relationship, provided they are properly managed, from a customer"s perspective.
According to the service marketing literature, the customers" perception of service quality is one of the most important factors that affect their attitude toward service providers and their degree of loyalty (Ansari and Riasi, 2016).Nevertheless, service quality can be considered a necessary but insufficient condition for ensuring positive results because there might be many other factors that could influence the service worker-customer relationships (Evans & Crosby, 1989).For example, while "first-come-first-served" is the norm in many service encounters, some customers try to cut in line.Usually, this behavior is considered unacceptable and unfair.In this situation, a service worker might remain silent.At that moment, customers face the situation of deciding whether to accept or reject such intrusions.Allon and Hanany (2012) state that customers might use community enforcement to sustain equilibria if the service worker does not grant priority.Thus, there might be a conflict and unhappy atmosphere affecting customers, resulting in a service failure situation.
In the above case scenario, a service worker"s prohibitive voice behavior in granting priority creates well-being for all the other customers, as well as the perception of fairness in service.A service worker might also need to speak up honestly about problems to inform customers, even if this approach might hamper short-term revenue.The advice can reduce uncertainty and create the images of integrity, reliability and fairness.By occupying a position close to that of the customer, the service worker is often best suited to employ the prohibitive voice.Nevertheless, in the service industry, service workers are expected to be polite at all times.With its potential risk of challenging the status quo, a prohibitive voice may bring the risk of being misunderstood, or other undesirable consequences (Morrison & Milliken, 2000); this creates an ethical dilemma in the service encounter.As a result, the employees might choose to be silent (Van Dyne, Ang & Botro, 2003).
Although the prohibitive voice has been recognized as playing the critical role in avoiding potential crises (Liang, Farh & Farh, 2012;Van Dyne, et al., 2003;Wei, Zhang & Chen, 2015), its role in regard to customer perceptions and CCB remains unexplored.Since the prohibitive aspect might potentially upset interpersonal relationships (Stamper & Van Dyne, 2001), it is important to examine whether the employees" prohibitive voice produces a positive outcome for the service organization.Therefore, the purpose of this study is to fill this gap in the literature.Specifically, our research question can be outlined as follows: Research Question: How does the use of prohibitive voice by service employees affect customer citizenship behaviors?
The rest of this paper is structured as follows.It begins with an overview of the relevant literature on prohibitive voice behavior, procedural justice and customer citizenship behaviors, followed by the development of the hypotheses.The "Method" section describes the data and measurements.The "Findings" section presents the results.This paper concludes with a discussion of the results and offers future research directions.

Prohibitive Voice
Studies on organizational behavior identify two different forms of voice behavior by employees: making constructive suggestions and prohibiting behavior (Farh, Zhong & Organ, 2004).Liang, et al. (2012) define these two types of voices as promotive voice and prohibitive voice.The former is defined as attempting to put forward new ideas/opinions in order to improve the overall functioning of the organization.In the workplace, the good intention behind constructive suggestions is easily recognized.As it does not directly threaten any person or the current power structure within the work unit, it carries little risk for the individuals who employ such voice, and is likely to be encouraged and appreciated by the management.In contrast, the prohibitive voice raises a direct challenge to others and to the authority.This voice expresses individuals" concern about existing or upcoming practices, events, or acts that may damage their organization.It involves speaking up about the dysfunctional aspects of work practices, with the intention of stopping them in order to improve the overall functioning of the work unit or organization.Thus, it carries more risk than the promotive voice does (Liang, et al., 2012).Both the promotive voice and prohibitive voice are helpful for organizations.The promotive voice can provide innovative solutions for management.It tends to increase or enhance the positive effects of process gains, while the prohibitive voice can prevent the negative effects of process losses (i.e.factors that reduce productivity).For that reason, both forms of voice are constructive to the organization (Liang, et al., 2012;Van Dyne, et al., 2003).
The service workers" behavior can be separated into in-role and/or extra-role behaviors (Bettencourt & Brown, 1997).Extra role behavior that goes beyond existing role expectations involves actions that attempt to benefit the organization.The promotive voice refers to new ideas or suggestions provided by employees that can directly benefit customers; thus, it is more related to in-role behavior.The prohibitive voice is used to caution against undesirable situations, or to advise customers against undesirable behaviors; thus, it is more related to extra-role behavior.Studies have found that service company employees" extra-role behavior (i.e.organizational citizenship behavior, OCB) can affect customer perceptions of service quality and company performance.It should be noted that the prohibitive voice can also be conceptualized as a form of OCB (Liang, et al., 2012) albeit emphasizing different features.

Procedural Justice
Procedural justice theory suggests that in making justice evaluations, people focus on how decisions are made, as well as on the decisions themselves (Lind & Tyler, 1988).People"s reactions or behavior have also been linked to procedural justice.For example, studies on organizational behavior have found a positive relationship between procedural justice and OCB (Tyler & Smith, 1999).The reason for this positive relationship is that people are very sensitive to signals that indicate whether or not they are valued.When employees perceive the procedures of their organization as being fair, they feel they are valued members; as a result, they are motivated to involve themselves in OCB (Tyler, 1999).
In service research, procedural justice focuses on the way that outcomes are reached, including the tools, policies and procedures that firms use to support their communication with customers (Davidow, 2003); it also involves the perceived fairness of how failures are resolved 200).Procedural justice has been recognized as a powerful predictor of customer response when customers experience service failure (Gelbrich & Roschk, 2011).
Conceptually, customer behaviors can be separated into in-role and/or extra-role behaviors (Ahearne, et al., 2005).In-role behaviors can be defined as customer coproductive behaviors, which are those required to perform successful production and/or delivery of the service (e.g., providing the right address/information for a purchase delivery).Extra-role behaviors involve sacrifices on the customers" time or effort (e.g., proactive communication of anticipated problems, recruiting other customers, service improvement suggestions, and positive word-of-mouth); these are collectively referred to as customer citizenship behaviors (Gruen, 1995;Ahearne, et al., 2005;Bettencourt, 1997).

Hypotheses Development
In a service context, procedural justice is considered as customers" subjective evaluations of the justice of service procedures, that is, whether they are fair or unfair, ethical or unethical (Tyler, 1999).Depending upon how customers perceive their procedural fairness, service results may be evaluated differently (Blodgett et al., 1997;Gelbrich & Roschk, 2011).This view indicates the positive relationship between customers" procedural justice evaluation and service workers" behavior.
Although employing the prohibitive voice can be risky behavior (i.e. because of its potential for generating a negative image), it is important for the perception of justice (Schweiger & DeNisi, 1991) and thus is required.In order to enhance the efficacy of service delivery, service providers should communicate with their customers with accurate, timely and helpful information, even if the results might not serve the best interests of the service organization.In this study, we posit that the potential benefit of the prohibitive voice can be categorized according to two effects: (1) it provides useful and timely information reduces information asymmetry, and helps customers make quality decisions, and (2) it enhances the justice perception when other customers" unsuitable behavior occurs.These advantages create the procedural justice perception.Therefore, we posit that: H1: Customers" procedural justice evaluation is positively associated with the service worker"s prohibitive voice behavior.
CCB involves voluntary actions by customers.Equity theory predicts that a person will try to resolve conditions of unfairness because this feeling will create tension within the person (Adams, 1965).Service worker"s prohibitive voice behavior can be considered as input into an individual"s equity ratio.When a service worker takes the initiative to inform his/her customer about a potential factor that might affect the service quality, the customer might perceive this service provider as a person with integrity and trustworthiness.When the service provider takes the initiative to prevent other customers" improper behavior, customers might perceive that this service organization values fairness in service delivery.Such behavior could enhance a good service worker-customer relationship and customer-company identification, and thus reinforce CCB (Ahearne, et al., 2005).Therefore, we suggest that: H2: CCB is positively associated with the service worker"s prohibitive voice behavior.Tansky (1993) states that OCB is one form of people"s extra-role behavior that intends to reward the organization that benefits them.Procedural justice is one type of benefit; people will feel obliged to reciprocate by providing something of value in return if they perceive that they are treated fairly (Masterson, 2001).Studies have reported a robust relationship between perceptions of procedural justice and OCB (e.g., Moorman, Blakely & Niehoff, 1998;Podsakoff, MacKenzie, Paine & Bachrach, 2000).
As with employees, procedural justice could be considered as input into customers" equity ratio.Customers will feel that they are valued clients if they are fairly treated.Consequently, they will exhibit loyalty to their service company and want to engage in voluntary actions to support this company, enhancing the incidence of CCB.Social exchange theory can help to explain this phenomenon, as it predicts that people seek to reciprocate to those who benefit them (Tansky, 1993).When customers perceive that the employees/organization exhibit integrity and justice, they will feel obliged to reciprocate by providing something of worth in return.Therefore, we posit the third hypothesis H3.H3: CCB is positively associated with procedural justice.MacNeil (1985) distinguishes two types of contracts: exchange contracts and relational (psychological) contracts.Exchange contracts are short-term in nature, and based on transactions (Konovsky & Pugh, 1994).Relational contracts, which go beyond economic exchange, are based on social exchange and relationship quality.Social exchange, i.e. relational contracts, can be used as explanations for the occurrence of CCB.Social exchange theory predicts that people will seek to reciprocate to those who benefit them.While service providers" prohibitive voice behavior might create a feeling of fairness and trustworthiness, speaking up might sometimes create the negative feeling of unfriendliness.It is hypothesized that the success of a service provider"s prohibitive voice behavior on customer CCB depends on the enhanced feeling of procedural justice.Hence, the fourth hypothesis is formulated as follows: H4: Procedural justice mediates the relationship between service workers" prohibitive voice behavior and CCB.

Target Sample
We conducted a visitor"s survey to test our hypothesized model.We contacted the store managers of ten franchised beverage shops in Taiwan.These beverage shops promise higher quality beverage service.We briefly explained our research purposes, and promised to keep the responses confidential and to provide them with a research report.Customers at these beverage shops were recruited as the samples.Since selected beverage shops differ in their service attributes, customers may experience different service features in each selected shop.Accordingly, it had good potential to generalize the findings from the samples to the franchised beverage shop industry.Each customer was approached and given consent to participate in the study in public areas of the stores.A total of 450 customers completed the questionnaires.Eliminating uncompleted responses among the questionnaires, 441 responses were useable for data analysis, achieving a response rate of 99%.Among the survey participants, 45.1% were male, and 54.9% were female.
Respondents" ages ranged from 18 to over 45 years old.Most respondents were from the middle age range (25-35 years of age, 49.4%).About 19% of survey participants were 18-24 years of age, and 19% were over 36 years.Regarding the educational level, most of the respondents had a bachelor/college degree (71.5%).High school (17.2%) and graduate school (9.8%) occupy smaller proportions.Approximately 29.5% of the respondents indicated that they come to the shop daily; 34.5% of respondents indicated that they come to the beverage shop 1-5 times per week.The cumulative proportion of these two groups is up to 64%, which is more than half of the entire sample.

Measures
1. Demographics.Age, gender, education level and frequency of purchase were assessed to evaluate the characteristics of the respondents.
2. Prohibitive voice.Prohibitive voice measures the customers" perception regarding how a service worker performs his/her prohibitive voice behavior.It was measured using four questions on a 7-point Likert scale (1= strongly disagree to 7= strongly agree).The measures were modified from Liang, et al. (2012).Customers were asked to indicate whether they agreed that the employee: (1) can advise other customers against undesirable behaviors that would hamper service delivery, (2) can explain opinions on things that might affect efficiency in the service delivery, even if that would embarrass others, (3) can point out problems when they appear in the service encounter, even if that would hamper relationships with customers, and (4) can proactively notify customers about problems in the service encounter.
3. Procedural justice.Procedural justice measures the customer perception regarding whether employees can look after things when the unexpected happens, in order to restore the unfair, or failure of, service (Colquitt, 2001;Karande, Magnini & Tam, 2007;Saxby, Tat & Johansen, 2000); it was measured using three questions on a 7-point Likert scale.Customers were asked to indicate whether they agreed that: (1) the store associate was willing to handle their service procedures to satisfy my needs, (2) the procedures used gave me more control over how well the service problem or failure was solved, and (3) overall, the procedures used by the associate were fair.4. Customer citizenship behavior.CCB is a concept including three dimensions: positive word of mouth, helping other customers and providing feedback (Groth, 2005).It was measured using a 7-point Likert scale.Measures of positive word-of-mouth were adopted from Bove, Pervan, Beatty & Shiu (2009).Customers were asked to indicate whether they: (1) would encourage friends and relatives to go to this store, (2) have actually recommended this store to others, (3) will recommend this store to people interested in the store" products/services.Measures of helping other customers were adopted from Bove et al., (2009) and Groth (2005).Customers were asked to indicate whether they would: (1) assist other customers during service, (2) help others with their shopping, and (3) teach someone how to use the service correctly.Measures of providing feedback were also adopted from Groth (2005).Customers were asked to indicate whether they would: (1) fill out a customer satisfaction survey, (2) provide helpful feedback to customer service, (3) provide information when surveyed by the business, and (4) inform the business about the great service received from an individual employee.
A pilot survey was conducted to evaluate the reliability of the items in the questionnaire; 51 respondents were invited to take the pretest.The results of pilot survey demonstrated that all of the constructs had good reliability.

Internal Consistency and Discriminant Validity
Confirmatory factor analysis (CFA) was performed to assess the construct reliability and construct validity.Table 1 reports the results from the CFA.CCB is the concept measured by three dimensions: positive word of mouth, helping other customers and providing feedback.This construct is modeled as a second order CFA.With regard to the sufficient sample to test model, both the 0.05 and 0.01 CN values for our hypothesized model were greater than 200 (200 and 218, respectively).Accordingly, the size of our sample was adequate.In addition, all values of average variance extracted (AVE) are greater than 0.5 (Bagozzi & Yi, 1988), ranging from 0.53 to 0.84, which indicates high reliability.All factor loadings are significant with p-value less than 0.0001.All of the standardized loading estimates are greater than 0.5; thus, convergent validity is acceptable.In Table 2, the square root of the AVE for each construct is greater than the correlations involving the construct, thus discriminant validity is supported.
We adopted some procedures suggested by Podsakoff, MacKenzie, Lee and Podsakoff (2003) to minimize the possibility of common method variance (CMV).First, the survey was conducted on an anonymous basis.Second, questions of different constructs were printed on different pages to create a psychological effect of subject separation.
Harman"s one-factor test was employed to test the potential for common method bias (Podsakoff & Organ, 1986).All measures were analyzed by factor analysis, which yielded three factors with eigenvalues greater than 1, and the cumulative loading was 72.03%.In addition, the study employed CFA to test the fit of a one-factor model (in which all items were loaded on a single factor) and a five-factor model.The results revealed that the fit of the one-factor model was inferior to that of the five-factor model (Chi-square = 2109.909,df = 119, GFI = 0.566, AGFI = 0.443, CFI = 0.679 and RMSEA = 0.116).The results confirmed that CMV was not a major problem in the study (Podsakoff et al., 2003).

Model Fit
Structural equation modeling (SEM) was performed to test the hypotheses.Customer OCB is modeled as a second order CFA.Prohibitive voice behavior and procedural justice are modeled as first order CFA. Figure 1 shows the result of standardized parameter estimates obtained from the SEM analysis.With regard to the model fit, the value of Chi-square to degree of freedom is 3.201.Chi-square =361.710(P<0.000),GFI=0.912,AGFI=0.881,CFI= 0.960, NFI=0.943 and SRMR=0.0524.The GFI, CFI and NFI are greater than 0.9, which indicates a good fit between data and model.

Hypothesis Testing
Table 3 summarizes the coefficients of hypothesis testing regarding H1 to H3.In H1, it is hypothesized that customer" procedural justice evaluation is positively associated with the service worker"s prohibitive voice behavior.In Table 3, the coefficients is 0.299 (p<0.001).Therefore, H1 is supported.In H2, it is hypothesized that CCB is positively associated with the service worker"s prohibitive voice behavior.In Table 3, the coefficients is 0.274 (p<0.001).Therefore, H2 is supported.In H3, it is hypothesized that CCB is positively associated with the procedural justice.In Table 3, the coefficients is 0.636 (p<0.001).Therefore, H3 is supported.To test the mediation of hypothesis H4, Both the Bootstrap and Sobel test were carried out.Table 4 shows the result of the Bootstrap analysis using the Bias-corrected percentile method, with 5000 Bootstrap samples.In Table 4, the lower and upper values of 95% confidence interval of Prohibitive voice→Procedural justice coefficient are 0.162 and 0.444; the 95% confidence interval of Prohibitive voice→CCB coefficient are 0.177 and 0.379; and the 95% confidence interval of Procedural justice→CCB coefficient are 0.539 and 0.726.All of these values are positive and differ from zero.The results demonstrate that the mediation effect is reliable.

Discussion and Conclusion
This study was performed to explore the effect of the prohibitive voice of service workers on customer citizenship behaviors.The four hypotheses were supported by the analysis.These findings help us to understand the relationships among employee"s prohibitive voice behavior on customer"s procedural justice perception and CCB behavior.Thus, this study enables a better understanding of consumers" justice perception and volunteer behavior.Three conclusions are noteworthy and deserve more attention.
First, the direct effects of service workers" prohibitive voice behavior on customers" perception of procedural justice (H1) and CCB behavior (H2) are supported.The results demonstrate that employees" prohibitive voice behavior could have a positive effect on both fairness perception and CCB behavior.This is important because CCB behaviors are the desired voluntary behaviors (e.g., positive word of mouth) and are beneficial to service organization (Organ, 1988;Bove et al., 2009).
Second, consistent with social exchange theory (Tansky, 1993), the direct effects of customers" perception of procedural justice on their CCB behavior (H3) is supported.This implies that CCB behavior can be enhanced when people perceive that the procedural justice of service is at a high level.This result attests that customers are more willing to exhibit citizenship behavior to reciprocate to those who are more just.
Third, customers" perception of procedural justice serves as the mediation role regarding the effect of service workers" prohibitive voice behavior on CCB.This result is consistent with the relational contracts theory (MacNeil, 1985).Service encounter involves a highly dynamic environment wherein service quality can be influenced by many unexpected internal (e.g., lack of space, materials and employees) and external factors (e.g., customers who jump the queue, or lack of time).When something unexpected happens, service workers should inform the customers so that the customers can make better decisions.While purchasing a beverage in a shop might be regarded as an exchange contract regarding the purchased product, the prohibitive voice of the service worker and the perception of employees" procedural justice can serve as a relational/psychological contract between employee and customer.The relational contract will have a long term effect in nature that enhances CCB.Prohibitive voice behavior from employees enables customers to know the service procedures, and helps customers understand the situation so as to reduce information asymmetric and help customers" decision-making.Customers, therefore, think they are being treated by fair procedures so that they are more willing to provide citizenship behaviors.
There are two theoretical contributions of this paper: first, it confirms that the organization can benefit from employee"s prohibitive voice behavior (Van Dyne, et al., 2003;Liang et al., 2012).The prohibitive voice of a service worker is also proved to have a positive role in service organization.Our findings extend the theory of employee"s prohibitive voice behavior in that we investigated this issue regarding the interpersonal relationship between service worker and customer.Prohibitive voice, recognized as playing the critical role in avoiding potential crises, and its role regarding CCB is supported.Second, it confirms the important role of procedural justice perception on citizenship behavior (e.g., Moorman, 1991;Podsakoff et al., 2000).This study extends the OCB theory from organization behavior to boundary spending behavior, which helps to illuminate how employee behaviors contribute to the desired customer behaviors.
This study offers strong managerial implications.As mentioned in the literature review, the prohibitive voice and promotive voice are two forms of voice.While the promotive voice has usually been of interest because of its advantages, the prohibitive voice has been neglected.Referring to the view of the prohibitive voice by Van Dyne, et al. (2003) and Liang, et al., (2012), the results of this research provide evidence that service workers" prohibitive voice can serve as a positive force for an organization through its positive effect on CCB, which is an interesting standpoint.Based on our findings, we recommend employees to employ prohibitive voice behavior when needed during their service.Furthermore, the success of the prohibitive voice behavior on CCB depends on the level of procedural justice perception being enhanced.Since the prohibitive voice might enhance the perception of unkindness, the instruction should consider the customer"s standpoint: whether this voice helps customers to make a better decision, or enhances procedural justice.For marketing managers, the goal is to cultivate a friendly environment to increase employees" prohibitive voice behavior and customers" procedural justice perception.Knowledge gained in this research can allow managers to succeed in those behaviors that their service workers and customers may engage in.
Finally, we would like to indicate some limitations of this article.First, the first limitation relates to the particular service context of the beverage shop.A useful extension would be to to test the effects of service worker" voice behaviors in different service contexts.For example, prohibitive voice behavior might be more important for those requiring high-speed delivery service industries (e.g, the airport counter, fast food restaurants, amusement park, etc.), and less important for those service industries who require high-speed delivery less (e.g., consultant).Future research could explore this framework in relation to different service types.Second, this research identified service workers" prohibitive behavior as an important antecedent for procedural justice and CCB.Future research could investigate the service workers" prohibitive behavior and promotive voice behavior simultaneously, and explore their antecedents and outcomes.Despite these limitations, this study made some strong theoretical and practical contributions.The current findings provide practitioners with a view that service organization can benefit from service workers" prohibitive voice behavior.This study is presented in the hope that it will stimulate additional research aimed at understanding how service workers affect their customers.

Table 1 .
(Kline, 1998)atory Assessment of the Measures good model should have good fit and ensure the construct reliability and construct validity.A Chi-square test can be applied to test whether there it is a good fit between data and model.A p-value greater than 0.05 is required.However, it should be noted that the Chi-square is very sensitive to sample size.When the sample size is large (i.e.400 or more), the Chi-square is almost always statistically significant.Since the sample size of this research is more than 400, we should refer to other indices to assess the fitness of the model.A value of Chi-square/df less than 3 but greater than 1 is considered as good(Kline, 1998).The GFI=0.926, NFI=0.953,CFI=0.970,SRMR 0.0373 indicate a good fit.All composite reliability (CR) values are greater than 0.8.These values range from 0.81 to 0.94, indicating high reliability. A

Table 2 .
Correlation Estimates, Means, SDs and Discriminant Validity Customer citizenship behavior is the concept measured by three dimensions: positive word of mouth, helping others customers and providing feedback.The diagonal elements are square root of the AVE.

Table 4 .
(MacKinnon, Warsi & Dwyer, 1995)ctThe results of the Sobel test are shown in Table 5.It should be noted that there are three versions of the Sobel test: Sobel test equation, Aroian test equation and Goodman test equation.The Sobel test and the Aroian test seem to perform best(MacKinnon, Warsi & Dwyer, 1995).The Sobel test equation has z-Statistics = 3.51 (p-value < 0.0001), and the other two tests are statistically significant.These results demonstrate that the mediation effect is statistically different from zero.Based on the results of the Bootstrap and Sobel test, we concluded that the mediation of procedural justice on employee" prohibitive voice on the CCB (H4) is supported.Note.a indicates raw regression coefficient between independent variable and mediator; b indicates raw coefficient between mediator and dependent variable.S a is standard error of a, and S b is standard error of b.