Who Chose this Restaurant Anyway? The Effect of Responsibility for Choice,Guarantees,and Failure Stability on Customer Complaining

Abstract

Customers often behave in the context of a group, with different behavior occurring in this context to that which transpires in an individual context. However, customer complaining behavior (CCB), including voice, negative word of mouth, in addition to that transmitted electronically, and exit, in a group setting has not been studied previously. A service failure during a group celebration at a restaurant and the pattern of CCB that ensues is examined. This is based on customers’ level of responsibility in restaurant selection on behalf of the group, the presence of an unconditional service guarantee, and the perceived stability of the failure. Findings suggest that customers are more inclined to exit when they have participated to a greater degree in choice and that the presence of an unconditional guarantee interacts with participation to influence negative word of mouth intentions. Perceived failure stability had the greatest influence on CCB.

Keywords

customer complaining behavior level of responsibility unconditional service guarantee failure stability service failure and restaurants

Service failure is inevitable in the restaurant context because uncontrollable problems can occur during a dining experience. These may include slow service due to unanticipated overcrowding (Folkes, Koletsky, & Graham, 1987) and disturbances caused by other customer behavior, such as noisy children or public drunkenness (Huang, Lin, & Wen, 2010). When restaurant customers experience service failure, they frequently warn family and friends about their dissatisfying encounter, such as negative word of mouth (WoM) and/or discontinue patronizing the restaurant, that is, exit (Lam & Tang, 2003). Increasingly, negative electronic WoM (e-WoM) is another form of customer complaining behavior (CCB) engaged in following failure, with customers sharing their disapproving evaluations of restaurants online (Zhang, Ye, Law, & Li, 2010). These types of CCB are detrimental for restaurants. Negative WoM can taint the opinions of other potential customers (Nyer & Gopinath, 2005) and is damaging to an organization’s reputation and its ability to acquire and retain customers (Lau & Ng, 2001). This effect is enhanced in the case of negative e-WoM because in the online environment customers can tell the world about their dissatisfaction and it is also relatively permanent (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004). If customers choose to exit, this is felt by the organization through declining market share, profit reduction, and loss of future revenues (Reichheld, Markey, & Hopton, 2000).

The CCB literature advocates that customer voice, whereby customers direct their complaints to the organization, needs to be facilitated by organizations. This is the only form of CCB that is beneficial to organizations (Halinen & Tahtinen, 2002). Voice provides the opportunity to restore and maintain customer relationships and to reduce private action, that is, negative WoM and exit (Nyer & Gopinath, 2005). This is important for restaurants because repeat customers account for between 60% and 75% of their sales (Tanyeri & Cobe, 2010). Furthermore, restaurateurs face intensive competition, with customers being able to choose between various alternatives (Han & Ryu, 2009). Among its other benefits, voice can be used to identify common service failures (Tax & Brown, 1998), improve service design and delivery (Marquis & Filiatrault, 2002; Tax & Brown, 1998), and develop new product ideas. In the light of this, restaurants are recommended to facilitate customer voice and minimize the harmful CCB types of exit and negative WoM.

Existing research that focuses on the antecedents of CCB suggests that they can be classified as organizational, situational, and personal variables (Marquis & Filiatrault, 2002). Our article contributes to this line of inquiry by being the first to examine whether the situational variable of the level of customer responsibility in restaurant choice for a group celebration influences CCB. Group dining is common in restaurants and the mere presence of others is expected to influence customer behavior (Tombs & McColl-Kennedy, 2010), although this has not been formerly examined in respect to CCB. We test the effect of customers’ level of responsibility (sole responsibility, joint responsibility, and group responsibility) for the choice of a restaurant for a group celebration on CCB, following a service failure during the group dining experience. It is noted that previous studies, drawing on attribution theory, have examined the effect of causal locus, that is, whether the cause of service failure is perceived to reside within or outside of the customer, on CCB. For example, customer self-blame has been found to be negatively associated with voice and exit intentions (Mattila & Ro, 2008). However, our study is unique in that it does not examine a situation where the customer is to blame for the service failure, but rather a scenario where the customer, to varying degrees, has selected a restaurant that has subsequently failed.

Restaurant service is characterized by intangibility, simultaneous production and consumption, high contact between service personnel and customers, and nonstandardization (Pedraja & Yagüe, 2001). This makes it difficult for customers to evaluate prepurchase, thereby leading to perceived risk. A customer, or customers, responsible for choosing a restaurant could use salient cues in the service environment, such as the presence of a service guarantee, to reduce perceived risk (Berry, 1995). Guarantees suggest that customers can expect a certain level of service, and even if the service fails, it promises a payout and/or that the failure will be fixed (Berry, 1995). Risk minimization would logically be important in the selection of a restaurant for a group celebration (Ariely & Levav, 2000). In respect to service failure and CCB, little research has investigated the role that guarantees play (see, for exception, McColl, Mattsson, & Morley, 2005; McQuilken & Robertson, 2011). Our study examines the effect of the presence of an unconditional service guarantee, that is, a promise that if customers are dissatisfied with the service for any reason, they are entitled to a full refund (Marvin, 1992) on CCB. Specifically, we anticipate that customers’ level of responsibility in restaurant choice and the presence of an unconditional guarantee will interact to influence CCB.

Finally, existing research suggests that anything unusual that brings customers’ attention to an outcome, such as the disconfirmation of expectations, results in causal search (Oliver, 1997). Customers attribute causes for service failure based on the classification of stability, locus, and controllability (Weiner, 1980). The stability dimension, which is the focus of the current study, relates to the customer-perceived likelihood that the negative event will recur (Weiner, 1980). Researchers have proposed that the presence of a service guarantee will emphasize to customers that the failure is an aberration, and, therefore, unlikely to recur, that is, it is unstable (Kashyap, 2001). While it follows that the harmful CCB types of negative WoM and exit may be reduced as a consequence, this association has not been examined empirically previously. Therefore, the third antecedent variable to be tested in this study is failure stability.

In summary, our study examines whether, following a failure at a restaurant during a group dining experience, CCB is influenced by customers’ level of responsibility in selecting the restaurant on behalf of the group, the presence of an unconditional service guarantee, and the perceived stability of the failure. The remainder of this article is organized as follows. First, a review of the literature leads to the development of the hypotheses. Second, the method is outlined. Third, the results of the analysis are reported and discussed. Fourth, the implications for managers based on the results are presented. Finally, the limitations of the study and directions for future research are provided.

Literature Review and Hypotheses Development

Customer Complaining Behavior

CCB refers to “a set of multiple (behavioral and nonbehavioral) responses, some or all of which are triggered by perceived dissatisfaction with a purchase episode” (Singh, 1988, p. 94). The CCB types of negative WoM, exit, and voice are those that appear most frequently in the literature (Marquis & Filiatrault, 2002), and are the focus of this study. Additionally, we are witnessing an increase in negative WoM communicated electronically. Negative e-WoM is defined as negative statements made by potential, actual, or former customers about a product or company that are made available to a multitude of people and institutions via the Internet (Hennig-Thurau et al., 2004). As this form of complaining is likely to become increasingly important (Harrison-Walker, 2001) and as negative e-WoM has not previously been considered as a type of CCB, we include it in our study.

Complaint statistics grossly understate customer dissatisfaction (Day, Grabicke, Schaetzle, & Staubach, 1981), with the typical disgruntled customer engaging in a variety of indirect activities. These include boycotting the organization or telling others about their dissatisfaction (Day et al., 1981). Negative WoM is an indirect form of CCB because confrontation with service personnel is avoided (Marquis & Filiatrault, 2002). Informing others about a service failure allows customers to vent, gain social support for the validity of their negative feelings, and is a form of retaliation as it discourages others from purchasing from the organization (Curren & Folkes, 1987). Similar to negative WoM, exit is a form of private CCB (Singh, 1988). It involves customers ceasing to patronize the organization. Exit is considered to be an antiorganization reaction, whereby customers intend to penalize the organization.

Negative WoM has increasingly spread online, making it ever more a public response to customer dissatisfaction as it reaches many people beyond customers’ friends and family (Vilpponen, Winter, & Sundqvist, 2006). It can be transmitted via avenues, such as social media, discussion boards, e-mail, product review sites, blogs, consumer opinion platforms, and boycott websites. In the restaurant context, for example, the introduction of various popular restaurant review sites, such as Urbanspoon and Eatability, has made it easy for customers to share their negative restaurant experiences online. Other differences between traditional and negative e-WoM have been noted, with e-WoM being usually unsolicited (Vilpponen et al., 2006), available to other customers for an indefinite time (Hennig-Thurau et al., 2004), potentially anonymous (Hennig-Thurau et al., 2004), and accessible by organizations (Hennig-Thurau et al., 2004).

Customer voice is the only form of CCB that is referred to largely throughout the literature as being beneficial to organizations because it provides them with the opportunity to analyze and rectify customer dissatisfaction. If customers are unwilling to voice, dissolution of the customer and organization relationship is likely (Halinen &Tahtinen, 2002). Therefore, studies on CCB have stressed that customers’ voice needs to be encouraged by organizations (Tax & Brown, 1998).

The Influence of Level of Responsibility in Restaurant Choice for a Group on CCB

A group is defined as “the assemblage of two or more people who share common interests or goals, perceive or may develop some form of cohesiveness and who interact with one another on a social and task-oriented level” (Finsterwalder & Tuzovic, 2010, p. 110). In the current study, we refer to a primary/personal group. It is well-established that consumer behavior in the group context is different to that which occurs in an individual context (Ariely & Levav, 2000); however, CCB in the group context has not been studied previously. Therefore, we examine the influence of customers’ varying levels of responsibility (sole responsibility, joint responsibility, and group responsibility) in selecting a restaurant for a celebratory dinner on behalf of a group. Here, the customer acts as an agent of, or an agent with, the group, in coordination and cooperation with its members (Bagozzi & Dholakia, 2006). Following a service failure at the restaurant of choice during the group celebration, we examine the CCB that ensues.

When making a decision on behalf of a group, the welfare of others is considered (Corfman & Lehmann, 1993), particularly in the case of primary groups. Customers are likely to endeavor to minimize regret and to avoid losses for all concerned (Ariely & Levav, 2000). This would suggest that the greater the level of customer participation in the choice of restaurant, the more responsible the customer would feel for the celebration being dampened by the service failure, at least in as far as the selection of the restaurant. The customer may feel obliged to “put things right” for the group by voicing the group’s displeasure (Kim & Chen, 2010). Similarly, the customer would be more inclined to “get back” at the restaurant by telling others of its poor performance, including increasingly via electronic means (see, e.g., Harrison-Walker, 2001; Hennig-Thurau et al., 2004), and by vowing to never return.

The customer trying to protect his or her ego can also explain this pattern of CCB. An ego-involved product decision is one where the product is viewed as an extension of the self (Iverson & Rueder, 1956). When a customer is responsible for selecting a restaurant for a group function, he/she will make a choice that enhances his or her self-image and conveys a positive image to others in the group (Belk, 1988). Should the choice be unsatisfactory, the customer’s self-identity, feelings of importance, personal status, and self-esteem will be threatened (Iverson & Rueder, 1956), resulting in CCB in an attempt to restore the ego. Therefore, the following hypotheses are advanced:

Hypothesis 1a: Customers who are solely responsible for the choice of restaurant will be more inclined to voice following service failure.

Hypothesis1b: Customers who are solely responsible for the choice of restaurant will be more inclined to spread negative WoM following service failure.

Hypothesis 1c: Customers who are solely responsible for the choice of restaurant will be more inclined to spread negative e-WoM following service failure.

Hypothesis 1d: Customers who are solely responsible for the choice of restaurant will be more inclined to exit following service failure.

The Influence of Unconditional Service Guarantees on CCB

From their review of the past 20 years of service guarantee research, Hogreve and Gremler (2009) concluded that little is known about the role of service guarantees in a failure context. A service guarantee is “[ . . . ] a statement explaining what service customers can expect (the promise) and what the company will do if it fails to deliver (the payout)” (Hart, Schlesinger, & Maher, 1992, p. 20). In the current study, we examine the unconditional guarantee, which promises that if customers are dissatisfied with the service for any reason, they are entitled to a full refund (Marvin, 1992). Unconditional guarantees have been argued to be particularly beneficial for organizations that rely on referrals and that are susceptible to negative WoM (Hart et al., 1992). Restaurants fall into this category of organization (Zhang et al., 2010).

In this article, we do not hypothesize for the main effects of guarantee on CCB; this association was tested in McQuilken and Robertson’s (2011) recent study, which was also set in the restaurant context. They found that while the unconditional guarantee encourages voice, it does not influence exit or negative WoM. We hypothesize an interaction between customers’ level of responsibility for restaurant choice and the presence of an unconditional guarantee on CCB. Hart (1993) suggested that the usefulness of a guarantee increases as the ego involvement of the customer increases. For a customer who has solely selected a restaurant on behalf of a group, the presence of a guarantee should strengthen his/her inclination to voice following failure (Walster, Berscheid, & Walster, 1976). A guarantee signals the probability of a successful complaint (McColl et al., 2005; McQuilken & Robertson, 2011), that is, that the restaurant will remedy the failure without hassle. Where customers perceive a greater probability of a successful complaint, lower levels of exit and negative WoM behavior are likely (Singh & Wilkes, 1996), so that the following hypotheses are raised:

Hypothesis 2a: The presence of an unconditional guarantee increases customers’ intentions to voice when they are solely responsible for the choice of restaurant.

Hypothesis 2b: The presence of an unconditional guarantee reduces customers’ intentions to spread negative WoM when they are solely responsible for the choice of restaurant.

Hypothesis 2c: The presence of an unconditional guarantee reduces customers’ intentions to spread negative e-WoM when they are solely responsible for the choice of restaurant.

Hypothesis 2d: The presence of an unconditional guarantee reduces customers’ intentions to exit when they are solely responsible for the choice of restaurant.

The Influence of Failure Stability on CCB

The stability dimension of attribution refers to customers’ perceptions regarding the permanency of the failure’s cause (Weiner, 1980). Customers will assess whether a failure is a random occurrence or is likely to happen again. For example, where a restaurant’s service is slow because one of the chefs had to leave to attend to a family emergency, the customer may conclude that the event is unlikely to recur. If, however, a service fails for stable reasons (e.g., a restaurant has slow service due to the owner’s policy of operating with minimal staff), the customer will be inclined to believe that the failure will happen again (Weiner, 1986). Customers who attribute failures to stable causes are more likely to warn friends and family than if they are uncertain about future performance (Blodgett & Granbois, 1992; Curren & Folkes, 1987), and will probably vow to never return to the organization (Blodgett & Granbois, 1992; Folkes et al., 1987). However, Curren and Folkes (1987) found that customer-perceived stable versus unstable causes do not influence voice; customers are equally inclined to voice. This gives rise to the following hypotheses:

Hypothesis 3a: A customer-perceived stable service failure will encourage dissatisfied restaurant customers to voice.

Hypothesis 3b: A customer-perceived stable service failure will encourage dissatisfied restaurant customers to spread negative WoM.

Hypothesis 3c: A customer-perceived stable service failure will encourage dissatisfied restaurant customers to spread negative e-WoM.

Hypothesis 3d: A customer-perceived stable service failure will encourage dissatisfied restaurant customers to exit.

Various studies have found that service guarantees heighten customers’ expectations of service quality, thereby reducing their risk perceptions and increasing their confidence in making a purchase decision (see, e.g., Hogreve & Gremler, 2009; Wirtz, 1998; Wong, Tsaur, & Wang, 2009). Customers are likely to reason that unless an organization is assured that it has achieved a high level of service quality, it will not offer an unconditional service guarantee (Tucci & Talaga, 1997). This is because it would be unable to afford the payouts associated with offering a guarantee (McCollough & Gremler, 2004). Indeed, why would a restaurant promise excellent service and offer a generous payout to customers if it cannot keep its promise? Therefore, it has been proposed that service guarantees influence customers’ attribution of blame in the context of service failure by increasing their perceptions of firm control over service delivery (Callan & Moore, 1998) and by destabilizing the cause of the failure (Kashyap, 2001). However, we argue that although an unconditional service guarantee buffers restaurants to some extent following a failure that is ascribed to an unstable cause, it has no influence on CCB following a stable failure. When customers experience a service failure in the context of a restaurant offering a 100% satisfaction guarantee, and the explanation suggests that the failure is unstable, they are likely to perceive that the restaurant is a high-quality provider that probably could not have anticipated the failure because it had not occurred before. As a result, customers will tend to be more lenient and forgiving of the failure (Kelley, Hoffman, & Davis, 1993), thus reducing their CCB intentions, and they will not expect to be compensated for it. This argument is consistent with the “benefit of the doubt” effect (Hess, 2008). Hess (2008) found that organizations with reputations for producing quality products are somewhat insulated from the effect of service failure. However, if the failure is ascribed to a stable cause, customers will begin to doubt the credibility of the unconditional guarantee; customers believe that high-quality providers should be able to eliminate stable failures (Nikbin, Ismail, Marimuthu, & Abu-Jarad, 2011). How could a restaurant that was genuine in its promise to provide excellent service fail so consistently? Was the guarantee simply a marketing ploy to get customers’ business? Based on this argument, we advance the following hypotheses:

Hypothesis 4a: The presence of an unconditional guarantee reduces customers’ intentions to voice following a customer-perceived unstable service failure.

Hypothesis 4b: The presence of an unconditional guarantee reduces customers’ intentions to spread negative WoM following a customer-perceived unstable service failure.

Hypothesis 4c: The presence of an unconditional guarantee reduces customers’ intentions to spread negative e-WoM following a customer-perceived unstable service failure.

Hypothesis 4d: The presence of an unconditional guarantee reduces customers’ intentions to exit following a customer-perceived unstable service failure.

Research Method and Preliminary Analysis

This study employed a 3 (responsibility for choice: sole, joint, or group) × 2 (guarantee: none or unconditional) × 2 (failure stability: unstable or stable) full factorial, between-subjects experimental design using scenarios. Respondents were assigned randomly to one of the 12 experimental treatments. Scenarios are used frequently in service research because of their advantages: control over how respondents perceive the independent variables, thereby improving internal validity (Cooper & Emery, 1995); and enabling the inclusion of a representative set of service failures (Smith, Bolton, & Wagner, 1999).

The context of this study was a full-service restaurant, where service failures are common (Namkung, Jang, & Choi, 2011). The scenarios depicted a failure for which the restaurant was at fault. Respondents were asked to imagine that they were the customer in the scenario and to respond as they would in a similar situation. Web-based self-report survey data were collected from Australian online panel members aged 18 years and over. The panel employed is recruited via e-mail and targeted website advertising. Its members participate in a diverse range of surveys, with a typical panelist partaking in six surveys per year. Average survey response rates for the panel range from 8% to 40%. For the current study, a random sample of 1,986 panel members was sent an opt-in email message inviting them to participate in the study, including a link to the questionnaire. Of those panelists invited to participate, 400 members completed the questionnaire, representing a response rate of 20.1%. Each respondent received $3 for participating in the survey.

The manipulations for choice of restaurant were achieved by altering the scenario descriptions:

You choose: It is your close friend’s birthday next Saturday evening and you are responsible for booking a restaurant for the celebration. You need to organize a booking for 10 people. Having considered various possible restaurants over the past couple of weeks, you decide to book a table at Eden Restaurant. You hope that everyone who has been invited will be happy with your choice of restaurant because you really want the night to be enjoyed by all, especially your friend.

You and a friend jointly choose: It is your close friend’s birthday next Saturday evening. Your friend asks for your advice on a restaurant for the celebration. Having both considered various possible restaurants over the past couple of weeks, you jointly decide to book a table at Eden Restaurant. You both hope that everyone who has been invited will be happy with your choice of restaurant because you really want the night to be enjoyed by all.

The group chooses: It is your close friend’s birthday and a group of 10 of you are out celebrating at a bar. As the evening passes, several members of the group mention that they are starting to feel hungry. As a group, you all agree to ring to see if there is a table available at Eden Restaurant. Fortunately, there is. Everyone hopes that the group’s choice of restaurant will be a good one so that the evening continues to be enjoyed by all, especially your friend.

Respondents in the unconditional guarantee condition were asked to look at a newspaper advertisement for a fictitious restaurant that describes a 100% Satisfaction Guarantee (see Figure 1).

Figure 1

Hypothetical Advertisement for Unconditional Guarantee

Scenarios under the guarantee treatment also included the following sentence, “the restaurant was chosen because you / you and your friend / members of the group remember seeing a newspaper advertisement promoting the restaurant’s 100% satisfaction guarantee.” Respondents were then presented with a failure involving slow service, meals not arriving at the same time, and overcooked food. The stability manipulations were described as follows:

Unstable failure: The waiter apologizes for the problems your group has experienced, explaining that he is usually the bartender and this evening he is filling in for two wait staff who called in sick last minute. The waiter adds that Eden Restaurant has been experiencing a problem with its oven this evening. This problem has never occurred before and is being rectified first thing tomorrow. Clearly, your group has been unlucky. He goes on to highlight Eden Restaurant’s long-established reputation for excellent service and points to various industry awards that it has attained. He claims that in the several years that he has worked at the restaurant, the problems your group has experienced have been incredibly rare, assuring you that they won’t occur again.

Stable failure: The waiter, appearing quite flustered, comments that while all of the staff at Eden Restaurant try to provide excellent customer service at all times, there are simply never enough staff on duty to serve the number of occupied tables. He goes on to say that, unfortunately, this is a common occurrence.

Scenario realism was measured using four items (on a 1-7 scale, from strongly disagree to strongly agree) developed by Wilson and McNamara (1982). A reported mean of 6.20 confirmed that respondents found the service encounter to be highly realistic and were able to adopt the role of the customer. To verify that respondents recognized who selected the venue for the birthday celebration, they were asked, “Who was responsible for the choice of restaurant?” The response options were, “I was responsible for the choice of restaurant”; “both my friend and I were jointly responsible for the choice of restaurant”; and “the group as a whole was responsible for the choice of restaurant.” The stability manipulation was tested via one item taken from Jorgensen (1996) and two additional items developed for our study. These items were designed to reflect the notion that a cause that is thought to be stable is perceived to persist over time and across situations while an unstable cause is perceived to be subject to changing temporal or situational conditions (Dixon, Spiro, & Jamil, 2001). Customers who perceive failures to be permanent are likely to feel that identical results will recur in the future. The additional items were worded as follows: “The cause of Eden Restaurant’s failure is stable, that is, likely to recur”; and “I think that a failure like the one our group experienced is likely to occur again at Eden Restaurant” (both measured on a 1-7 scale, from strongly disagree to strongly agree). An independent-sample’s t test revealed a difference between the unstable, M = 4.34, SD = 1.40, and stable failure conditions, M = 6.34, SD = 0.90, t(343) = −15.62, p = .000. A guarantee manipulation check question was not included, as it would potentially bias the responses of the no-guarantee group. For the dependent variables, voice intentions were measured using five items adapted from Singh (1988) and Liu and McClure (2001), exit intentions were measured via three items (Liu & McClure, 2001; Ping, 1993; Singh, 1990), and negative WoM intentions by four items (Athanassopoulos, Gounaris, & Stathakopoulos, 2001; Huefner et al., 2002; Singh, 1990). As we were unable to locate a multiple-item instrument to measure negative e-WoM, Churchill’s (1979) instructions for generating items to measure it were followed, that is, the items were developed to capture the domain of the construct using a literature search. Four items were developed to measure it using this approach, “share your bad experience online for everyone to see”; “provide a negative review of the restaurant online”; “vent your negative feelings about your bad experience online”; and “convince other customers not to patronize the restaurant by posting the details of your bad experience online.” All CCB items were measured on a 1-7 scale, anchored at extremely unlikely to extremely likely. Attitude toward complaining was included as a covariate in the analysis as it has been found to influence CCB (DeWitt & Brady, 2003; Richins, 1983; Smith et al., 1999). While some customers are assertive and seek redress whenever they are dissatisfied with a service, others are reluctant to complain no matter how disgruntled they might be. If the influence of attitude toward complaining is not removed, it is difficult to identify the effects of the treatment variables. Attitude toward complaining was measured using eight items taken from Richins (1982), Singh (1990), and Moorman (1998). Respondents also provided demographic data.

Data Analysis and Results

Of the 400 questionnaires administered, the following cases were removed: those with particularly short questionnaire completion times (6), cases in which the response to the “responsibility for choice of restaurant” manipulation check was incorrect (6), and 43 univariate outliers (i.e., standard scores of ±3.0 or beyond). Where the sample size permits, as was the case in the current study, removal of such outliers has been shown to greatly improve accuracy and reduce errors (Osborne & Overbay, 2004). This left 345 usable responses, with each of the 12 experimental cells containing between 26 and 30 responses. Of these, 42.3% of respondents were male and 57.7% were female. With respect to age, 11.9% of respondents were aged between 18 and 34 years, 30.7% were between 35 and 54 years, 41.2% were within the age range of 55 to 74 years, and 16.2% of the sample was over 75 years of age. Hence, respondents aged between 18 and 34 years are underrepresented in our sample.

Confirmatory factor analysis (CFA) was employed to test the validity of the types of CCB and attitude toward complaining constructs. The measurement model fit the data adequately (see Table 1) following the deletion of two items measuring voice, and one item for negative WoM, and three items that measured attitude toward complaining. The removal of these items could be theoretically justified (Bagozzi & Yi, 1988). Composite reliability (CR) and average variance extracted (AVE) were calculated per construct, all of which were found to be greater than 0.5 (Fornell & Larcker, 1981), with the exception of the AVE values for negative e-WoM (0.47) and attitude toward complaining (0.45), although these values are very close to the recommended cutoff. It is noted that some researchers (see, e.g., Bagozzi & Baumgartner, 1994) advocate an AVE value of 0.40 or greater, which is met in this case, thereby demonstrating convergent validity. The constructs had adequate discriminant validity (see Table 2), as the square root of the AVE value for each construct was larger than the correlation between them.

Table 1

Final Measurement Model Results

Model Fit	χ²	df	GFI	CFI	RMSEA
	254.36	125	.92	.96	.05
Variable		Mean	SD	Standardized Loading	CR
Customer voice intentions		5.4	1.2		.87
Report the problem to the restaurant to get what was fair		5.4	1.4	0.90
Complain to the restaurant to get what you deserved		5.1	1.4	0.86
Inform the restaurant so that it can take action		5.7	1.2	0.71
Customer exit intentions		6.4	0.7		.87
Not go back to this restaurant again		5.9	1.4	0.65
Eat at another restaurant next time		6.4	0.8	0.89
Look for another restaurant next time around		6.4	0.7	0.86
Customer negative WoM intentions		5.5	1.1		.87
Speak to friends and relatives (who were not at the birthday celebration) about your bad experience		5.7	1.1	0.74
Convince your friends and relatives (who were not at the birthday celebration) not to go to this restaurant		5.0	1.5	0.79
Warn friends and family so that they would not have the same problem		5.7	1.2	0.82
Customer negative e-WoM intentions		3.0	1.4		.84
Share your experience online for everyone to see		2.9	1.6	0.87
Provide a negative review of the restaurant online		3.2	1.7	0.78
Vent your negative feelings about your bad experience online		2.9	1.6	0.79
Convince other customers not to patronize the restaurant by posting the details of your bad experience online		2.9	1.5	0.71
Attitude toward complaining		4.9	1.1		.82
It feels good to get my frustration with an unsatisfactory good or service off my chest by complaining		4.3	1.6	0.62
People should always complain about unsatisfactory goods or services		5.5	1.2	0.80
I often complain when I am dissatisfied with a good or service because I feel that it is my duty to do so		4.6	1.5	0.75
I like people who complain to firms because usually their complaints are reasonable		4.8	1.3	0.66
I usually complain about unsatisfactory goods or services		5.2	1.4	0.67

Note. GFI = goodness-of-fit index; CFI = comparative fit index; RMSEA = root mean square error of approximation; CR = composite reliability; WoM = word of mouth.

Table 2

Correlation Matrix and AVE Statistics

Construct	1	2	3	4	5
1. Customer voice intentions	0.73
2. Customer exit intentions	0.67	0.73
3. Customer negative WoM intentions	0.23**	0.58**	0.73
4. Customer negative e-WoM intentions	0.21**	0.27**	0.34**	0.69
5. Attitude toward complaining	0.55**	0.16**	0.34**	0.26**	0.67

Note. WoM = word of mouth; e-WoM = electronic word of mouth; AVE = average variance extracted. Diagonal elements shown in boldface are square roots of the AVE values of the constructs.

p < .01.

Between-groups multivariate analysis of covariance (MANCOVA) was run to examine the influence of responsibility for choice, guarantee, and failure stability on the combined dependent CCB construct (refer to Table 3). Box’s M test of equality of covariance matrices indicates that the assumption of homogeneity of variance–covariance matrices has not been violated, p = .057 (Tabachnick & Fidell, 2001). The covariate, attitude toward complaining, was found to have a main effect on the combined dependent variables: F(4, 329) = 44.58, p = .000, Wilks’s Λ = .648, partial η² = .352. There was a main effect for level of responsibility for choice—F(8, 658) = 2.18, p = .027, Wilks’s Λ = .95, partial η² = .026—and for stability on the combined dependent variables—F(4, 329) = 10.43, p = .000, Wilks’s Λ = .89, partial η² = .112. An interaction effect was also found for guarantee and responsibility for choice on the combined dependent variables: F(8, 658) = 2.66, p = .007, Wilks’s Λ = .94, partial η² = .031. The interaction effects for guarantee and stability failed to reach statistical significance: F(8, 658) = 1.24, p = .843, Wilks’s Λ = .99, partial η² = .00; therefore, Hypotheses 4a, 4b, 4c, and 4d were rejected.

Table 3

MANCOVA for Voice, Negative WoM, Negative e-WoM, and Exit

Effect	Wilks’s Λ	F	df	Significance	Partial η²
Attitude toward complaining	0.65	44.53	4	.000	.35
Guarantee	0.98	1.82	4	.124	.02
Responsibility	0.95	2.18	8	.027	.03
Stability	0.89	10.43	4	.000	.11
Guarantee × responsibility	0.94	2.67	8	.007	.03
Guarantee × stability	1.00	0.35	4	.843	.00
Responsibility × stability	0.97	1.24	8	.273	.02
Guarantee × responsibility × stability	0.98	1.06	8	.389	.01

Note. MANCOVA = multivariate analysis of covariance; WoM = word of mouth. Computed using α = .05. n = 345.

Next, the results for the dependent variables were considered separately using ANCOVA analysis (see the appendix for cell means and Table 4 for ANCOVA results). Levene’s test for homogeneity of variance indicated that it has not been violated for intention to voice, F(11, 333) = 0.62, p = .811; negative WoM, F(11, 333) = 1.62, p = .092; or negative e-WoM, F(11, 333) = 1.78, p = .057. However, the significant result for exit intentions, F(11, 333) = 4.25, p = .000, indicates that the assumption of homogeneity of variance has not been met. Hence, a more stringent p value of .01 is applied to this variable. While the covariate, attitude toward complaining, has a significant, but small influence on negative e-WoM and exit (partial η² = .06 and .02, respectively), it has a moderate influence on negative WoM (partial η² = .11) and a large influence on voice (partial η² = .31), as per the theory. After adjusting for respondents’ attitude toward complaining, significant results were achieved across several hypotheses.

Table 4

ANCOVA for Voice, Negative WoM, Negative e-WoM, and Exit

	Voice						Negative WoM						Negative e-WoM						Exit
Test	Sum of Squares	df	Mean Square	F	Significance	η²	Sum of Squares	df	Mean Square	F	Significance	η²	Sum of Squares	df	Mean Square	F	Significance	η²	Sum of Squares	df	Mean Square	F	Significance	η²
Attitude toward complaining	151.95	1	151.95	145.75	.000	.31	36.91	1	36.91	39.53	.000	.11	34.88	1	34.88	19.98	.000	.06	2.79	1	2.79	0.02	.020	.02
Guarantee	6.60	1	6.60	6.33	.012	.02	0.04	1	0.04	0.04	.840	.00	0.43	1	0.43	0.25	.621	.00	0.38	1	0.38	0.74	.390	.00
Responsibility	1.33	2	0.67	.64	.528	.00	2.07	2	1.03	1.11	.332	.01	3.83	2	1.92	1.10	.335	.01	7.55	2	3.78	7.34	.001	.04
Stability	5.58	1	5.58	5.35	.021	.02	20.90	1	20.90	22.39	.000	.06	5.26	1	5.26	3.01	.083	.01	16.15	1	16.15	31.40	.000	.09
Guarantee × responsibility	5.90	2	2.95	2.83	.060	.02	8.26	2	4.13	4.42	.013	.02	7.55	2	3.78	2.16	.117	.01	0.07	2	0.03	0.07	.936	.00
Guarantee × stability	1.07	1	1.07	1.03	.312	.00	0.13	1	0.13	0.14	.709	.00	0.29	1	0.29	0.17	.685	.00	0.09	1	0.03	0.17	.677	.00
Responsibility × stability	0.72	2	0.36	.35	.707	.00	0.15	2	0.07	0.08	.925	.00	6.69	2	3.34	1.92	.149	.01	1.54	2	0.77	1.50	.225	.01
Guarantee × responsibility × stability	3.10	2	1.55	1.49	.227	.01	1.67	2	0.84	0.90	.410	.01	2.97	2	1.48	0.85	.429	.01	0.23	2	0.11	0.22	.803	.00
	Adjusted R² = .32						Adjusted R² = .18						Adjusted R² = .08						Adjusted R² = .12

Note. ANCOVA = analysis of covariance; WoM = word of mouth; e-WoM = electronic word of mouth. Computed using α = .05. n = 345.

The main effects for responsibility for choice on voice, F(2, 333) = 0.64, p = .528, partial η² = .00; negative WoM, F(2, 333) = 1.11, p = .332, partial η² = .01; and negative e-WoM, F(2, 333), = 1.01, p = .335, partial η² = .01, failed to reach statistical significance; therefore, Hypotheses 1a, 1b, and 1c are rejected. A significant main effect was found for responsibility for choice on exit, F(2, 333), = 7.34, p = .001, partial η² = .04. Post hoc comparisons using the Games–Howell test indicated that the mean score for the group treatment (M = 6.18, SD = 0.90) was significantly different from the individual (M = 6.51, SD = 0.61) and the pair treatment groups (M = 6.50, SD = 0.72). However, there was no difference between the individual and pair groups. Therefore, Hypothesis 1d is supported.

A two-way interaction was found between guarantee and responsibility for choice on negative WoM, F(2, 333) = 4.42, p = .013, partial η² = .03 (see Figure 2). Simple effects analysis revealed that there is a reduced intention to engage in negative WoM following a failure when a guarantee is offered, versus when it is not, when the individual is solely responsible for the choice of venue, F(1, 333) = 6.02, p = .016, Ms 5.34 versus 5.77. This finding lends support to Hypothesis 2b. There was no difference found for the pair, F(1, 333) = .06, p = .815, Ms 5.60 versus 5.63, or the group treatments, F(1, 333) = 3.00, p = .087, Ms 5.27 versus 5.59. Post hoc comparisons revealed that when a guarantee is not present, the mean score for negative WoM is significantly different across the individual (M = 5.77) and group treatments (M = 5.27); however, there is no difference between the pair (M = 5.60) and the individual or group treatments. In contrast, when the restaurant offered an unconditional guarantee, post hoc tests revealed that there is not a significant difference between any of the individual, pair, or group treatment groups (Ms of 5.34, 5.63, and 5.59, respectively). Significant interaction effects were not found between guarantee and responsibility for choice on voice, F(2, 333) = 2.83, p = .06; negative e-WoM, F(2, 333) = 2.16, p = .117; or exit, F(2, 333) = 0.07, p = .936; therefore, Hypotheses 2a, 2c, and 2d were not supported.

Figure 2

Interaction Between Level of Responsibility for Choice of Restaurant and Guarantee on Negative Word of Mouth

Main effects were found for stability on voice—F(1, 333) = 5.35, p = .021, partial η² = .02; negative WoM—F(1, 333) = 22.39, p = .000, partial η² = .06; and exit—F(1, 333) = 31.40, p = .000; partial η² = .09, providing support for Hypotheses 3a, 3b, and 3d. The main effect for stability on negative e-WoM failed to reach statistical significance, F(1, 333) = 3.02, p = .083, partial η² = .01; hence, Hypothesis 3c is rejected. Table 5 presents a summary of the results obtained for each of the hypotheses tested.

Table 5

Summary of Results for Hypothesized Associations

Hypothesis	F Statistic	Significance Level	Test Result
1a: Sole responsibility for choice increases voice	0.64		Not supported
1b: Sole responsibility for choice increases negative WoM	1.11		Not supported
1c: Sole responsibility for choice increases negative e-WoM	1.01		Not supported
1d: Sole responsibility for choice increases exit	7.34	***	Supported
2a: An unconditional guarantee increases voice when customers are solely responsible for choice	2.83		Not supported
2b: An unconditional guarantee reduces negative WoM when customers are solely responsible for choice	4.42	**	Supported
2c: An unconditional guarantee reduces negative e-WoM when customers are solely responsible for choice	2.16		Not supported
2d: An unconditional guarantee reduces exit when customers are solely responsible for choice	0.7		Not supported
3a: A customer-perceived stable failure encourages voice	5.35	***	Supported
3b: A customer-perceived stable failure encourages negative WoM	22.39	***	Supported
3c: A customer-perceived stable failure encourages negative e-WoM	3.02		Not supported
3d: A customer-perceived stable failure encourages exit	31.40	***	Supported
4a-4d: An unconditional guarantee reduces (a) voice, (b) negative WoM, (c) negative e-WoM, and (d) exit following a customer-perceived unstable failure	1.24		Not supported

Note. WoM = word of mouth; e-WoM = electronic word of mouth.

p ≤ .01. ***p ≤ .001.

Discussion

The mean scores for the CCB types across all scenarios (refer to the appendix) indicate that customers are more inclined to defect than to engage in other types of CCB if a restaurant’s service fails (main effect means for exit range from 6.17 to 6.63), as per the theory; although it is noted that customers’ intentions to voice are also high (Ms of 5.22 to 5.55) in our study. This is in contrast to previous study findings that suggest the association between customer dissatisfaction and voice is weak (McColl et al., 2005). While customers also reported a strong inclination to warn family and friends about the restaurant (Ms of 5.25 to 5.82), negative e-WoM intentions (Ms of 2.81 to 3.14) are low comparatively. Our study is the first to consider negative e-WoM as a CCB type, with findings suggesting that customers have not fully embraced the online medium for voicing. They may be unaware of dedicated complaint or restaurant review sites, for example. Alternatively, they might feel it is inappropriate to share publicly negative comments that they did not enjoy their friend’s birthday celebration. It is also noted that our sample underrepresents the 18- to 34-year-old age-group and that e-WoM is especially prevalent among young people (Okazaki, 2009). However, being online panelists, our sample is likely to be regular Internet users, irrespective of their age.

Our study is the first to consider CCB in a group context by testing the influence of customers’ level of responsibility for restaurant selection on behalf of a group on CCB. No significant main effects were found for participation in choice on either voice or negative e-WoM. Respondents report equally high voice intentions and equally low negative e-WoM intentions, regardless of whether they were solely or partly responsible for restaurant selection. However, the greater the level of customer responsibility (sole or joint) in the selection of a restaurant, the more inclined the customer was to exit than when the decision was made via a group consensus. It appears that when customers feel responsible for choosing a venue that ultimately disappoints the group, they will be more likely to boycott the restaurant in the future. This finding is of importance to restaurateurs because exit is considered to be an anti-organization reaction where the intention of dissatisfied customers is to punish the firm (Diaz & Ruiz, 2002). Indeed, customer exit has been found to have a strong influence on the future revenue, profit, and market share of organizations (Reicheld & Sasser, 1990).

Although no main effect was found for the level of responsibility for choice on negative WoM, the significant two-way interaction between responsibility for choice and guarantee on negative WoM (see Figure 2) suggests that an unconditional guarantee reduces customers’ negative WoM intentions if the restaurant is chosen independently. This is a benefit of offering a guarantee as negative WoM can taint others’ attitudes and intentions toward a restaurant (Nyer & Gopinath, 2005). The offer of a 100% satisfaction guarantee justifies the customer’s choice of restaurant, so that the customer is not as inclined to engage in negative WoM for self-validation. However, the presence of an unconditional guarantee does not influence negative WoM intentions when the choice of restaurant is made with others, that is, either with a friend or with the group. It appears that where the responsibility for choice is reduced, the effect of the guarantee on negative WoM diminishes. Under the no guarantee condition, a customer is less likely to vent his or her displeasure to others when the entire group chooses the restaurant (as opposed to when choosing individually). The dissatisfying experience is less threatening to an individual when everyone in the group has contributed to the choice of restaurant. Therefore, the customer is less inclined to spread negative WoM for the purpose of restoring the ego or testing for consensus in a group choice context.

Researchers have argued that service guarantees encourage dissatisfied customers to voice (see, e.g, McQuilken & Robertson, 2011; Wirtz, 1998). Although we did not hypothesize this association, it is important to note that we did not find it (see Table 3) as per the results of McColl et al.’s (2005) study, that is, an unconditional guarantee did not facilitate voice. Our results also suggest that a guarantee does not interact with responsibility for choice to influence voice; ours is the first study to examine this interaction. Proponents of service guarantees have long espoused that unconditional guarantees encourage dissatisfied customers to voice to the service provider following a failure, providing the offending organization with an opportunity to reduce consumer dissatisfaction and maintain loyalty (see, e.g., Hart, 1988; Spreng, Harrell, & Mackoy, 1995; Wirtz, 1998). However, our findings support the conclusion drawn by McColl et al. (2005) that when used in isolation of additional quality cues, service guarantees do not encourage customer voice.

As hypothesized, when customers expect the restaurant to fail again in the future, they are more inclined to engage in negative WoM than if they are unsure about future performance, supporting past findings (Blodgett & Granbois, 1992; Curren & Folkes, 1987). A small main effect was also found for stability on customers’ voice intentions. This finding is in contrast to that of Curren and Folkes (1987), who found that customers are equally likely to voice whether the cause is perceived to be stable or unstable. Our results provide strong support for the notion that customers will vow to never return to the restaurant if they deem the cause of the failure to be stable (Blodgett & Granbois, 1992; Folkes et al., 1987). Indeed, customers who perceive failures to be stable are likely to feel that identical results will recur in the future. Stability did not influence negative e-WoM.

Although we hypothesized that the presence of an unconditional guarantee would destabilize the cause of the service failure—the restaurant will be perceived as having control over the service and customers will deem the failure to be an aberration (Kashyap, 2001; Weiner, 1986)—and, therefore, minimize CCB, we failed to find an interaction between guarantee and stability on any of the CCB types. When customers have been advised that a failure is unlikely to recur, the presence (vs. absence) of an unconditional service guarantee does not add credibility to this statement. We do note, however, that while the failure stability manipulation check was significant and the effect size was large (partial η² = .42), the mean value for the unstable service failure manipulation (M = 4.34) was above the mid-point on the 1- to 7-point scale employed. This suggests that respondents deemed the service failure more likely to recur than we had expected. Had the unstable service failure manipulation been expressed in such a way that respondents perceived the failure as being even less likely to occur at a later date, we might have achieved even stronger results for Hypotheses 3a, 3b, and 3d. It is also possible that the inclusion of an apology in the unstable service failure condition may have influenced the manipulation check result and possibly respondents’ CCB intentions. However, it is difficult to say this with any certainty, as the influence of an apology on customers’ perceptions of service failure stability has not been examined empirically previously.

Managerial Implications

Our findings suggest that a greater level of customer participation in selecting a restaurant positively influences customer exit. This suggests that service personnel need to be trained to initially establish who is responsible for restaurant choice and regularly check with this customer that everything is satisfactory. As the customer responsible for choosing a venue for a group celebration will feel pressure to please everyone, he or she will perceive a restaurant that promises a certain level of service as being a less risky option. Our results also suggest that when a service failure occurs, an unconditional service guarantee shields the organization to some extent from the negative consequences of failures. The service guarantee tempers negative WoM intentions for those customers who have complete control over the choice of restaurant.

The findings suggest that an unconditional guarantee does not interact with either responsibility for choice or stability to influence voice; nor does it have a main effect on voice. This result is of concern for restaurants offering service guarantees given that they have been long touted to encourage voice. Restaurants should inform customers about the existence and nature of their guarantees via multiple communication vehicles, including advertising and their service personnel (Hocutt & Bowers, 2005). As voice involves customers’ time, effort, and possibly fear of embarrassment or confrontation (Hocutt & Bowers, 2005), the guarantee invocation and collection process needs to be clear, simple, and nonthreatening. For example, it might be useful for restaurants to place brochures on all tables encouraging customers to invoke the guarantee if they are dissatisfied. A brochure could guide dissatisfied customers through the invocation and collection process, in a similar fashion to the brochures adopted by banks and hospitals that assist customers in making a “complaint, suggestion, or compliment.” The brochure could also include supporting cues about a restaurant’s commitment to service quality. For example, any changes made to improve the service based on the feedback obtained from guarantee invocations could be outlined in the brochure. Restaurant staff can also be trained to identify service failures and to invoke the guarantee on behalf of customers. Finally, internal marketing should be employed to train, motivate, and empower employees to deliver the promises made to the customer via the guarantee to reduce the occurrence of service failure and to minimize payouts.

Overall, failure stability has the largest effect on CCB. For an unstable failure, our findings suggest that service personnel need to be trained to explain to customers the cause(s) of the failure and to stress that it is highly unlikely to recur; this can prevent customers from seriously doubting the restaurant’s motives and intentions. It is important that this information is both truthful and adequate (Folger & Cropanzano, 1998). If managed correctly, such practices will minimize organization-specific attribution, and hence CCB. Of course, organizations that pride themselves on the delivery of a high-quality service need to do their utmost to avoid the types of failures produced by stable causes. Techniques, such as blueprinting, the use of control charts, fishbone diagrams, and Pareto analysis are all powerful tools to monitor service quality and to get to the root cause of dissatisfying incidents (Lovelock, Patterson, Walker, 2011).

Limitations and Suggestions for Future Research

Our results need to be considered in the light of several limitations. First, the experiment was conducted solely in a full-service restaurant context, for which results should not be generalized beyond. Second, the scenarios involved a relatively large gathering of friends to celebrate a special occasion. Had the restaurant been chosen for a different purchase occasion, for example, an intimate dinner for two, the results might have been different. A third limitation relates to respondent self-selection, which reduces the randomness of the survey and may skew the results (Eaton, 1997). Fourth, respondents did not perceive the service failure to be as unlikely to recur as was anticipated. While this is a recognized limitation of our study, the unstable service failure appears to have been perceived by respondents as intended; it prompted them to report a significantly lower intention to engage in voice, negative WoM, and exit than did the stable service failure manipulation. Fifth, locus and control attributions are known to influence CCB (Folkes et al., 1987), yet they were held constant in this study. Sixth, the inclusion of an apology in the unstable service failure manipulation might have influenced customers’ perceptions of whether the failure was likely to recur and their CCB. Yet this is questionable at best, as these associations have not been examined empirically previously. Finally, the convergent validity of the measures for negative e-WoM and attitude toward complaining were questionable.

Various avenues for future research are open to pursue, particularly those that are stimulated by this study’s limitations. First, future experimental work in this area could employ videos of a simulated service failure using professional actors to increase realism. Second, to extend this study’s generalizability, future research could replicate it in other contexts where individuals are often responsible for selecting a service on behalf of a group (e.g., package tours and live shows). Third, the moderating influence of group power dynamics on CCB could be explored. For example, consider the scenario of a boss who has made a poor choice of restaurant for a staff function and the pattern of CCB that might ensue. Fourth, the attribution dimensions of locus and controllability could also be examined and in the same vein, blame contagion might be tested in a group service failure setting (see, e.g., Fast & Tiedens, 2010). Fifth, customers are more likely to engage in CCB as problem severity increases (see, e.g., Singh & Wilkes, 1996). Future research might examine whether failure severity moderates the influence of level of responsibility in restaurant choice for a group on CCB. Finally, future research might examine whether the level of responsibility for service selection influences positive WoM, following a successful service encounter. This is in-line with the self-serving attributional bias, which suggests that people tend to take credit for good outcomes (Fiske & Taylor, 1991).

Footnotes

Appendix

Mean Voice, Negative Word of Mouth (WoM), Negative e-WoM, and Exit Values by Experimental Condition

			Voice		Negative WoM		Negative e-WoM		Exit
Guarantee	Responsibility	Stable	Mean	SD	Mean	SD	Mean	SD	Mean	SD
No	You	No	5.26	1.23	5.42	1.04	2.83	1.52	6.27	0.73
		Yes	5.25	1.25	6.09	1.01	3.54	1.71	6.83	0.40
		Total	5.26	1.23	5.77	1.07	3.20	1.65	6.56	0.64
	You/friend	No	5.67	5.41	5.44	0.83	2.64	1.28	6.39	0.58
		Yes	5.41	1.04	5.85	1.11	3.16	1.48	6.64	0.52
		Total	5.54	1.07	5.64	0.99	2.90	1.39	6.51	0.56
	Group	No	5.13	1.32	4.88	1.18	2.82	1.32	5.90	0.97
		Yes	4.64	1.39	5.56	1.01	2.67	1.20	6.54	0.61
		Total	4.88	1.37	5.24	1.14	2.74	1.25	6.24	0.86
	Total	No	5.36	1.23	5.25	1.04	2.76	1.36	6.19	0.79
		Yes	5.10	1.27	5.84	1.05	3.12	1.51	6.67	0.52
		Total	5.22	1.25	5.55	1.09	2.95	1.45	6.44	0.71
Yes	You	No	5.44	1.02	4.96	1.10	2.46	1.06	6.28	0.62
		Yes	5.31	1.24	5.65	0.85	3.25	1.28	6.66	0.49
		Total	5.38	1.13	5.31	1.04	2.85	1.23	6.47	0.58
	You/friend	No	5.72	1.23	5.40	1.37	3.43	1.40	6.34	0.91
		Yes	5.43	1.21	5.97	0.97	3.24	1.30	6.63	0.81
		Total	5.57	1.21	5.70	1.20	3.33	1.33	6.49	0.86
	Group	No	5.36	1.30	5.40	0.80	2.76	1.10	5.84	1.01
		Yes	5.89	1.29	5.78	0.90	2.99	1.51	6.44	0.79
		Total	5.63	1.31	5.59	0.86	2.87	1.31	6.14	0.95
		Yes	5.54	1.26	5.80	0.90	3.16	1.35	6.58	0.71
		Total	5.52	1.22	5.52	1.05	3.01	1.30	6.37	0.82
Total	You	No	5.35	1.12	5.19	1.09	2.64	1.31	6.27	0.66
		Yes	5.28	1.24	5.88	0.95	3.40	1.51	6.75	0.45
		Total	5.32	1.18	5.54	1.07	3.02	1.46	6.52	0.61
	You/friend	No	5.70	1.52	5.42	1.11	3.02	1.38	6.36	0.75
		Yes	5.42	1.12	5.91	1.03	3.20	1.38	6.63	0.67
		Total	5.55	1.14	5.67	1.09	3.11	1.38	6.50	0.72
	Group	No	5.25	1.30	5.15	1.03	2.79	1.20	5.87	0.98
		Yes	5.26	1.47	5.67	0.96	2.83	1.36	6.50	0.70
		Total	5.25	1.38	5.42	1.02	2.81	1.28	6.19	0.90
	Total	No	5.43	1.20	5.25	1.07	2.81	1.30	6.17	0.83
		Yes	5.32	1.28	5.82	0.98	3.14	1.43	6.63	0.62
		Total	5.37	1.24	5.53	1.06	2.98	1.38	6.40	0.77

Lisa McQuilken (e-mail: lisa.mcquilken@deakin.edu.au) is a senior lecturer in marketing in the School of Management and Marketing, Deakin University, Melbourne, Australia. She holds a PhD in managing service recovery and has published in the International Journal of Hospitality Management, Journal of Tourism and Travel Marketing, and the Journal of Financial Services Marketing, among others. Her research interests include service guarantees as a recovery tool, the influence of justice on recovery, self-service technologies, and consumer complaining behavior. Nichola Robertson, PhD (e-mail: nichola.robertson@deakin.edu.au), is a senior lecturer in marketing at the Deakin Graduate School of Business, Deakin University, Melbourne, Australia. Her research interests include self-service technologies, consumer complaining behavior, service recovery, and customer dissatisfaction. Her research has been published in the Journal of Service Research, International Journal of Hospitality Management, Journal of Retailing and Consumer Services, and Services Marketing Quarterly. She serves on the Editorial Advisory Board of Managing Service Quality: An International Journal.

References

Ariely

Levav

(2000). Sequential choice in group settings: Taking the road less traveled and less enjoyed. Journal of Consumer Research, 27, 279-290.

Athanassopoulos

Gounaris

Stathakopoulos

(2001). Behavioral responses to customer satisfaction: An empirical study. European Journal of Marketing, 35, 687-707.

Bagozzi

R. P.

Baumgartner

(1994). The evaluation of structural equation models and hypothesis testing. In Bagozzi

R. P.

(Ed.), Basic principles of marketing research (pp. 386-422). Oxford, England: Blackwell.

Bagozzi

R. P.

Dholakia

U. M.

(2006). Antecedents and purchase consequences of customer participation in small group brand communities. International Journal of Research in Marketing, 23, 45-61.

Bagozzi

R. P.

(1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16, 74-94.

Belk

R. W.

(1988). Possessions and the extended self. Journal of Consumer Research, 15, 139-168.

Berry

L. L.

(1995). On great services: A framework for action. New York, NY: Free Press.

Blodgett

J. G.

Granbois

D. H.

(1992). Toward an integrated conceptual model of consumer complaining behavior. Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 5, 39-103.

Callan

R. J.

Moore

(1998). Service guarantee: A strategy for service recovery. Journal of Hospitality & Tourism Research, 22, 56-71.

10.

Churchill

G. A.

(1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16, 64-73.

11.

Cooper

D. R.

Emery

C. W.

(1995). Business research methods (5th ed.). New York, NY: McGraw-Hill.

12.

Corfman

K. P.

Lehmann

(1993). The importance of others’ welfare in evaluating bargaining outcomes. Journal of Consumer Research, 20, 124-137.

13.

Curren

M. T.

Folkes

V. S.

(1987). Attributional influences on consumers’ desires to communicate about products. Psychology & Marketing, 4, 31-45.

14.

Day

R. L.

Grabicke

Schaetzle

Staubach

(1981). The hidden agenda of consumer complaining. Journal of Retailing, 57, 86-106.

15.

DeWitt

Brady

M. K.

(2003). Rethinking service recovery strategies: The effect of rapport on consumer responses to service failure. Journal of Service Research, 6, 193-207.

16.

Diaz

A. B. C.

Ruiz

F. J. M.

(2002). The consumer’s reaction to delays in service. International Journal of Service Industry Management, 13, 118-140.

17.

Dixon

A. L.

Spiro

R. L.

Jamil

(2001). Successful and unsuccessful sales calls: Measuring salesperson attributions and behavioral intentions. Journal of Marketing, 65, 64-78.

18.

Eaton

(1997). The problems and pitfalls of net survey methodologies. Petaluma, CA: Creative Research Systems.

19.

Fast

N. J.

Tiedens

L. Z.

(2010). Blame contagion: The automatic transmission of self-serving attributions. Journal of Experimental Social Psychology, 46, 97-106.

20.

Finsterwalder

Tuzovic

(2010). Quality in group service encounters: A theoretical exploration of the concept of a simultaneous multi-customer co-creation process. Managing Service Quality: An International Journal, 20, 109-122.

21.

Fiske

S. T.

Taylor

S. E.

(1991). Social cognition. New York, NY: McGraw-Hill.

22.

Folger

Cropanzano

(1998). Organizational justice and human resource management. Thousand Oaks, CA: Sage.

23.

Folkes

V. S.

Koletsky

Graham

J. L.

(1987). A field study of causal inferences and consumer reaction: The view from the airport. Journal of Consumer Research, 13, 534-539.

24.

Fornell

Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 28, 39-50.

25.

Halinen

Tahtinen

(2002). A process theory of relationship ending. International Journal of Service Industry Management, 13, 163-180.

26.

Han

Ryu

(2009). The roles of the physical environment, price perception, and customer satisfaction in determining customer loyalty in the restaurant industry. Journal of Hospitality & Tourism Research, 33, 487-510.

27.

Harrison-Walker

L. J.

(2001). E-complaining: A content analysis of an Internet complaint forum. Journal of Services Marketing, 15, 397-412.

28.

Hart

C. W. L.

(1988). The power of unconditional service guarantees. Harvard Business Review, 66(4), 54-62.

29.

Hart

C. W. L.

(1993). Extraordinary guarantees: A new way to build quality throughout your company and ensure satisfaction for your customers. New York, NY: American Management Association.

30.

Hart

C. W. L.

Schlesinger

L. A.

Maher

(1992). Guarantees come to professional service firms. Sloan Management Review, 33(3), 19-29.

31.

Hennig-Thurau

Gwinner

K. P.

Walsh

Gremler

D. D.

(2004). Electronic word-of-mouth via consumer-opinion platforms: What motivates consumers to articulate themselves on the Internet? Journal of Interactive Marketing, 18, 38-52.

32.

Hess

R. L.

Jr. (2008). The impact of firm reputation and failure severity on customers’ responses to service failures. Journal of Services Marketing, 22, 385-398.

33.

Hocutt

M. A.

Bowers

M. R.

(2005). The impact of service guarantees on consumer responses in the hotel industry. Journal of Hospitality & Leisure Marketing, 13, 5-23.

34.

Hogreve

Gremler

D. D.

(2009). Twenty years of service guarantee research. Journal of Service Research, 11, 322-343.

35.

Huang

W-H.

Lin

Y-C.

Wen

Y-C.

(2010). Attributions and outcomes of customer misbehavior. Journal of Business and Psychology, 25, 151-161.

36.

Huefner

J. C.

Payne

B. L.

Otto

C. R.

Huff

S. C.

Swenson

M. J.

Hunt

H. K.

(2002). Consumer retaliation: Confirmation and extension. Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 15, 114-127.

37.

Iverson

M. A.

Rueder

M. E.

(1956). Ego involvement as an experimental variable. Psychological Reports, 2, 147-181.

38.

Jorgensen

B. K.

(1996). Components of consumer reaction to company-related mishaps: A structural equation model approach. In Corfman

K. P.

Lynch

J. G.

(Eds.), Advances in consumer research (Vol. 23, pp. 346-351). Provo, UT: Association for Consumer Research.

39.

Kashyap

(2001). The effects of service guarantees on external and internal markets. Academy of Marketing Science Review, 1(10).

40.

Kelley

S. W.

Hoffman

K. D.

Davis

M. A.

(1993). A typology of retail failures and recoveries. Journal of Retailing, 69, 429-452.

41.

Kim

J. H.

Chen

J. S.

(2010). The effects of situational and personal characteristics on consumer complaint behavior in restaurant services. Journal of Travel & Tourism Marketing, 27, 96-112.

42.

Lam

Tang

(2003). Recognizing customer complaint behavior: The case of Hong Kong hotel restaurants. Journal of Travel & Tourism Marketing, 14, 69-86.

43.

Lau

G. T.

(2001). Individual and situational factors influencing negative word of mouth behavior. Canadian Journal of Administrative Sciences, 18, 163-178.

44.

Liu

R. R.

McClure

(2001). Recognizing cross-cultural differences in consumer complaint behavior and intentions: An empirical examination. Journal of Consumer Marketing, 18, 54-74.

45.

Lovelock

C. H.

Patterson

P. G.

Walker

R. R. H.

(2011). Services marketing: An Asia-Pacific and Australian perspective (5th ed.). Frenchs Forest, New South Wales, Australia: Pearson.

46.

Marquis

Filiatrault

(2002). Understanding complaining response through consumers’ self-consciousness disposition. Psychology & Marketing, 19, 267-292.

47.

Marvin

(1992). Exemplary service guaranteed. Restaurants & Institutions, 102(21), 108-113.

48.

Mattila

A. S.

(2008). Discrete negative emotions and customer dissatisfaction responses in a casual restaurant setting. Journal of Hospitality & Tourism Research, 32, 89-101.

49.

McColl

Mattsson

Morley

(2005). The effects of service guarantees on service evaluations during a voiced complaint and recovery. Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 18, 22-50.

50.

McCollough

M. A.

Gremler

(2004). A conceptual model and empirical examination of the effect of service guarantees on post-purchase consumption evaluations. Managing Service Quality, 14, 58-74.

51.

McQuilken

Robertson

(2011). The influence of guarantees, active requests to voice and failure severity on customer complaint behavior. International Journal of Hospitality Marketing, 30, 953-962.

52.

Moorman

(1998). Market level effects of information: Competitive responses and consumer dynamics. Journal of Marketing Research, 35, 82-98.

53.

Namkung

Jang

Choi

S. K.

(2011). Customer complaints in restaurants: Do they differ by service stages and loyalty levels? International Journal of Hospitality Management, 30, 495-502.

54.

Nikbin

Ismail

Marimuthu

Abu-Jarad

I. T.

(2011). The impact of firm reputation on customers’ responses to service failures: the role of failure attributions. Business Strategy Series, 12, 19-29.

55.

Nyer

P. U.

Gopinath

(2005). Effects of complaining versus negative word of mouth on subsequent changes in satisfaction: The role of public commitment. Psychology & Marketing, 22, 937-953.

56.

Okazaki

(2009). Social influence model and electronic word-of-mouth: PC versus mobile Internet. International Journal of Advertising, 28, 439-472.

57.

Oliver

R. L.

(1997). Satisfaction: A behavioral perspective on the consumer. Boston, MA: McGraw-Hill/Irwin.

58.

Osborne

J. W.

Overbay

(2004). The power of outliers (and why researchers should always check for them). Practical Assessment, Research & Evaluation, 9(6). Retrieved from http://pareonline.net/getvn.asp?v=9&n=6

59.

Pedraja

Yagüe

(2001). What information do customers use when choosing a restaurant? International Journal of Contemporary Hospitality Management, 13, 316-318.

60.

Ping

R. A. J.

(1993). The effects of satisfaction and structural constraints on retailer exiting, voice, loyalty, opportunism and neglect. Journal of Retailing, 75, 218-241.

61.

Reichheld

F. F.

Markey

R. G.

Jr. Hopton

(2000). E-customer loyalty: Applying the traditional rules of business for online success. European Business Journal, 12, 173-179.

62.

Reichheld

F. F.

Sasser

W. E.

Jr. (1990). Zero defection: Quality comes to services. Harvard Business Review, 68, 105-111.

63.

Richins

M. L.

(1982). An investigation of consumers’ attitudes toward complaining. In Mitchell

(Ed.), Advances in consumer research (Vol. 9, pp. 502-506). Ann Arbor, MI: Association for Consumer Research.

64.

Richins

M. L.

(1983). Negative word-of-mouth by dissatisfied consumers: A pilot study. Journal of Marketing, 47, 68-78.

65.

Singh

(1988). Consumer complaint intentions and behavior: Definitional and taxonomical issues. Journal of Marketing, 52, 93-107.

66.

Singh

(1990). Voice, exit, and negative word-of-mouth behaviors: An investigation across three service categories. Journal of the Academy of Marketing Science, 18, 1-15.

67.

Singh

Wilkes

R. E.

(1996). When consumers complain: A path analysis of the key antecedents of consumer complaint response estimates. Journal of the Academy of Marketing Science, 24, 350-366.

68.

Smith

A. K.

Bolton

R. N.

Wagner

(1999). A model of customer satisfaction with service encounters involving failure and recovery. Journal of Marketing Research, 36, 356-372.

69.

Spreng

R. A.

Harrel

G. D.

Mackoy

R. D.

(1995). Service recovery: Impact on satisfaction and intentions. Journal of Services Marketing, 9, 15-23.

70.

Tabachnick

B. G.

Fidell

L. S.

(2001). Using multivariate statistics (4th ed.). Needham Heights, MA: Pearson.

71.

Tanyeri

Cobe

(2010). Won’t you be my regular? Restaurant Business, August, 26-34.

72.

Tax

S. S.

Brown

S. W.

(1998). Recovering and learning from service failures. Sloan Management Review, 40, 75-88.

73.

Tombs

A. G.

McColl-Kennedy . (2010). Social and spatial influence of customers on other customers in the social-servicescape. Australasian Journal of Marketing, 18, 120-131.

74.

Tucci

Talaga

(1997). Service guarantees and consumers’ evaluation of services. Journal of Services Marketing, 11, 10-18.

75.

Vilpponen

Winter

Sundqvist

(2006). Electronic word of mouth in online environments: Exploring referral network structure and adoption behavior. Journal of Interactive Advertising, 6(2), 71-86.

76.

Walster

Berscheid

Walster

G. W.

(1976). New directions in equity research. In Berkowitz

Walster

(Eds.), Advances in experimental social psychology (pp. 1-42), New York, NY: Academic Press.

77.

Weiner

(1980). Human motivation. New York, NY: Holt, Rinehart & Winston.

78.

Weiner

(1986). An attributional theory of motivation and emotion. New York, NY: Springer-Verlag.

79.

Wilson

McNamara

J. R.

(1982). How perceptions of a simulated physician-patient interaction influence intended satisfaction and compliance. Social Science & Medicine, 16, 1699-1704.

80.

Wirtz

(1998). Development of a service guarantee model. Asia Pacific Journal of Management, 15, 15-75.

81.

Wong

J-Y.

Tsaur

S-H.

Wang

C-H.

(2009). Should a lower-price service offer a full satisfaction guarantee? Service Industries Journal, 29, 1261-1272.

82.

Zhang

Law

(2010). The impact of e-word of mouth on the online popularity of restaurants: A comparison of consumer reviews and editor reviews. International Journal of Hospitality Management, 29, 694-700.