How do cynical employees serve their customers? A multi-method study

ABSTRACT

In this multi-method study, we investigate how social job demands (i.e., social interruptions) and resources (i.e., colleague support) in the service context influence employee (negative) (re)actions to customers through cynicism towards the job. In addition, we investigate why customers are less satisfied with the provided service when employees endorse a cynical attitude. To test the hypothesized process, we used observer ratings of the employee–customer interactions regarding the number of interruptions and employee negative (re)actions during service encounters, employee self-reports of overall colleague support and daily cynicism, and customer-ratings of service quality. Participants were 48 service employees and 141 customers. Results of multi-level structural equation modelling analyses showed that whereas the number of observed social interruptions during service encounters related positively to cynicism, social support related negatively. Cynical employees exhibited more negative (re)actions towards their customers (e.g., expressed tension, were unfriendly). Consequently, the more negative (re)actions employees showed towards their customers, the less satisfied customers were with the service quality. The study contributes to the literature by explaining what makes service employees cynical about their work, and why cynical employees provide low-quality services.

KEYWORDS:

• Cynicism
• interruptions
• negative reactions
• service quality
• social support

Service industries dominate modern economies around the globe. Virtually, all humans consume services and about 70% of the employees in Western countries are employed in service jobs (Hülsheger & Schewe, 2011). Ever since the seminal work of Hochschild (1983) on service occupations, two major approaches have emerged. On the one hand, the literature on work psychology and management has focused on the stressors inherent in service work, as well as their outcomes for employees (e.g., Dormann & Zapf, 2004; Grandey, Dickter, & Sin, 2004). This literature has largely ignored customers’ reactions and their satisfaction with the provided services. On the other hand, the service management literature focused particularly on customer evaluations of the quality of service work and has largely ignored the working conditions under which service work is performed and how service providers function (e.g., Davidow, 2003; Parasuraman, Zeithaml, & Berry, 1988). As company profitability and employee functioning may be potentially competing goals, scholars are urged to address service work from both the customer and the employee perspectives simultaneously (Bradley, McColl-Kennedy, Sparks, Jimmieson, & Zapf, 2010; Macintosh & Stevens, 2007).

Some studies have linked organizational characteristics and customer satisfaction through either employee well-being indicators (Moliner, Martínez-Tur, Ramos, Peiró, & Cropanzano, 2008) or relational service competence (Gracia, Salanova, Grau, & Cifre, 2013). However, these studies used only subjective or unit-level data, thus failing to provide insights into the service encounter – which ideally requires data from both employees and their own customers. In this study, we adopt a multi-level and multi-method perspective by studying customers nested in service employees in an attempt to investigate how employees’ work-related experiences relate to customers’ evaluations of service quality through the observed reactions of employees during the service encounter. By focusing on employee–customer encounters, we respond to Ashforth and Humphrey’s (1993) call to zoom in on the micro-level of service interactions.

Social information processing (SIP) theory (Salancik & Pfeffer, 1978) and the social interaction model (Côté, 2005) provide the basis to argue that social job demands and the lack of social job resources influence service quality (as rated by customers) unfavourably through the activation of a callous attitude of employees towards their customers (i.e., cynicism) and the enactment of this attitude in the form of negative (re)actions during service encounters. Because service encounters are social in nature and in line with SIP theory, we focus on two social job characteristics: (1) the number of observed interruptions during service encounters, which is a social job demand, and (2) colleague support, which represents a social job resource (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). Whereas social interruptions concern the temporary suspension of a person’s goal-directed action during service encounters (Baethge, Rogotti, & Roe, 2015), jobs high on social support elicit pleasant emotions and facilitate understanding employees’ emotions, which makes it relevant for the management of emotions inherent in service encounters (Côté, 2005).

Contrary to most studies that focused on positive encounters between employees and their customers (Barger & Grandey, 2006; Tsai, 2001), we focus on negative aspects of the encounters and the display of negative (re)actions of employees to customers. These negative (re)actions usually concern a violation of the display rules set by the organization and are referred to as emotional deviance (Rafaeli & Sutton, 1987; Thoits, 1990). Neglecting the role of negative (re)actions in service encounters is a significant gap in the literature because research suggests that not all employees are willing or capable of showing the organizationally required emotions and act accordingly (Pugh, Groth, & Hennig-Thurau, 2011). Richard and Converse (2016) also call for more research to understand the micro-level processes that drive emotional deviance at the work-event level.

We suggest that negative encounters and the associated display of negative reactions may result from high employee cynicism, which represents a state of disengagement and loss of concern that is manifested in the form of negative attitudes towards the recipients of one’s work (Schaufeli, Leiter, Maslach, & Jackson, 1996). In the service context, studying cynicism is relevant because it represents a way to protect oneself from being overly involved in customers and to minimize the discrepancy between felt and shown emotions (Ashforth & Humphrey, 1993). Thus, feeling cynical on a day at work is likely to result in emotional deviance in the form of negative (re)actions towards customers. As Thoits (1990) argued, understanding the causes and consequences of emotional deviance is of theoretical importance because it helps explain what makes individuals, who as social actors usually aim for social approval, to act inappropriately and what are the outcomes of these inappropriate actions. Although there is research on the predictors of cynicism within the service context, we lack knowledge on the predictors of daily cynicism and even more on objective outcomes, such as customer-rated service quality.

The study contributes to the literature by uncovering the process through which social job demands and resources influence employee (negative) attitudes (i.e., cynicism) and (re)actions during encounters with customers and ultimately service quality by focusing on the situational (micro) and social level (cf. Ashforth & Humphrey, 1993). Investigating the immediate effects of emotional deviance (i.e., negative (re)actions) due to state (daily) cynicism on service quality contributes to understanding and refining the mechanisms that link emotional labour to potential outcomes (Tschan, Rochat, & Zapf, 2005). To unravel this social process, we use data from three sources of information: observer ratings, employee self-reports, and customer ratings, and focus both on general and daily job characteristics (social support and social interruptions, respectively), as well as state cynicism and the accompanying negative (re)actions that are crucial for employee–customer service encounters (Ashforth & Humphrey, 1993; Dormann & Zapf, 2004).

Service context and employee (re)actions

Emotional labour has been described as the requirement to display specific emotions at work and to manage one’s own emotions to adhere to the organizationally required emotional displays (Grandey, Diefendorff, & Rupp, 2013; Hochschild, 1983; Rafaeli & Sutton, 1987; Zapf, 2002). Thus, the quality of service encounters between employees and their customers is central in emotional labour (Zapf, 2002). Inherent in the definition of emotional labour is the requirement from employees to invest effort in dealing with the demands that relate to the interaction with customers to create or maintain a pleasant encounter (Dormann & Zapf, 2004).

In his social interaction model, Côté (2005) explains how the service employee’s display of emotions affects how the customer responds in his/her display of emotions and, ultimately, how this response may feed back and affect the service employee’s emotional state. Accordingly, employees’ (i.e., senders’) public displays of emotions communicate rich and valuable information during social encounters to customers (i.e., receivers) about their goals, intentions and attitudes. As such, senders’ emotion regulation influences their own public displays of emotion and, consequently, receivers’ responses in the form of behaviour that, consequently, determine senders’ outcomes. Whereas displays of emotions like happiness and embarrassment elicit favourable response from receivers, display of emotions like anger or contempt elicit an adverse response as they indicate that the employee will not accommodate the receiver (Côté, 2005), thus initiating a vicious circle of negative reactions and emotions. The social interaction model clearly illustrates the main mechanisms proposed in this study: what happens in the service context determines employees’ attitudes and reactions that, in turn, affect customers’ behaviour.

Although the consequences of emotional labour depend on the emotion-regulation strategies that employees use (see meta-analysis, Hülsheger & Schewe, 2011), it is generally found to cause unfavourable outcomes for employees since it has been related to higher levels of exhaustion, i.e., the stress-related component of job burnout (see meta-analysis, Kenworthy, Fay, Frame, & Petree, 2014) and cynicism, i.e., the attitudinal or motivational component of burnout (e.g., Castanheira & Chambel, 2010). The requirement to follow organization-set emotional displays to fulfil role expectations is generally demanding for employees – particularly when employees use surface acting (Grandey et al., 2013). Hochschild (1983) recognized that the problem with emotional labour is not only that it is energy-depleting, but also that it creates a feeling of deteriorating self-respect for employees and a resentment over the roles one has to play at work. The latter states, in turn, lead to feelings of inauthenticity, cynicism, and recognition that our desire for respect must be subordinated to the desire of others (Godwyn, 2006). In line with this assumption, examining cynicism in relation to service encounters advances our understanding of the interplay between employees and customers by recognizing that emotional labour can influence beliefs, affect and behavioural intentions reflecting this attitude (Wilkerson, Evans, & Davis, 2008). In this way, we add to previous studies that have mainly focused on emotional labour as an energy-depleting process and focused on its relation with exhaustion (e.g., Grandey, 2003; Tschan et al., 2005).

Cynicism is suggested to occur in stressful interactions with customers that make employees feel emotionally overextended and physically drained (Abraham, 2000). Cynicism represents a coping strategy for thwarted competence, causing individuals to distance themselves from their customers. It is characterized by emotional numbness, detachment, callousness (Cordes & Dougherty, 1993), and a lack of caring (Leiter, 1988). ‘Cynicism takes the form of disgust at the hypocrisy of maintaining a separation between the inauthentic self, which acts to portray commercialized feeling, and the true self, which must be suppressed’ (Abraham, 2000, pp. 273–274). Researchers have treated cynicism as a ‘learned attitude that is an outcome of exposure to disillusioning, unfair organizational events’ (Wilkerson et al., 2008, p. 2274).

Despite dominant approaches that view cynicism as an attitude that develops over time, recent diary studies have shown that cynicism levels vary within the same employee from 1 day to another (e.g., Van Gelderen, Bakker, Konijn, & Binnewies, 2014). According to Côté’s (2005) social interaction model, the state of cynicism can be externalized to customers as negative emotional (re)actions, which serve a signal function. This is because cynicism is expressed destructively (e.g., uncooperative) and actively (e.g., employing voice but in a disparaging way) (Wilkerson et al., 2008). Empirical evidence showed that employee cynicism associates negatively with organizational commitment (Andersson & Bateman, 1997), organizational citizenship behaviour (Abraham, 2000; Bedeian, 2007), job satisfaction (Abraham, 2000), and positively with grievance (Wanous, Reichers, & Austin, 2000), badmouthing (Wilkerson et al., 2008), and employee turnover (Bedeian, 2007). Therefore, it is highly relevant to investigate its role in employee–customer service encounters (Ashforth & Humphrey, 1993; Dormann & Zapf, 2004).

Social demands and resources

To examine the predictors of daily cynicism in the service context, we draw on SIP theory (Salancik & Pfeffer, 1978), which states that individuals develop attitudes as a function of the social information available to them. Work-related attitudes are not given but socially constructed. Salancik and Pfeffer suggest that ‘people learn what their needs, values, and requirements should be in part from their interactions with others’ (p. 230). The social context has two general effects on attitude and need statements: (1) it provides a direct construction of meaning through communicating socially acceptable beliefs, attitudes and needs, and acceptable reasons for action, and (2) it shifts an individual’s attention to certain information, making that information more salient and providing expectations concerning individual behaviour.

Applying SIP theory (Salancik & Pfeffer, 1978) to the social context of service encounters suggests that employees interpret the information that they receive by other people (e.g., colleagues, supervisors, customers) and depending on this interpretation, they form negative or positive attitudes towards their work. This information may be communicated through trusting relationships that entail receiving support from those who interact directly with them (Bandura, 1986), which functions as a coping resource (Edwards & Cooper, 1988). For instance, advice from a more experienced colleague on how to manage demanding customers or how to maintain smooth interactions with customers may help employees avoid cynical attitudes towards their customers. In contrast, when such support is lacking, employees will be more likely to develop callous feelings towards their customers and become cynical. Indeed, social support was found to relate negatively to cynicism in service employees working in home care (Xanthopoulou et al., 2007).

Whereas social support may influence employee attitudes in a favourable way, interruptions may influence attitudes in an unfavourable way. During a working day, employees are confronted with several external interruptions that originate from calls by other people (i.e., clients, colleagues, supervisors, or noise) and that are usually unintended and uncontrollable (Jett & George, 2003). Given that these interruptions are caused by social actors and postpone the activity at hand, most interruptions are social in nature. Although social interruptions, under certain conditions, may have positive effects (e.g., stimulation during a boring task; Zijlstra, Roe, Leonora, & Krediet, 1999), in the specific context of service work, interruptions caused by other individuals are more likely to interfere with goal-attainment, thus leading to distancing and cynicism towards customers (e.g., Lee & Ashforth, 1996; Peeters & Le Blanc, 2001). The problem is that the interruption requires employees to shift their attention on someone or something else than the customer, which implies that there will be insufficient resources for serving the client.

Interruptions call for activity regulation since the person who is interrupted should modify the action plan and change the strategy to achieve the original goal within the new constraint (i.e., the interruption) or even delay goal accomplishment (Zijlstra et al., 1999). Each action (e.g., serving a customer) requires developing or recalling an action plan and executing an action plan (Frese & Zapf, 1994). Interruptions of the ongoing action are hindrances in this regulation process and may elicit negative feelings such as irritation or anger. Also, interruptions create a motivational burden because the employee should stay committed to the ongoing task (i.e., serving the customer) while regulating goal-irrelevant requirements (Baethge et al., 2015). In addition, going back to the initial task after the interruption costs effort and time, thus demotivating employees. In line with SIP, interruptions from others ‘cue an individual as to what to consider’ (p. 229) in the work environment. Employees, who experience a lot of interruptions, may become increasingly doubtful that all these interruptions are necessary and attribute negative motives to the interruptions (cf. Brown, Kulik, Cregan, & Metz, 2017). They may form the belief that others at work are self-serving and are putting their own interest above the interests of the employee (Wilkerson, 2002), resulting in cynicism.

Negative (re)actions

Consistent with the attitude–behaviour framework (Ajzen, 1991), cynicism as an attitude held by an employee about his/her work will produce behaviours critical of the work. In the context of service work, we expect cynical employees to show more negative (re)actions towards their customers, as negative (re)actions represent behavioural manifestations of cynical attitudes (Schaufeli et al., 1996). By expressing negative feelings linked with cynicism, employees minimize the discrepancy between felt (negative) emotions and their expression. Service providers’ psychological state of cynicism results in immediate, impulsive actions or public displays of emotion, e.g., being unfriendly or showing tension during the service encounter (Côté, 2005; Weiss & Cropanzano, 1996).

By combining SIP theory (Salancik & Pfeffer, 1978) and the social interaction model (Côté, 2005), we argue that social interruptions will relate positively to daily cynicism (because they interfere with employees’ goal-directed action during the service encounter; Baethge et al., 2015). In contrast, social support will relate negatively to daily cynicism (because it signals understanding for employees’ emotions), which will be displayed to customers in the form of negative (re)actions. Hence, we hypothesize:

Hypothesis 1:

There is a negative indirect relationship between social support and employee negative (re)actions to customers through cynicism.

Hypothesis 2:

There is a positive indirect relationship between number of social interruptions and employee negative (re)actions to customers through cynicism.

Employee negative (re)actions and service quality

Parasuraman et al. (1988) found that customers desire the following in the provided services: (a) to receive the promised services dependably and accurately (reliability); (b) to receive the promised services promptly (responsiveness); (c) value for money and use of high quality equipment (assurance); (d) physical facilities, equipment, personnel, and written material that look good (tangibles); and (e) individualized attention and caring (empathy). Past research has shown that employee affective expressions in service contexts shape customer satisfaction with the provided service. For instance, Tsai (2001) showed that positive emotional displays from employees increased customer willingness to return to the store and to speak positively about the provided services to their friends (see also, Tsai & Huang, 2002). Similarly, Pugh (2001) found that observed emotional displays by employees (i.e., greeting, smiling, making eye contact, and thanking the customer) related positively to customer evaluations of service quality. Also, Barger and Grandey (2006) performed a study where independent coders observed the smiling strength of employees and customers before, during, and after real service encounters, while customers rated their satisfaction with the encounter. Results indicated that employees’ overall smile strength predicted customers’ overall smile strength (i.e., the smile was mimicked by the customer) and higher customer satisfaction with the encounter. Nevertheless, most past research investigated the role of employee positive displays on customer satisfaction and service quality, thus neglecting the role of negative reactions. Furthermore, the few studies that focused on negative emotional displays mainly concern the alignment to negative, organizationally set display rules (e.g., bill collectors; Sutton, 1991), and not negative employee (re)actions.

In line with Côté (2005), we suggest that employees who display authentic negative reactions provoked by cynicism, like anger and contempt, are judged by their customers less favourably. Displaying authentic cynicism, anger, or contempt emotions in the form of negative (re)actions to customers triggers receivers to reciprocate senders’ negative behaviours as they conclude from such displays that the service provider will not accommodate them. When a customer is the target of an employee’s angry or antagonistic expression, (s)he may infer that (s)he did something wrong (Van Kleef, 2009). This may upset the customer and activate certain behaviours like leaving the store and never come back. Employees’ expressions may also elicit negative reactions in customers, either through an emotional contagion process (e.g., mimicry; see also, Barger & Grandey, 2006; Hatfield, Cacioppo, & Rapson, 1994) or by affecting impressions and interpersonal (dis)liking. For instance, expressions of irritability were found to decrease liking in exchange relationships (Clark & Taraban, 1991). Therefore, we expect that the more cynical the employee, the more negative (re)actions s/he will show towards customers, who will consequently rate the quality of the received service more negatively (see Figure 1).

Hypothesis 3:

There is a negative, indirect relationship between employee cynicism and service quality through employees’ negative (re)actions to customers.

Figure 1. The hypothesized model and study design.

Taken all hypotheses together, we suggest that the social context (social support and interruptions) communicates supporting and undermining signals, respectively, and influences employee attitudes (cynicism). Consequently, cynicism externalizes itself in the form of negative (re)actions towards customers, who eventually rate the quality of the received service less favourably. Therefore, our final hypotheses are as follows:

Hypothesis 4:

There is a positive sequential indirect relationship of social support to service quality through cynicism and employee negative (re)actions to customers.

Hypothesis 5:

There is a negative sequential indirect relationship of social interruptions to service quality through cynicism and employee negative (re)actions to customers.

Method

Procedure and participants

Data were collected by four groups of undergraduates (consisting of three students each) as a part of their Bachelor thesis. Each group collected data from a minimum of 10 employees. Prior to data collection, all students had followed training, where they familiarized with observation methods in general and practised with Bales’ (1950) Interaction Process Analysis (IPA) method and the utilized observation checklist. Despite expressed criticisms (McGrath & Julian, 1963), Bales’ (1950) coding scheme was considered the most appropriate for this study because it is accurate for contexts where social interactions with customers are viewed as a form of problem-solving and not as form of conflict resolution (Angelmar & Stern, 1978). Observers were not aware of the specific study hypotheses. The students were instructed to approach organizations that were easily accessible, and where direct interaction with customers took place. Moreover, they were directed to look for customer–employee interactions that were relatively short and for situations where unobtrusive observations would be possible. Prior to the observations, the students asked permission from the direct supervisor of the participating employees to investigate their interactions with some of their customers. Within each organization, the students asked the supervisor to allow observing two employees, whose customer interactions were also observed. After obtaining the supervisors’ consent, the observers approached the employees and asked for their permission to observe them. Two students observed each employee for 45 min.

Employees were informed about the study procedure and were reassured about the anonymity of the data. Also, they were offered the possibility to receive a brief report with the results of the study. Upon their agreement, the observers distributed a one-time questionnaire to the employees and asked them to fill it in at the end of the observation. With this questionnaire, employees’ perceived overall social support from colleagues and their cynicism levels up to that point of the shift were evaluated. Afterwards, the observers found a place within ear shot and with view of the customer and the employee (to observe verbal and non-verbal (re)actions). Students recorded the number of social interruptions during the 45-min observation period (during which employees may have interacted with more than one customer), as well as the number of employee negative (re)actions. During the observation, the third student of each group approached random customers, who had just been served by the observed employee. To randomize the customer sample such that the students/observers would select customers in an unbiased way (i.e., to avoid selecting only the satisfied or dissatisfied ones, or only those customers who were not in a hurry), every 10 min a customer was approached with a request to fill in a short questionnaire about the quality of the provided services. If a customer refused to collaborate, the next customer was approached. In this way, social support, social interruptions, cynicism, and negative reactions were measured with the employee as the reference point, while service quality was measured with the employee–customer interaction as the reference point.

A total of 48 employees and 141 customers filled out and returned the questionnaires. The employee sample includes 54% females. The sample’s age ranged from 19 to 62 years with an average of 28 years (SD = 9.4). Most participants (40%) were high-school graduates or had fulfilled secondary vocational education, while 23% of the participants were university graduates. Organizational tenure was on average 7.4 years (SD = 8.2), and the average working hours per week was 28 h (SD = 10.9). Note that part-time jobs are not uncommon in The Netherlands (Statistics Netherlands, 2014). Most participants worked in customer service jobs (24%) or as sales persons (24%), while 17% worked as waiters/waitresses in restaurants/cafes. We did not collect information about the sociodemographic characteristics of the customers.

Measures

Observations Employees’ negative (re)actions

The observations were scored using Bales (1950) IPA form. According to this theoretical framework, interactions with customers are viewed from a problem-solving perspective that include six phases, namely, problems of orientation, problems of evaluation, problems of control, problems of decision, problems of tension-management, and problems of integration. Bales’ coding system consists of 12 categories that refer to either task-oriented problems or socioemotional-oriented problems. The task-oriented problems are dealt with, primarily, by the expression of attempted answers and questions, while the socio-emotional problems are handled basically by the expression of (positive and negative) reactions (Rogers & Farace, 1975). In this study, we focused on how employees handled socio-emotional problems by manifesting negative (re)actions, because we are interested in the emotional aspect of service work that is largely neglected (i.e., negative vs. positive).

During the 45 min that the observation lasted, the observers marked the form every time the observed employee showed a negative action or reaction to customers. These (re)actions agree with “rejecting” and “reacting” behaviours that have been uncovered by Rafaeli (1989) during unstructured observations of cashiers during their interactions with customers. We used the following three items/categories of the IPA form capturing negative reactions: disagrees (e.g., shows passive rejection, formality, withholds help), shows tension (e.g., asks for help, withdraws out of field), and shows antagonism (e.g., deflates other’s status, defends/asserts self). As each observation was conducted by two observers, we estimated the inter-rater agreement or reliability by calculating the intra-class correlation coefficient (ICC). ICC values per item ranged from .96 to .99 indicating high inter-rater agreement (Shrout & Fleiss, 1979). Therefore, for each item, we used the mean score of both observers for negative (re)actions exhibited by the employees in further analyses.

Social interruptions

We measured the total number of observed social interruptions during the 45 min that the observations lasted. The observers had to mark every time that the observed employee was interrupted by, for example, the telephone, other people or colleagues, or noise. The ICC was .98. Again, we used the average score of the two observers for the analysis.

Employee self-reports

Social Support was assessed with the four-item scale of Bakker, Demerouti, and Verbeke (2004), including: “Can you ask your colleagues for help if necessary?” Participants were asked to indicate the frequency with which they experience each statement using a 5-point rating scale ranging from 1 = never to 5 = always (Cronbach’s alpha = .80).

Cynicism

We assessed day-level cynicism with three of the five items from the Maslach Burnout Inventory–General Survey (Schaufeli et al., 1996). Cynicism reflects whether employees experienced indifference or a distant attitude towards work on the day of the observation. An example item is: “Today, I doubted the importance of my work”. Participants were asked to indicate the extent to which they agreed with each statement using a 4-point rating scale (1 = totally disagree, 4 = totally agree; Cronbach’s alpha = .66).

Customer reports

Service Quality was measured with the Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality (SERVQUAL; Parasuraman et al., 1988). The scale consists of 22 items that capture the following five dimensions: tangibles (4 items; “The company’s physical facilities are visually appealing”), reliability (5 items; “The company provides its services at the time it promises to do so”), responsiveness [4 items; “Employees of the company are not always willing to help customers” (reversed)], assurance (4 items; “Employees of the company are polite”), and empathy [5 items; “Employees of the company do not know what your needs are” (reversed)]. Instead of a 7-point answer format, we used a 6-point answer format (1 = strongly disagree, 6 = strongly agree) in order to avoid the undecided answer. Participants had to answer the items in response to the contact that they had with the employee that just served them. Following the scale instructions, we reversed the scores on responsiveness and empathy. We used the total score for the analysis; the combined index proved to be reliable (Cronbach’s α = .88).

Strategy of analysis

In order to capture the hierarchical structure of our data (i.e., customer reports nested within employees/observations), we performed multi-level structural equation modelling (MSEM) with Mplus 7 (Muthén & Muthén, 2011). We tested a two-level model with the customer reports at the first-level of analysis (N = 141) and the individual employees/observations at the second level of analysis (N = 48). All predictor variables were centred to the grand-mean as they represent level-2 variables (i.e., employee level). The decision to use a two-level model was supported by the ICC (ρ), which indicates the amount of variance in the dependent variable that can be attributed to the different levels of analyses. Results suggested that 9% of the variance in service quality could be attributed to the differences between employees, while the rest of the variance could be attributed to within-employee fluctuations (i.e., the different customers). This result calls for multi-level analyses.

The significance of indirect effects was estimated with the Monte Carlo Method for Assessing Mediation (MCMAM; Selig & Preacher, 2008), which is appropriate for multi-level analysis and performs better than the widely used Sobel test (MacKinnon, Lockwood, & Williams, 2004). This method of estimation provides a confidence interval for the indirect effect. If the confidence interval does not contain zero, mediation (or any form of indirect effect) is supported. Overall model fit was assessed with the chi-square and the related degrees of freedom, the Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI), the Akaike Information Criterion (AIC), the Bayesian Information Criterion (BIC), and the Root Mean Square Error of Approximation (RMSEA). A cut-off value of .06 for RMSEA and .95 for CFI and TLI indicates a good fitting model (Hu & Bentler, 1999). In addition, models with lower AIC and BIC values indicate superior model fit.

Results

Hypotheses testing

Table 1 presents mean scores, standard deviations, and correlations between the study variables both at the between-person and the within-person level of analysis. The theoretical model displayed in Figure 1, which included all hypothesized paths yielded a good fit to the data (see Table 2). Next, we compared the hypothesized model to a model that was not only similar to the hypothesized model but also included the direct path from social support to employee negative (re)actions (M2) and a model that included the direct path from social support to service quality (M3). Table 2 presents the fit indices of all alternative models. We found that neither of these models fit better than the hypothesized model. Similarly, the fit of a model including the direct path from number of social interruptions to employee negative (re)actions (M4) or the model with the direct path from number of social interruptions to service quality (M5) did not fit better to the data than the hypothesized model. The final alternative model (M6) that was not only similar to the hypothesized model but also included the direct path from cynicism to service quality was also not significantly better than the hypothesized model. Moreover, none of the direct paths added in M2–M6 were significant. Also, the AIC and the BIC of the hypothesized model were lower than the AIC and the BIC of all alternative models, indicating that the proposed model is the most parsimonious and best-fitting model.

Table 1. Means, standard deviations, and correlations between the study variables (N = 48 employees and 141 customers).

	Variables	Mean	SD	1	2	3	4	5
1	Number of social interruptions	2.38	2.04	–	.26**	.25**	−.05	−.05
2	Social support	3.46	.85	.26**	–	−.29**	−.22**	.24**
3	Cynicism	1.82	.63	.25**	−.29**	–	.29**	−.15
4	Negative (re)actions	.32	.71	−.05	−.22**	.29**	–	−.28**
5	Service quality	4.69	.59	−.03	.15	−.09	−.18*	–

Correlations below the diagonal are between-employee correlations (N = 48), where customer-level data were averaged across each employee; correlations above the diagonal concern within-employee correlations; **p < .01; *p < .05.

Table 2. Goodness-of-fit indices and chi-square difference tests of alternative multi-level structural equation models (N = 48 employees and 141 customers).

Model	χ^2	df	Comparison^1	Δχ^2	AIC	BIC	CFI	TLI	RMSEA
M1: Hypothesized	4.21	5			774.66	821.84	1.00	1.00	.00
M2: Hypothesized and social support → negative (re)actions	3.19	4	M1–M2	1.02 ns	775.64	825.77	1.00	1.00	.00
M3: Hypothesized and social support → service quality	2.36	4	M1–M3	1.85 ns	774.81	824.94	1.00	1.00	.00
M4: Hypothesized and social interruptions → negative (re)actions	3.01	4	M1–M4	1.20 ns	775.46	825.59	1.00	1.00	.00
M5: Hypothesized and social interruptions → service quality	4.01	4	M1–M5	.20 ns	776.46	826.59	.99	.99	.01
M6: Hypothesized and cynicism → service quality	3.97	4	M1–M6	.24 ns	776.42	826.55	1.00	1.00	.00

¹For all comparisons; Δdf = 1; ns: not significant.

We hypothesized that social support (Hypothesis 1) will relate indirectly and negatively to employee negative (re)actions towards customers via cynicism. Inspection of the path coefficients of the hypothesized model showed that there was a negative path from social support to cynicism (estimate = −.27, SE = .10, t = −2.69, p < .01). Although social support was measured at the general rather than the day level, its relationship with daily cynicism represents an occasion-specific association between both constructs (Eid & Diener, 2004). Moreover, cynicism related positively to employee negative (re)actions (estimate = .34, SE = .16, t = 2.17, p < .05). The MCMAM showed that the indirect effect from social support to employee negative (re)actions through cynicism was significant and negative (95% confidence interval: LL = −.218, UL = −.003) providing support to Hypothesis 1.

Hypothesis 2 suggested that social interruptions relate indirectly and positively to employee negative (re)actions to customers via cynicism. Results of the hypothesized model showed that there was a positive path from the number of social interruptions to cynicism (estimate = .11, SE = .04, t = 2.74, p < .01). In addition, the indirect effect from number of social interruptions to employee negative (re)actions through cynicism was also significant (95% confidence interval: LL = .002, UL = .091). Thus, Hypothesis 2 was also supported.

Hypothesis 3 suggested an indirect and negative effect of cynicism on service quality through employee negative (re)actions. Next to the positive relationship between cynicism and employee negative (re)actions, employee negative (re)actions related negatively to service quality (estimate = −.15, SE = .07, t = −2.02, p < .05). Hypothesis 3 was supported since the indirect effect of cynicism to service quality through employee negative (re)actions was significant and negative (95% confidence interval: LL = −.115, UL = −.0003).

Finally, we examined the significance of the two sequential indirect effects: social support (Hypothesis 4)/interruptions (Hypothesis 5) → cynicism → employees’ negative (re)actions → service quality. Both indirect effects were non-significant (95% confidence interval for the sequence starting from social support: LL = −.001, UL = .042; 95% confidence interval for the sequence starting from social interruptions: LL = −.0003, UL = .017), thus rejecting Hypotheses 4 and 5. In an additional analysis, we also tested whether the interaction between social support and number of social interruptions predicted cynicism. This interaction was not significant.

Discussion

In the present study, we used observer ratings, employee self-reports, and customer service quality evaluations in an attempt to investigate the micro (i.e., situational) and social processes that explain which conditions make service employees cynical towards their work, and why on days employees are cynical, their customers are less satisfied with the provided services. In line with our hypotheses, results showed that in jobs where social support is lacking, and on days that employees are frequently interrupted during their service encounters, they are more likely to endorse a cynical attitude towards the recipients of their services. Thus, on days that employees are cynical about their work, they are more likely to exhibit negative (re)actions towards their customers, which translate into less satisfied customers. We showed that daily and structural social aspects of the job do translate into employee negative re(actions) towards their customers, even though negative emotional displays are usually prohibited by the display rules set by service organizations. We found that employees do show negative (re)actions towards their customers that signal their cynical state towards their customers and function as self-protection (cf. Ashforth & Humphrey, 1993). In these micro and meso-processes, cynicism was the linking mechanisms that connects lack of social support and frequent interruptions during customer encounters with negative (re)actions to customers. This is because employee cynicism did not remain an internalized experience but it externalized towards the service recipients, who eventually became dissatisfied with the service.

Theoretical contributions

In this study, we examined service work by considering the employee and the customer perspective simultaneously, thus responding to the call to integrate employee functioning and service management literatures (Bradley et al., 2010; Macintosh & Stevens, 2007). Synthesizing both perspectives allowed revealing the underlying mechanisms that explain why service employees become cynical, and how cynicism converts into low service quality. In this context, the study findings contribute to the literature in the following ways.

First, this study contributes to the burnout literature since it tests the assumption that cynicism is expressed through negative (re)actions towards the recipients of one’s work (Schaufeli et al., 1996). Although many studies showed that cynicism results in unfavourable outcomes for both employees and organizations (Lee & Ashforth, 1996), there is very limited evidence linking cynical attitudes and employee negative (re)actions. Moreover, the studies that provide support for this relationship (e.g., Evans, Goodman, & Davis, 2010; Wilkerson et al., 2008) investigated cynicism towards the organization and not cynicism towards one’s job or the recipients of one’s work. Thus, in this study, we show that cynical employees are less likely to align with the positive emotional display rules that are usually set by service organizations (Morris & Feldman, 1996). Rather, they are more likely to be authentic and (re)act negatively towards their customers. Importantly, we supported this relationship by combining employees’ self-reports of their state of cynicism and independent observations of their negative (re)actions during service encounters. Previous research has shown that positive affective/attitudinal responses including positive affect (Barger & Grandey, 2006) and work engagement (Salanova, Agut, & Peiró, 2005) can have a positive relationship with customer satisfaction. Our study showed that under specific conditions (frequent interruptions by other people) and in specific jobs (with low levels of social support), employees become cynical towards their job, which, contrary to organizational norms, is externalized towards customers and consequently, has unfavourable effects to service quality. Given that this is a rather negative process, future studies should focus on specific personal or contextual factors that may act as moderators and mitigate the negative impact of service work on employee cynicism.

Second, we make a conceptual contribution by extending SIP theory (Salancik & Pfeffer, 1978) to emotion work and the context of service encounters. We demonstrated that the social context of service work, by communicating supporting signals and expectations concerning individual behaviour, influences employee daily experiences of cynicism that consequently are externalized in the form of negative (re)actions towards customers. Noteworthy, this process that is in line with SIP theory was supported when combining self-report and observation data. The information provided by people at work (in the form of support or expectations) relates to how employees react to customers. SIP theory explains how others can form employee attitudes through making aspects of the social environment salient, influencing employees in how to interpret the environmental cues and their needs in the work context (Brown et al., 2015; Salancik & Pfeffer, 1978). SIP theory has usually been applied to explain positive employee attitudes, such as job satisfaction and organizational commitment (e.g., Pollock, Whitbred, & Contractor, 2000). Our study, similar to the study of Brown et al. (2015), not only shows that SIP theory is useful in explaining a negative employee attitude, cynicism, but also shows that SIP theory can explain daily experiences of cynicism as they unfold on a working day.

A third contribution of the present study is that it established the relationship between cynicism and negative (re)actions towards clients, thus explaining why cynical employees provide low-quality services. In line with previous studies on the role of emotional contagion between employees and customers during service encounters and the impact of emotional contagion processes on service quality (e.g., Barger & Grandey, 2006; Tsai, 2001; Tsai & Huang, 2002), our findings suggest that employees’ negative (re)actions are perceived as inappropriate or unfair by customers, thus determining their attitudes (i.e., low service quality). This is again in line with SIP theory (Salancik & Pfeffer, 1978) as employees, through their negative reactions, give social signals to customers and determine their attitudes. Adding to previous evidence, we did not focus on positive but rather on negative (re)actions. In line with Côté’s (2005) social interaction model, we proposed and found that employees’ negative work attitudes affect customers’ attitudes (i.e., low satisfaction with the provided service) because cynical employees are likely to show negative (re)actions to customers. Côté (2005) suggests that customers infer that they did something wrong in order to receive a negative reaction, or they just mimic employees’ negative (re)actions (Hatfield et al., 1994). Thus, cynical employees’ negative (re)actions are publicly displayed to customers during the service encounters that, in turn, determine customers’ (negative) emotional states and related attitudes (i.e., low service quality).

It is important to note that in our study, service quality correlated significantly only with social support (positively) and with negative (re)actions (negatively). Also, when accounting for the relationship between social support and service quality in our path analyses, this relationship was no longer significant. These weak relationships are in line with earlier research. For instance, Gracia et al. (2013) found that dedication (which can be considered as the theoretical opposite of cynicism) was unrelated to customer-rated service quality, whereas Wilkerson et al. (2008) found that cynicism was unrelated to supervisor ratings of organizational citizenship behaviour. Also, in the study of Moliner et al. (2008), the correlation between cynicism and (self-rated) extra-role customer service was similar to the correlation found in our study (although the correlation in that study was significant probably due to the larger sample size). These findings indicate that cynicism (or employee well-being) is not sufficient to inhibit performance during service encounters. Employees need to behave in ways that externalize cynicism to customers. We showed that the negative relationship between cynicism and service quality may be explained by employees’ negative (re)actions to customers, whereas Gracia et al. focused on relational service competence. Searching for constructs that act as linking mechanisms in connecting cynicism to performance outcomes is a promising way to uncover their relation and to overcome the generally low correlations (e.g., Demerouti & Cropanzano, 2010). As Wilkerson et al. (2008) concluded, the cynicism–performance relationship is rather complex and suggested that factors like impression management, contingent rewards, accountability, and self-discipline may explain the influence of cynicism.

Strengths, limitations, and avenues for future research

An unquestionable strength of the present study is the multi-method design that was adopted. The hypothesized relationships were supported with data coming from three unique sources of information: observer ratings, employee self-reports, and customer evaluations. In this way, we overcame potential shortcomings that could be attributed to common method biases, adding to the robustness of our findings, the validity of the hypothesized process and contributing to overall research rigour (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Nevertheless, it is also important to note that we used a cross-sectional design since information from the different sources was collected at the same point in time. This means that the relationships between the variables are correlational and conclusions about causality should be made with caution. To counteract this concern, we made sure that data were collected in a way that was in line with the order of the variables in the proposed process. Namely, data on number of social interruptions, negative (re)actions, and service quality were collected during the service encounters (but by different raters), while social support and cynicism were evaluated by employees after the observations were over. However, alternative explanations are also possible as, for instance, dissatisfied customers could also interrupt service employees. In this way, service quality could be a hypothetical predictor rather than an outcome in the model. In a similar vein, frequent interruptions may attenuate boredom at work. A more detailed observation of the employee–customer interactions combined with experience sampling methods (that allow following people over short time-periods) could clarify these issues.

Another strong point of our study is that the variables were measured at different levels of specificity that identified with the hypothesized process. This adds to the robustness of the study design because it allows capturing reality in the most accurate manner. Since we were interested in testing how cynicism relates to negative (re)actions during the service encounters and in turn, to poor service quality, it was logical to assess daily instead of overall cynicism. In this way, we captured the most proximal predictor of (re)actions and, consequently, service quality (Sonnentag, Dormann, & Demerouti, 2010). Similarly, we measured social interruptions as observed during the service encounters, as a proximal antecedent and of daily cynicism. Instead, we focused on employee perceptions of the support they receive from their colleagues in general as we could not be sure that participants would have the chance to receive social support prior or during the observation (which could have occurred at the beginning of the working day). Although it has been shown that employees who work in a generally supportive climate are more likely to receive adequate support on a daily basis (Xanthopoulou, Bakker, Demerouti, & Schaufeli, 2012), future studies could focus on daily support over and above general support, in order to show whether the effects we found remain even when the most proximal antecedent is taken into account. Furthermore, although our findings imply that employees’ negative (re)actions may elicit negative emotions in customers that, in turn, determine their attitudes (i.e., low satisfaction with the provided service), we did not measure customers’ reactions or other linking mechanisms next to daily cynicism. Future studies could also incorporate such variables in the proposed process.

A drawback of the study is the relatively small sample size. Although our sample at level 2 (N = 48) comes close to the recommendation of Maas and Hox (2004), who suggested that 50 cases are needed at the highest level of analysis, the small sample size may be responsible for certain non-significant findings. It is indicative that even though MSEM analyses supported most hypothesized indirect effects, the analyses did not provide evidence for the full process from social support/interruptions to service quality through first, cynicism and them employee negative reactions to customers. Some authors (e.g., Mathieu & Taylor, 2007) imply that supportive evidence for each intermediate process (i.e., as in our case: social support/interruptions → cynicism → negative reactions and cynicism → negative reactions → service quality) indicates a full sequence of effects. Although this is not a very robust approach, it does suggest that the full chain of effects could have been supported in our study, if the sample had been larger.

Practical implications and overall conclusion

Our study suggests that cynical employees are likely to provide low-quality services to customers because they are less likely to regulate their cynical attitudes, thus exhibiting negative (re)actions towards their customers. The findings of this study are informative for organizations that aim at preventing service employees from becoming cynical towards their work and their customers. Based on our findings, it may be suggested that redesign programs should place special emphasis on the social characteristics of the service context: social interruptions and support. Although it may be difficult to reduce the frequency of occurred interruptions from the side of customers, employees themselves may try to avoid interrupting their colleagues, when they interact with customers. Furthermore, promoting a supportive work environment seems to help employees to effectively handle the emotionally demanding aspects of service work (e.g., Xanthopoulou, Bakker, Demerouti, & Schaufeli, 2009) and avoid cynical attitudes.

In line with the suggestions of Gracia et al. (2013), we think that this can happen by asking work teams directly ‘what it is that helps them to perform better’ (p. 52) in their interaction with customers. As aspects of the social environment were important to influence social encounters, we think that unit-level interventions shared by colleagues within the same unit could be most effective. Such interventions may take the form of human resources practices to create fair and clear procedures to promote employee well-being (Moliner et al., 2008) or to develop unit-level strategies that help to generate positive unit experiences while improving coordination on work tasks (Gracia et al., 2013). However, employees may also be trained so that they are able to arrange for themselves the conditions that are needed to perform well during their service work. Job crafting (i.e., individual adjustments of the job such that it is less demanding and more motivating and challenging) has been found to help individuals being more creative (Demerouti, Bakker, & Gevers, 2015; Van Wingerden, Bakker, & Derks, 2017) and to show more task and extra-role performance (Demerouti, Bakker, & Halbesleben, 2015). Interventions on the unit-level combined with individual-level interventions may help to create the social context that facilitates service work. This study provides evidence that it is important for organizations to care for the occupational well-being of their employees and to value its strategic importance as it influences the service-oriented effectiveness reported by customers.

Disclosure statement

No potential conflict of interest was reported by the authors.