Service failure recovery: The moderating impact of individual-level cultural value orientation on perceptions of justice

Service failure recovery: The moderating impact of individual-level cultural value orientation on perceptions of justice

Intern. J. of Research in Marketing 23 (2006) 263 – 277 www.elsevier.com/locate/ijresmar Service failure recovery: The moderating impact of individua...

367KB Sizes 0 Downloads 34 Views

Intern. J. of Research in Marketing 23 (2006) 263 – 277 www.elsevier.com/locate/ijresmar

Service failure recovery: The moderating impact of individual-level cultural value orientation on perceptions of justice Paul G. Patterson a,⁎, Elizabeth Cowley b , Kriengsin Prasongsukarn c a b

School of Marketing, University of NSW, Sydney 2052, Australia Faculty of Economics and Business, University of Sydney, Australia c Graduate Business School, Assumption University, Thailand

Abstract It is now well recognized that an effective service recovery program is essential to generating customer satisfaction and loyalty. A number of studies have investigated the impact of service recovery efforts (compensation, speed of response, etc) on post-recovery satisfaction. However, despite the importance of global markets, none have examined the impact of customers' cultural value orientation (i.e., cultural values measured at the individual level) in implementing effective service recovery programs. In this research we use the individual rather than the nation as the unit of analysis. Using an experimental design with data from both Eastern and Western cultures, we investigate how customer evaluations of recovery efforts are influenced by interplay of consumers' value orientation and service recovery attributes (apology, cognitive control, and recovery initiation). The results reveal that cultural values of individual Power Distance, Uncertainty Avoidance and Collectivism do indeed interact with a firm's recovery tactics to influence perceptions of fairness (justice). Finally, all three forms of justice (distributive, procedural, interactional) positively impact on overall service recovery satisfaction. © 2006 Elsevier B.V. All rights reserved. Keywords: Service recovery; Individual cultural values; Satisfaction; Justice theory; Fairness

1. Introduction In today's increasingly competitive global environment the success of service firms largely depends on their capacity to consistently deliver satisfying consumption experiences. Providing consistent and error-free service, while a noble goal, is difficult to achieve in service industries. The high level of human involvement (both service provider and customer) in the ‘manufacture’ and delivery of many services means the quality of service delivery is often dependent on the attitude and behaviour of front-line employees, the expectation of customers, and even the behaviour of other customers. In addition, production and consumption of many retail services generally occur at the same time, meaning that customers interact directly with the service provider at the time service is delivered (Bloemer & Kasper, 1995; de Ruyter & Wetzels, 2000;

⁎ Corresponding author. Tel.: +61 2 9385 1105; fax: +61 2 9663 1985. E-mail address: [email protected] (P.G. Patterson). 0167-8116/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ijresmar.2006.02.004

Zeithaml, 1981). As a result, service delivery leaves little space for inspection or revision before it is provided to the customer. Even the most customer-oriented organisation and the strongest quality program will not entirely eliminate mistakes during service delivery (Kelly & Davis, 1994). Given that quality is notoriously difficult to manage in services enterprises, it is important to have an effective service recovery program in place. When a service failure occurs, the organisation's response has the potential either to restore customer satisfaction and reinforce loyalty or to exacerbate the situation and drive the customer to a competing firm (Smith, Bolton, & Wagner, 1999). Effective customer complaint handling and service recovery, can turn angry and frustrated customers into loyal ones (Boshoff, 1997; Boshoff & Leong, 1998; Michel, 2001). Evidence indicates that well considered and well enacted service recovery can overcome disappointment and enhance relationships, whereas “ineffective” service recovery can severely damage satisfaction, trust and commitment and lead to customers switching service providers (Keaveney, 1995;

264

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

Michel, 2001; Smith & Bolton, 1998; Tax, Brown, & Chandrashekaran, 1998). However, given that service encounters and service recovery are essentially social exchanges, perceptions of recovery efforts will be affected by an individual's cultural orientation. With the globalisation of markets, and technology services now more easily crossing geographical as well as cultural boundaries, there is a compelling rationale for considering cultural effects (van Birgelen, de Ruyter, de Jong, & Wetzels, 2002). Furthermore, there is considerable evidence in social psychology indicating that social exchanges and processes are culturally contingent (Fiske, Kitayama, Markus, & Nesbitt, 1998). “Culture” is defined here as the sum of all behavioural norms and patterns collectively shared by a social group (Usunier, 1996). The objective of this paper therefore is to examine the influence of customers' cultural value orientation (i.e., cultural values measured at the individual, not nation, level) and service recovery processes on the perception of fairness (justice) and post recovery satisfaction in a medium-contact service contact (resort accommodation). 1.1. Service recovery A failed service encounter is an exchange where a customer perceives a loss due to a failure on the part of the service provider. At which point, a sensitive service provider attempts to provide a gain via some recovery effort to offset the customer's loss. This view is consistent with social exchange and equity theories (Adams, 1965; Homans, 1958). Service recovery refers to the actions an organization takes in response to a service failure (Gronroos, 1988). Service recovery strategies involve actions taken by an organisation and its employees to return the customer to a state of satisfaction (Danaher & Mattsson, 1994; Sparks & McColl-Kennedy, 2001). These strategies may include acknowledgment of the problem, prompt rectification of the problem, providing an explanation for the service failure, apologising, empowering staff to resolve issues on the spot, making offers of compensation (i.e., refunds, price discounts, upgrade services, free products or services), and being courteous and respectful during the recovery process (for examples, see Bitner, 1990; Blodgett, Hill, & Tax, 1997; Hoffman & Kelly, 2000; Sparks & Callan, 1996). Clemmer and Schneider (1996) emphasised that it is important to focus on the process and relational aspects of service recovery (i.e., how the customer is treated during the recovery process), as well as the outcomes (i.e., rectifying the problem, providing compensation). Hence, both what is done (tangible compensation) and how it is done (employee interaction with the customer) influence customer perceptions of justice (Levesque & McDougall, 2000). And in the marketplace, the customer is both judge and jury when it comes to perceptions of justice. The balance of the paper is organised as follows: first we present a theoretical framework based on equity and social exchange theory within a service recovery context. Next the methodology is discussed including a description of the scenarios used in the research. Following a discussion of the

results, we conclude with key managerial and research implications. 2. Justice theory The leading theoretical perspectives of service recovery, embracing both processes and outcomes, have centred on equity theory (Bowen & Johnston, 1999; Clemmer, 1993; Smith et al., 1999; Tax et al., 1998; Tyler, 1994). Equity theory concerns individuals' perceptions of the fairness of a situation or decision (Adams, 1965). More specifically, people compare the ratio of their outputs (what they receive) to inputs (what they pay — financial and non-financial) to the ratio of the other party. If the difference is in the individual's favour, the result may be a feeling of guilt or regret. If the difference is to the individual's detriment, the result may be a feeling of disappointment or anger. The theory predicts that in both cases the individual will act to bring about a state of equilibrium. Justice theory is adapted from social exchange and equity theories. The three dimensions of justice include distributive, procedural, and interactional (refer Table 1). Distributive justice focuses on the perceived fairness of the outcome of the process and involves the policies and rules governing how recovery decisions are made (Deutsch, 1975; Smith et al., 1999). Procedural justice, which involves the perceived fairness of the means by which decisions are made and conflicts are resolved (Lind and Tyler, 1988) while interactional justice involves the manner in which the customer is treated throughout the process (e.g., information is exchanged and outcomes are communicated) (Bies & Shapiro, 1987). Research involving the three dimensions has been developed independently, and only recently been integrated into service evaluations in general and service recovery contexts specifically (see Smith et al., 1999; Tax et al., 1998). Each of the three justice dimensions is driven by various actions taken by the firm in effecting service recovery. For example, customer perceptions of fairness following compensation and an apology impact overall distributive justice, while providing cognitive control (i.e., keeping the customer informed of what is happening) impacts procedural justice. Finally, the manner in which service personnel treat a customer (politeness, respect, and courtesy) during the recovery process affects perceptions of interactional justice. A number of studies have examined the justice dimensions in various service recovery settings employing field survey or experimental methodologies. Generally, the studies have Table 1 Justice definitions Justice dimension

Definition

Distributive justice Procedural justice

Perceived fairness of the outcome (e.g., problem rectified, refund, compensation) Perceived fairness of the process employed in resolving the service failure (e.g., speed of recovery, keeping customer informed) Perceived fairness of the manner in which the customer is treated (e.g., respect, empathy, courtesy)

Interactional justice

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

concluded that higher perceived levels of justice (either alone or in combination) are positively correlated with higher postrecovery satisfaction (Hoffman, Kelly, & Rotalsky, 1995; Sparks & McColl-Kennedy, 2001; Tax et al., 1998).

265

of higher (lower) Power Distance, higher (lower) Collectivism, stronger (weaker) Uncertainty Avoidance, higher (lower) Masculinity, and higher (lower) Confucian Dynamic. 3.2. Predictions

3. Cultural orientation Hofstede (1991) defined culture as the “training or refining of one's mind from social environments in which one grew up” (p.4). Service encounters (including service recovery efforts) are social exchanges. The interaction between service-provider and customer is a crucial component of satisfaction and provides a motive to continue the relationships. As Czepiel (1990) noted “The social content of service encounters often seems to overshadow the economic”. Because a degree of social interaction between provider and customer is necessary for service recovery, it stands to reason that societal values and norms as well as expectations of both parties are likely will be influenced by one's cultural background. Prior research suggests that the importance of culture in shaping people's attitudes, values, and preferences should be particularly observable for services that require a medium-high degree of customer contact with service personnel and facilities (Choi & Markus, 1998; Mattila, 1999; Schutte & Ciarlante, 1998). More specifically, recent cross-cultural studies in marketing have shown that cultural norms and values have a strong impact on service encounter evaluations (Winsted, 1997, 1999), service quality expectations (Donthu & Yoo, 1998), referral behaviour (Money, Gilly, & Graham, 1998), and the nature and strength of customer relationships (Patterson & Smith, 2001, 2003). Given the social interaction involved in most service recovery episodes, it is logical that the cultural orientation of a consumer will impact how the customer perceives such recovery efforts. 3.1. Measuring cultural values and norms at the individual level Keillor, D'Amico, and Horton (2001) contend that any chances for marketing success will be greatly improved if attention is directed at consumer, rather than country, characteristics. And as suggested by Dawar and Parker (1994) and Leung, Bhagat, Buchan, Erez, and Gibson (2005), the term culture is not necessarily synonymous with the term country. Studies based on cultural dimensions, rather than nationality should provide greater explanatory power because they enable us to attribute differences to cultural norms and traits, and to extrapolate beyond the countries included in any sample. In this study we do not focus on individual nations, but rather we use the cultural values of the individual respondent as the unit of analysis. Using national generalisations to explain individual behaviours is an ecological fallacy because country-level relationships are interpreted as if they are applied to individuals (Yoo & Donthu, 2002). Applying Hofstede's cultural typology at the individual level is reasonable since the values of an individual person are identified in terms of the selected dimensions of culture (Donthu and Yoo, 1998). For example, a person can be described from the cultural perspective as being

Satisfaction with service recovery efforts has been shown to depend on customers' expectations and perceptions of compensation, apology, empathy and politeness displayed attributions of blame, as well as accessibility and speed of recovery in a Western context (Andreassen, 2000; Sparks & McCollKennedy, 2001; Tax et al., 1998). We argue that the impact of these attributes on justice perceptions is value dependent, and thus consumer expectations and perceptions of recovery efforts will vary, depending on cultural value orientation. As noted by Donthu and Yoo (1998), “When marketing efforts fit the culture, their impact on service quality should be greater or more noticeable” (p. 59). To build a theory taking into account the influence of culture, we must first link observed cultural value orientation differences to specific dimensions of culture that are hypothesised to have produced the differences (Leung & Bond, 1989). Our model places the justice framework in a cultural context. We hypothesise that the success of service recovery tactics is influenced by an individual consumer's value orientation along the dimension of Power Distance, Individualism–Collectivism, and Uncertainty Avoidance. These cultural dimensions interact with service recovery attributes to impact customer perceptions of distributive, procedural, and interactional justice, and thus satisfaction with the overall service recovery effort. 3.3. Cultural orientation and service recovery Power Distance is defined as the extent to which the less powerful members of a society expect and accept that power is distributed unequally (Hofstede, 1991). Power typically symbolizes a higher status, respect, more rights and wealth. In lower Power Distance cultures inequalities among people will tend to be minimized, decentralization of activities is more likely, subordinates expect to be consulted by superiors, and privileges and status symbols are less evident. On the other hand, in higher Power Distance cultures inequalities among people are considered desirable, there is greater reliance by the less powerful on those who hold power, centralization is the norm, and subordinates are likely to be separated from the superiors by wide differentials in salary, privileges and status symbols. Furthermore, the concept of ‘face’ is a fundamental norm that often accompanies high Power Distance. Face refers to social and professional position, reputation, and self image. For some individuals therefore, face is of critical importance and loss of face has disastrous personal consequences and thus avoided at all costs. The intense desire for gaining (and not losing) face suggests that an apology from a person of higher status would be more highly valued than an apology from a lower ranking employee (e.g., receptionist). Thus consumers with a higher Power Distance value orientation are more sensitive to the status of the employee

266

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

delivering an apology. Therefore we expect them to be more satisfied when an apology is given by an employee they perceive to be of higher status. Conversely, consumers with a lower Power Distance value orientation will not be as concerned whether the source of the apology is a higher or lower status representative of the firm. H1. An apology from a service provider with more status will have a greater effect on perceptions of distributive justice for customers with a higher Power Distance value orientation, compared to customers with lower Power Distance value orientation. A Collectivist (versus an Individualist) orientation is represented by a social pattern consisting of closely linked individuals who see themselves as part of one or more collectives and are willing to give priority to the collective goals over their own personal goals. Collectivists are susceptible to social influence and are concerned with their selfpresentation. Collectivists value harmony and avoid confrontation at all costs. Since complaining requires a customer to express their dissatisfaction directly to the service provider thus breaking the harmony of the situation, customers with a Collectivist orientation, therefore, should find it disturbing to voice their complaints directly and risk confrontation. Rather they would prefer the organisation to recognise that a service failure has occurred and the firm to voluntarily initiate recovery efforts (Triandis, 1995). Thus: H2. Customers with a higher Collectivist value orientation will perceive more interactional justice when there is an organisation-initiated recovery (versus self-initiated recovery) than customers with a lower Collective orientation. Uncertainty Avoidance describes the extent to which people feel threatened by ambiguous or an unknown situations (Hofstede, 1980; Patterson & Smith, 2001). Higher Uncertainty Avoidance is associated with a desire for reduction of ambiguity and a need for predictability, written rules, and structured relationships. Lower Uncertainty Avoidance is associated with a propensity to engage in risk taking behaviour. Individuals with a higher Uncertainty Avoidance are active, emotional, and security-seeking. On the other hand, individuals with lower Uncertainty Avoidance are more contemplative, less aggressive, unemotional, and generally accepting of personal risk (Triandis, 1995). Consequently, customers with a higher Uncertainty Orientation place a higher value on having cognitive control (a result of reliable follow-up, being kept informed during the recovery process) over the situation, than do their counterparts with a lower Uncertainty Avoidance orientation. Thus Hypothesis 3 is: H3. Customers with a higher Uncertainty Avoidance orientation, will, when given cognitive control over the recovery process, perceive higher levels of procedural justice than customers with a lower Uncertainty Avoidance orientation. We now include two general hypotheses. First, we predict that each of the three justice dimensions will be positively

associated with overall satisfaction with service recovery efforts, irrespective of an individual's cultural orientation. Hence our first general hypothesis is: H4a,b,c. In a service recovery situation, irrespective of an individual's cultural orientation, satisfaction will be related positively to perceptions of (a) distributive (b), procedural (c), and interactional justice. Next, consistent with the satisfaction literature, we include disconfirmation of expectations as an antecedent of overall satisfaction. Consumers are known to judge performance against their prior expectation in making satisfaction evaluations (Liljander & Strandvik, 1994; Patterson, 1993; Smith et al., 1999). When perceptions of performance exceed expectations, positive disconfirmation occurs; while negative disconfirmation results when performance outcomes are less than expected. For this reason, an overall disconfirmation variable is included as a predictor of overall satisfaction with service recovery. Further, by controlling for expectations, we provide a stronger test of our main hypotheses for the three justice dimensions. Hence: H5. In a service recovery situation, irrespective of value orientation, disconfirmation of expectation will be related positively to customer's evaluation of satisfaction with service recovery. 4. Methodology 4.1. Design To test the first three hypotheses, we used three single factor experimental designs. Each design has a separate manipulation of the service recovery. In all three experiments, the respondent reads the same description of the service failure. The manipulations appear in the description of the service recovery. Participants randomly assigned to Manipulation 1 were informed that they received an apology from either the duty manager (higher status) or the receptionist (lower status). Participants randomly assigned to Manipulation 2 were informed that the service recovery had been initiated either by the service provider or themselves. Participants randomly assigned to Manipulation 3 were informed that they had been kept informed (high cognitive control) or had not been kept informed (low cognitive control) by the service provider while waiting for a resolution. In each case, the service provider was the receptionist, except in Manipulation 1. In each case, the receptionist initiates the investigation into the source of the service failure, except in Manipulation 2. In each case, no cognitive control is present, except in Manipulation 3. The service failure was rectified in all cases. All dependent and value orientation measures were identical regardless of the manipulation. For analysis purposes the appropriate independent measure (value orientation) was used in a median split resulting in three separate 2 × 2 designs. Regression analysis on the pooled data

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

(from all manipulations) was then employed to test the relationship between the three types of justice and disconfirmation, and post-recovery satisfaction. 4.2. Sample Previous research has shown that members of a western country are typically more Individualistic, lower in Power Distance, and lower in Uncertainty Avoidance both when the dimensions are measured at country and at an individual level (for a review see Triandis & Suh, 2002). In order to maximise the variance within each dimension, we collected data in both an eastern and a western country. We used a separate design and sample to test H1 (n = 150), H2 (n = 165), and H3 (n = 172). We then pooled the data to test H4 and H5 (n = 487). We collected data from undergraduate students in Thailand (n = 246) and in Australia (n = 241). Only participants reporting English as their mother tongue in Australia and Thai in Thailand were included in the analysis. This resulted in the exclusion of 45% of the Australian sample (original n = 438). We selected these two countries because they have very different profiles using Hofstede's (1980) national scores. Without a two country sample, it could be argued that the effects we present here would hold only in a western country had we collected an Australian sample, or only in an eastern country had we collected a Thai sample. A two country sample strengthens the argument for using individual level indicators of cultural value orientation. 4.3. The stimulus In each design one aspect of the scenario was manipulated (the status of the service provider, the source of recovery initiation or the degree of the customer's cognitive control). A scenario approach is used because it avoids the problem of intentionally imposing service failures on customers and minimises memory-biases, which are common in self-reports of past service failures as Smith et al. (1999) reported. The scenarios were developed over several stages following the procedure used by Sparks and McColl-Kennedy (2001). First, depth interviews were conducted to generate service breakdown ideas suitable for the project. The scenarios developed met both of our criteria; 1] an undergraduate student sample would be familiar with the setting and 2] the service failure would be considered a core service failure. Second, scripts were drafted by the researchers. The manipulations of the status of the apologising service provider, the source of the recovery initiation, and the cognitive control of the customer were pretested on 188 undergraduate students (see Section 4.3.2). The final scenarios described a service situation where the customer made a booking in advance at a resort/hotel. The guest, or the participant, was informed upon arrival that the room had been double-booked and no other rooms were available. The scenario was written through the eyes of the reader. In addition, instructions encouraged the respondents to imagine that they were the person involved in the service encounter. Finally, the scenarios were translated into Thai and pilot tested on 198 undergraduate students (see Section 4.3.1).

267

4.3.1. Measurement equivalence To ensure item equivalence, a critical consideration in a multi-lingual study of this nature, the English version was forward translated by bilinguals whose mother language was Thai (Hambleton, 1993), and then back translated by bilingual authors whose mother language was English (Brislin, Lonner, & Thorndike, 1973). Depth interviews were conducted to establish meaning equivalence of key concepts and phrases with Thai undergraduate students. Similar reactions to the service failure in the scenario were reported by Thai (n = 100) and Australian (n = 98) undergraduate students. Both groups agreed that the scenario was realistic, that the problem presented in the scenario was major, and that they would be irritated, annoyed, and angered by the situation. This further supports the assertion that the translation resulted in psychologically equivalent stimuli. See Table 2 and Section 4.3.2 for the results of a pilot test. Before comparing the two country samples and pooling the data, we established construct and measurement equivalence between the Australian and Thai data following the procedure suggested by Steenkamp and Baumgartner (1998). The regression coefficient of one item (the marker) is set equal to one and its intercept to zero to identify a base model. Further restrictions impose configural, metric and scalar invariance of scales. Measures may be completely or only partially invariant. All scales discussed satisfy complete metric invariance. Steenkamp and Baumgartner require that the intercept of at least one item besides the marker item must be invariant in order to satisfy partial scalar invariance. At least partial scalar invariance is necessary for means of latent variables to be comparable across countries. All scales satisfy at least partial scalar invariance. Further, the scales meet at least partial error invariance and complete variance invariance. The results are shown in the Appendix. Hence measurement invariance is, by and large, established. When combined with the qualitative phase and rigorous translation and back translation procedures adopted, there is a legitimate platform to pool the country samples for analysis (Craig & Douglas, 2000; Mullen, 1995; Steenkamp & Baumgartner, 1998). 4.3.2. Manipulation checks The manipulations were pre-tested on a sample of undergraduate students in Thailand (n = 100) and in Australia Table 2 Test of equivalence of scenario between Thailand and Australia Test of equivalence

Australia N = 98

Thailand N = 100

Tvalue

Last stayed at a hotel or resort (months) How realistic was the problem that described to you? (1 = not at all realistic and 7 = very realistic) How do you view this problem? (1 = minor and 7 = major problem) How irritated would you be? (1 = not at all and 7 = very irritated) How annoyed would you be? (1 = not at all and 7 = very annoyed) How angry would you be? (1 = not at all and 7 = very angry)

10 5.30

9.2 5.22

0.43 0.46

5.94

5.87

0.38

5.95

6.21

1.64

6.05

6.07

1.43

5.52

5.99

1.80

268

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

(n = 88). The manipulation of the status of the service provider offering the apology was operationalised using the statements: “The duty manager personally apologised on the behalf of the hotel” (higher) or “The receptionist personally apologised on the behalf of the hotel” (lower). Participants read one of the two versions of the scenario and rated their agreement that the status of the person apologising was higher on a seven-point Likert scale. T-tests revealed that in the first version of the scenario participants agreed that the person apologising had higher status, but not in the second version (Thailand: Manager = 5.18, Receptionist = 3.11, t = 9.73, p < .001 and Australia: Manager = 5.08, Receptionist = 4.34, t = 12.22, p < .001). The results are presented by country because the stimuli were presented to the Thai subjects in Thai and the Australian subjects in English. The manipulation of service recovery initiation was operationalised using the statements: “before you respond or complain, the receptionist immediately acted on your behalf” (organisation) or “you complain to the front desk receptionist” (customer). Participants read one of the two versions of the scenario and rated their agreement that the organisation initiated the recovery on a seven-point Likert scale. T-tests revealed that in the first version of the scenario participants agreed that the service recovery was initiated by the organisation, but not in the second version (Thailand: organisation = 5.73, customer = 3.33, t = 41.89, p < .001 and Australia: organisation = 5.02, customer = 2.86, t = 38.28, p < .001). The manipulation of cognitive control was operationalised using the statements: “During the time you were waiting, the receptionist came to you twice to keep you informed about what was being done” (more control) or “During the time you were waiting, the receptionist did not speak to you or keep you informed” (less control). Participants read one of the two versions of the scenario and rated their agreement that they

felt they could predict what was going to happen on a sevenpoint Likert scale. T-tests revealed that in the first version of the scenario participants agreed that they had enough knowledge (cognitive control) to predict the outcome of the encounter, but not in the second version (Thailand: more control = 4.65, less control = 2.58, t = 7.44, p < .001 and Australia: more control = 5.03, less control = 2.59, t = 8.45, p < .001). 4.4. Measures 4.4.1. Independent measures The CVSCALE (Donthu & Yoo, 1998; Yoo & Donthu, 2002; Yoo, Donthu, & Lenartowicz, 2001) was used to investigate cultural orientation at the individual level. This scale has previously been used successfully to capture Hofstede's (1991) five cultural dimensions at the individual level (Donthu & Yoo, 1998; Yoo & Donthu, 2002). Only three of the five cultural dimensions were employed in this study to support the hypotheses (Confucianism and Masculinity/ Femininity culture dimensions were not relevant to the hypotheses). Power Distance, Uncertainty Avoidance and Collectivism were captured using the 16 items of the original CVSCALE. These items were then subjected to both an exploratory factor analysis (principal components analysis and varimax rotation) and a confirmatory factor analysis (CFA). Refer Table 3 for the exact wording of the items. The exploratory factor analysis resulted in a clean solution (with only one item cross loading > 0.45) and all 16 items loading on the dimensions as reported in Donthu and Yoo (1998), thus indicating convergent and discriminant validity. Table 3 reports factor loadings and variance explained. The Cronbach alpha statistic for Power Distancei, Uncertainty Avoidancei, and Individualism–Collectivismi ranged from 0.61 to 0.80. All of the distributions were approximately normal.

Table 3 Exploratory factor analysis results CVSCALE Factor loadings

1 People in higher positions should make most decisions without consulting people in lower positions 2 People in higher positions should not ask the opinion of people in lower positions too frequently 3 People in higher positions should avoid social contact with people in lower positions 4 People in lower positions should not disagree with people in higher positions 5 People in higher positions should not delegate important tasks 6 It is important to have instructions spelled out in detail 7 It is important to closely follow instructions and procedures 8 Rules and regulations are important because they inform me of what is expected 9 Standardised work procedures are helpful 10 Instructions for operations are important. 11 Individuals should sacrifice self-interest for the group 12 Individuals should stick with their group even through difficulties 13 Group welfare is more important than individual rewards 14 Group success is more important than individual success 15 Individuals should only pursue their personal goals after considering group goals 16 Group loyalty should be encouraged even if individual goals suffer Variance explained Cumulative variance explained: 57.4% Cronbach alpha

1

2

3

− .122 .193 .328 .114 .060 .010 .128 .055 .087 .243 .473 .743 .890 .851 .745 .617 29.8%

.105 .068 −.065 −.064 .203 .699 .790 .766 .817 .777 .420 .136 .098 .030 .126 .116 15.6%

.667 .513 .648 .720 .707 .017 .122 .048 .088 −.013 .112 −.176 .075 .167 .277 .297 11.9%

.73

.80

.61

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

The sixteen items were then subjected to a confirmatory factor analysis. This resulted in four items being omitted. After refinement, a final CFA model was estimated that demonstrated good measurement properties. 4.4.2. Measurement properties A CFA was conducted on the CVSCALE using LISREL 8 (Jöreskog & Sörbom, 1993) and, after refinement, showed that the hypothesized model fit the data reasonably well (see Table 4). The overall fit of the model was adequate: the Jöreskog and Sörbom (1993) GFI was .97, and the CFI was .98; only 4 (2%) of the standardized residuals were over the absolute value of three (largest positive standardized residual was 3.36 and largest negative standardized residual was − 3.45); and the chi-square was 75.3 with 40 degrees of freedom. Second, as evidence of convergent validity, the measurement factor loadings were all significant (t-values between 12.11 and 23.18), the construct reliabilities were acceptable (ranging from 0.61 to 0.80), and the Variance Extracted (Fornell & Larker, 1981) indicated that in two of the three cases the variance captured by the construct was greater than the variance due to measurement error (0.54 and 0.56) for the Power Distance dimension the variance extracted Table 4 Confirmatory factor analysis CVSCALE Construct and measures

Standardized T-value Variance parameter extracted estimate

Power distance People in higher positions should avoid 0.74 social contact with people in lower positions People in lower positions should not 0.61 disagree with people in higher positions People in higher positions should not 0.60 delegate important tasks Uncertainty avoidance It is important to have instructions spelled out in detail It is important to closely follow instructions and procedures Standardised work procedures are helpful Instructions for operations are important Collectivism Individuals should stick with their group even through difficulties Group welfare is more important than individual rewards Group success is more important than individual success Individuals should only pursue their personal goals after considering group goals Group loyalty should be encouraged even if individual goals suffer

0.43 15.3

12.7 12.1

0.54 0.56

12.9

0.79

16.8

0.74 0.92

17.4 21.4

0.57

12.9

0.87

23.2

0.86

19.2

0.76

15.0

0.64

12.7

0.56

Goodness of fit statistics: χ2 (df: 40) = 75.3, p < 0.05, GFI = 0.97, AGFI = 0.95, NFI = 0.97, CFI = 0.98, RMSEA = 0.043.

269

Table 5 Measurement items for perceived justice Variable

Item

Distributive

1. The outcome I received was fair 2. I did not get what I deserved (R) 3. In resolving the problem, the service organisation gave me what I needed 4. The outcome I received was not right (R) 1. The length of time taken to solve my problem was longer than necessary (R) 2. The service organisation showed adequate flexibility in dealing with my problem 1. The employees were appropriately concerned about my problem 2. The employees did not put the proper effort into resolving my problem (R) 3. The employees' communications with me were appropriate 4. The employees did not give me the courtesy I was due (R)

Procedural

Interactional

(R) = reverse scored.

was 0.42. This could have been increased with the deletion of an additional item, however it was retained for theoretical reasons. Finally, to test for discriminant validity, the procedure described by Fornell and Larker (1981) was used. As an indication of discriminant validity, the average variance (AVE) extracted for each construct should be higher than the squared correlation between that construct and any other construct. This test holds, since in no cases is there a squared correlation between any two constructs that is higher than either of the constructs' AVE. 4.4.3. Cultural value orientation groups Each respondent was assigned to a Power Distance group (high, low), a Collectivist group (Collectivist, Individualist), and Uncertainty Avoidance group (high, low) on the basis of three median splits. The responses for each of the three dimensions were weighted by the factor scores and summed for use in the median split. It is interesting to note that the medians for the dimensions were not significantly different between countries for any of the dimensions. One explanation for the similarity in distributions across cultures is that the population sampled was young and educated, both of which affect the degree to which values are culturally determined as Triandis and Suh (2002) reported. 4.4.4. Dependent measures The perceptions of justice and satisfaction with the service recovery were measured using the scales adapted from Smith et al. (1999) and Oliver and Swan (1989), respectively. The perception of justice has three dimensions: distributive justice, procedural justice, and interactional justice (see Table 5 for the exact wording of the questions). Distributive justice was measured with 4 items (α = .89). Procedural justice was measured with two items (r = .73). Interactional justice was measure with four items (α = .83). A CFA was also conducted with resulting measurement properties exceeding all minimum acceptable standards, and demonstrating sound convergent and discriminant validity (Fornell & Larker, 1981). Overall satisfaction with the service recovery was measured with three seven-point scales anchored with “very dissatisfied” and

270

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

“very satisfied”; “did a poor job” and “did a good job”; and “happy” and “unhappy” (α = .86). 4.5. Procedure Versions (6 in total) of the scenario were randomly assigned to participants. The first question asked how many months it had been since the respondent had last stayed in a resort/hotel. If the respondent had never stayed in a hotel or resort, and therefore would not have experienced this service setting, they would have been excluded, however there were no such instances. Respondents were asked to read the scenario and try to imagine themselves in the setting. Participants then responded to the dependent measures (perceived justice and satisfaction items) and independent measures (CV scale dimensions). The entire procedure took approximately 15 min. 5. Results 5.1. Analysis plan Past research suggests a relationship between cultural value dimensions and perceived justice. Hence, the cultural value orientations that were not included in the hypothesised interaction in each manipulation were considered a possible source of extraneous variance, and employed as covariates. This allowed us to control for the fact that sensitivity to feelings of perceived justice may differ based on other aspects of value orientation, independent of the experimental manipulations. If the covariates were significant, they remained in the analysis. For example, when testing Manipulation 1, perceived distributive justice may be affected by Individualism–Collectivism and/or Uncertainty Avoidance. In order to present a more accurate analysis of the relationship between the status of the apologizing service provider and the Power Distance orientation of the customer, both Collectivism and Uncertainty Avoidance were initially included in the analysis as covariates. Controlling for these variables provides a stronger test of our hypotheses. A MANCOVA (or MANOVA) was used to test each hypothesis with all three justice dimensions as the dependent measures. Although we did not have specific hypotheses about each justice dimension, previous research has found them to be correlated (Tax et al., 1998). Regression results for each manipulation are also provided. The regressions were run on the relevant justice dimension with the manipulation as a dummy variable, the relevant mean-centred value orientation score and the interaction of the two as independent factors. It should be noted that the regression results did not change significantly with or without mean centring. 5.2. Manipulation 1 This manipulation was designed to test the moderating influence of the Power Distance orientation on the relationship between the status of the service provider offering an

Fig. 1. Two way interaction: apology status × power distance.

apology and perceived distributive justice. The manipulation included two levels of status of employee delivering apology (higher, lower). Students (n = 150) from Australia and Thailand were randomly assigned to one of the two experimental conditions. To test H1, a MANOVA with Power Distance group and status of the person offering the apology was run on the multiple dependent measures of distributive, procedural and interactive justice.1 There was a main effect for status (F(1, 145) = 4.12, p < .05) revealing that respondents felt a greater sense of distributive justice when the apology came from an employee of higher status, regardless of their Power Distance orientation. As expected was a significant interaction with status (F(1, 145) = 21.72, p < .0001). Higher Power Distance oriented customers felt a greater sense of distributive justice when the apology was provided by an employee with higher status (Manager = 5.17, Receptionist = 4.47, t = 3.36, p < .001), whilst lower Power Distance oriented customers reported no difference in feelings of distributive justice when an apology was provided by employee with either higher or lower status (Manager = 4.44, Receptionist = 4.64). See Fig. 1. H1 is supported. 2 The manipulation also interacted with Power Distance orientation for perceptions of procedural (F(1, 145) = 13.56, p < .0001) and interactional justice (F(1, 145) = 5.35, p < .05). In these cases, lower Power Distance oriented individuals perceived more procedural and interactional justice when the receptionist (or lower status employee) delivered the apology. The result is interesting because it provides evidence for independence of distributive justice. In addition, the main effect for Power Distance orientation was significant (F(1, 144) = 6.28, p < .05). Power Distance was not hypothesised to have an effect on distributive justice here,

1 Value orientation for individualism and uncertainty dimensions were initially included as covariates in the analysis, but were non-significant and therefore removed. 2 A regression run on perceived distributive justice with the mean-centred power distance score, the status manipulation as a dummy variable, and the interaction of the two reveals the same pattern of results. Status (b = .22, p < .0001) and the interaction term (status * power distance) (b = .20, p < .05) were significant, whilst power distancei orientation was not. (b = − .05, p = .48).

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

271

although the effect has been found elsewhere (see Fields, Pang, & Chiu, 2000). 5.3. Manipulation 2 This manipulation was designed to test the moderating influence of Individualism–Collectivism orientation on the relationship between recovery initiation and perceived interactional justice. The manipulation included two levels of recovery initiation (organization, customer). Respondents (n = 165) were randomly assigned to one of the two experimental conditions. To test H2 a MANCOVA with Individualism–Collectivism and recovery initiation with Power Distance score as a covariate was run on the multiple dependent measures of interactional, procedural, and distributive justice.3 There was a main effect for initiation (F(1, 160) = 12.65, p < .0001) revealing that respondents felt a greater sense of interactional justice when the service recovery was initiated by the organization, in this case the service provider, regardless of their Individualism–Collectivism orientation. There was a weak interaction between service recovery initiation and Individualism–Collectivism orientation (F(1, 160) = 2.64, p = .10). Customers with a Collectivist orientation felt a greater sense of interactional justice when service recovery was initiated by the organization (organization = 5.72, customer = 4.97, t = 4.28, p < .0001). Individualist oriented customers, on the other hand, did not report any significant increase in feelings of interactional justice whether recovery was initiated by the organization or customer (organization = 5.11, customer = 5.01). H2 is supported.4 See Fig. 2. The manipulation also interacted with Collectivism for perceptions of procedural justice. Customers with a Collectivist orientation felt a greater sense of procedural justice when service recovery was initiated by the organization (F(1, 160) = 6.71, p < .05). There was no effect for perceptions of distributive justice. 5.4. Manipulation 3

Fig. 2. Two way interaction: recovery initiate × collectivisn/individualism.

tional justice.5 There was a main effect for control (F(1, 168) = 17.42, p < .0001) revealing that respondents felt a greater sense of procedural justice when they had more cognitive control, regardless of their uncertainty orientation. There was no effect for Uncertainty Avoidance orientation (F(1, 168) = 0.04, p = .84) but, as expected, there was a significant interaction between control and Uncertainty Avoidance orientation (F(1, 168) = 8.80, p < .01). Customers with a higher Uncertainty Avoidance orientation felt a greater sense of procedural justice when the employee kept them informed of what was happening during the service recovery process (high = 5.21, low = 3.96, t = 4.37, p < .0001), whilst those with lower Uncertainty Avoidance orientation did not report a significantly greater feeling of procedural justice when the employee kept them informed (high = 4.73, low = 4.52). H3 is supported.6 See Fig. 3. The manipulation also had an effect on perceptions of distributive justice. Customers with a stronger Uncertainty Avoidance orientation felt a greater sense of distributive justice (F(1, 168) = 5.97, p < .05) when kept informed. There was no effect for perceptions of interactional justice. This provides evidence for the independence of the interactional justice dimension. 5.5. The influence of perceived justice on service recovery satisfaction

This manipulation was designed to test the moderating influence of Uncertainty Avoidance on the relationship between cognitive control and perceived procedural justice. The manipulation included two levels of cognitive control (higher, lower). Respondents (n = 172) were randomly assigned to one of the two experimental conditions. To test H3 a MANOVA with Uncertainty Avoidance orientation and cognitive control was run on the multiple dependent measures of procedural, distributive, and interac-

To test for replication of the relationship between perceived justice and satisfaction found in previous research, the pooled data from Australia (n = 241) and Thailand (n = 246) was analysed using multiple regression analysis. The adjusted R2 values for the service recovery satisfaction equations are .64 for pooled data and .65 and .63 for Australia and Thailand, respectively, (see Table 6). The regression results which demonstrate a highly consistent pattern of results for both

3 The covariate for power distancei also had a significant effect on interactional justice, as has been found in past research (White, Tansky, & Baik, 1995). The coviariate for uncertainty orientationi was initially included in the analysis, but was non-significant and was, therefore, removed. 4 A regression run on perceived interactional justice with the mean-centred collectivismi score, the initiation manipulation as a dummy variable, and the interaction of the two reveals the same pattern of results. Initiationi (b = .20, p < .01) and the interaction term (b = .15, p < .05) were statistically significant, whilst collectivismi orientation was not (b = .12, p = .13).

5 Power distancei and individualism–collectivismi scores were initially included as covariates in the analysis, but were non-significant and were, therefore, removed. 6 A regression run on perceived procedural justice with the mean-centred uncertainty avoidancei score, the cognitive control manipulation as a dummy variable, and the interaction of the two reveals the same pattern of results. Cognitive controli (b = .30, p < .0001) and the interaction term (b = .16, p < .05) were significant, while the cultural orientation of uncertainty avoidancei (b = −.07, p = .33) was not.

272

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

Fig. 3. Two way interaction: cognitive control × uncertainly avoidance.

countries, together with a Chow test (Table 6), justify pooling the data. For the pooled model, perceptions of distributive, procedural, and interactional justice were all significantly positively associated with customer satisfaction. Distributive justice had the largest standardised coefficient (β = .292) followed by interactional and procedural justice (β = .230 and .142, respectively). As predicted, the three justice dimensions and disconfirmation (of expectations) explain a considerable amount of variance in satisfaction with the service recovery effort, across each country sample and for the pooled data (adjusted R2 for Australia = .65, Thailand = .63 and pooled data = .64). Thus H4 and H5 are supported. 6. Discussion This research contributes to the services marketing and consumer behaviour literature by shedding light onto the role of cultural value orientation in influencing customer responses to service failures. We focus on service failures because mistakes are an unavoidable feature of all human endeavours, including service delivery (e.g., Boshoff, 1997). Although variations in customers' perceptions of service recovery efforts might sometimes be caused by differences in customs, we suggest here that they might also be produced by fundamental differences in orientations toward the world. Hofstede's (1980) cultural values typology is a useful tool for understanding an individual's fundamental cultural orientation. In this research three key dimensions of that typology (Individualism–Collectivism, Power Distance, and Uncertainty Avoidance) proved valuable in understanding a customer's perceptions of an organisation's service recovery efforts. The implications for marketing theory are fourfold. First, a general, but nonetheless important theoretical contribution lies in use of the individual as the unit of analysis when investigating cultural value orientations. Had we used the country as the unit of analysis, the pattern of results would have suggested that value orientations did not impact perceptions of justice. The relationship between justice perceptions and value orientations emerges only when value orientations are investigated at the individual level. This supports an ecological fallacy that in today's global economy the term culture is synonymous with the term country. Instead, as noted by Kwok and Uncles

(2005), some individuals in every country hold beliefs and values that are different from those typical of the country. Hence the chances of success will be improved if marketing effort is directed at the individual rather than the country stereotype. The next contribution is both methodological and conceptual. Although we did not have specific hypotheses for all three justice dimensions for each manipulation, we analysed the data using MANOVAs or MANCOVAs. It is interesting that the status manipulation affected perceptions of distributive justice differently than it affected perceptions of procedural and interactional justice. It is also noteworthy that the recovery initiation manipulation affected perceptions of distributive and procedural, but not interactional justice, while the cognitive control manipulation affected perceptions of distributive and interactional, but not procedural justice. Previous research has suggested that the high correlations between the justice dimensions indicates that they are not independent (Tax et al., 1998), conversely, the results presented here provide evidence for their independence. Our results however are consistent with Smith et al. (1999) who also found all recovery attributes had some effect on each of the justice elements. A third contribution of the research reported here is that we included the remaining two cultural values as covariates in testing each hypothesis. In H2 where we predicted that Individualism–Collectivism orientation would interact with recovery reaction, we controlled for Uncertainty Avoidance and Power Distance, with the latter being statistically significant. In this case at least two cultural norms simultaneously helped in explaining customers perceptions of service recovery. However in the analysis for H1 and H3, the unhypothesised cultural value orientations did not explain justice perceptions. This is interesting because cultural values have often been shown as not being independent of each other (Tax et al., 1998). Perhaps the independence of the cultural values is more pronounced when measured at the individual level. A fourth contribution is that the results show that distributive and interactional justice (fairness) account for a relatively large percentage of the explained variance in perceptions of overall satisfaction with recovery efforts. This is partly consistent with the findings of Smith et al. (1999) and Oliver (1997) who found distributive justice to be far and away the strongest predictor. Our results suggest that (at least in a medium contact hotel/ resort setting) distributive is only marginally more powerful than interactional fairness. We suggest that this occurs because Table 6 Regression analysis — service recovery satisfaction model Variable

Disconfirmation Distributive justice Procedural justice Interactional justice Adjust R2 F-statistic

Pooled data

Australia

Thailand

β

Sig.

β

Sig.

β

Sig.

.336 .292 .142 .230 .64 220.01

.00 .00 .00 .00

.294 .298 .136 .211 .65 115.01

.00 .00 .00 .00

.374 .205 .124 .233 .63 104.97

.00 .00 .02 .00

.00

.00

.00

A Chow test was used to test for any significance differences in the form (or slope) of the two regression models (Australia and Thailand). Results were not significantly different thus justifying pooling of the two country data (F 2.37, p >0.05).

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

it is easier for customers to access information on outcomes and assess how they are treated (interactional fairness) than procedural considerations (Leventhal, 1980). Interestingly however, our results and those of Smith et al. (1999) are at odds with those of Bowman and Narayandas (2001) who found procedural fairness to be three and a half times more powerful as a predictor of satisfaction than either procedural or interactional fairness. Their study focused on consumer products which may reflect a longer term perspective taken by customers regarding their relationship with the brand in question; or distributive outcomes (replacement of defective product, face value coupon, etc) may be taken for granted in a consumer products complaint situation. Perhaps the impact of the three justice elements differs across contexts. Future research could examine whether these effects occur in professional services (legal, financial, architectural) where outcomes are typically critical, but where there is still considerable face to face interaction between the parties, or in a B2B setting where relationships are important, but the core product is of considerable financial value. 6.1. Managerial implications The increasing globalization of markets and the ease with which services now cross national boundaries provides a compelling reason for understanding the cultural context of consumer behaviour (Maheswaran & Shavitt, 2000). Because the quality of interpersonal interaction between the customer and front-line employee drives customer evaluations service organizations, be they operating in multi-ethnic regions/ countries such as Europe, North America or the Asia–Pacific, or involved in international ventures, need to be particularly sensitive to the cultural diversity of their customer base. Our results clearly show that customers' perceptions of recovery efforts are significantly affected by the interaction effect of a firm's recovery tactics with their own value orientation. In other words, consumer expectations and perceptions of recovery efforts vary, depending on an individual's value orientation. An implication of this general finding is simply that firms have a better chance of implementing more appropriate recovery actions if employees are sensitive to individual customers' cultural value orientations. While it is unrealistic to expect front line employees to instantly assess a (complaining) customer's orientation, such information can be embedded in CRM systems. Today technology has enabled firms to segment customers into finer and finer categories to the extent that customers are no longer equal in the eyes of the firm. Technology has created a radical new business model that alters the dynamic of customer service. It is commonplace now for companies to measure how much each customer spends as well as other behavioural, demographic and psychographic data (Brady, 2000). As Brady (2000) noted technology has a dark side — where firms amass a mountain of information that provides an almost Orwellian view of individual buyers. Starwood Hotels and Resorts, Charles Schwab, Qantas and Singapore Airlines, just to name a few, systematically capture a wide array of such information so that responses to service requests can be segmented and customized.

273

Hence with sophisticated CRM systems and instant online access to customer profiles, there is no reason why cultural value orientations could not be captured via survey research (using for example the CVSCALE) and added to individual customer profiles. This could then be instantly accessed whenever a customer makes a booking, an enquiry or a complaint, thus providing guidance as to the appropriate action to adopt. This profile might only be kept for high net worth or frequent customers. In this way the findings of our study might usefully guide management actions. Each of the key findings is now discussed in turn. The three cultural orientation hypotheses were all supported. First, the interaction between Power Distance and the status of an employee delivering an apology was significant in explaining higher levels of perceived distributive justice. That is, consumers with a high Power Distance orientation were considerably more likely to view a fair outcome when an apology came from an employee seen as having higher (e.g, duty manager), rather than lower status in the organisation. On the other hand, respondents with a low Power Distance orientation are more egalitarian in their outlook: it does not matter to them whether the apology comes from an employee of higher or lower status. The management implications are clear. When offering an apology to consumers with a high Power Distance orientation, that apology should be from a high status employee (store manager, duty manager, high ranking supervisor) rather than a lower level employee (receptionist, waiter). Such actions meet the intense desire for regaining ‘face’ by the offended party. From the point of view of training front line personnel, the first step is to identify customers with a high Power Distance orientation. Of course this may not be possible in a brief service encounter episode. It might also be assumed that some 5-star service providers (5-star hotels, flying first class, upmarket private hospitals, exclusive club memberships, etc.) typically attract status-conscious people, who are more prone to perceiving unequal power and influence distributions in society. Hence it these types of establishments, good practice might be that final apology comes on from a senior manager, even if the initial apology comes from a front-line employee. Second, while there was a main effect when the organisation initiated the service recovery (rather than the customer), the inclusion of Collectivist versus Individualist orientation brings into stark contrast the role of cultural values. That is, as hypothesised, consumers with a Collectivist orientation perceived greater interactional fairness (than did their more Individualistic counterparts) when the recovery was initiated by the offending organisation. In short, Collectivists were happier when the organisation took the initiative. From a managerial standpoint, front-line staff need to be trained to be proactive in initiating recovery efforts. In the case of consumers with a high Collectivist orientation they may in fact not complain due to cultural norms that inhibit complaining behaviour. But a non-complaint response does not necessarily mean customers are happy — they may leave unhappy, spread negative word-of-mouth, and engage a different service supplier next time. Hence it is in the organisations best interests to be proactive and initiate service recovery. This may be particularly

274

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

6.2. Future research

relevant to retail service providers offering low risk investments products, banks for example, because Collectivists are more cautious about saving money (Dutta-Bergman & Wells, 2002). Next, we found that respondents with a higher Uncertainty Avoidance orientation were more likely than their lower uncertainty counterparts, to experience procedural fairness when they were kept informed of the progress being made to resolve their complaint. If upon accessing a (complaining) customer's profile, it showed a high Uncertainty Avoidance orientation, then the appropriate action could immediately be flagged in the online system to front line employees. In more general terms, providing information that encourages feelings of cognitive control may be most useful in high credence services (Zeithaml, 1981) such as professional services (financial planning, health, education, child minding, legal), or for services such as travel agents and tour operators because consumers high in Uncertainty Avoidance are most likely to use these services (Money & Crotts, 2003).

Clearly, there is a need for further research in order to resolve the issue of whether other recovery attributes (i.e., when compensation is provided and not provided, when recovery is provided in a formal or less formal manner) interact with value orientations and whether the interaction of related service recovery attributes (i.e., apology from high/low status with/ without compensation; recovery initiate by firm/customer with formal/less formal manner provided by the frontline staff) combine to impact on consumer expectations and perceptions of recovery efforts. The moderating impact of other individual differences (criticality of the service encounter, personality, etc.) might also be examined, as might a study of service types with different characteristics (e.g., professional services). Research efforts of this nature will further aid our understanding in designing appropriate service recovery procedures that transcend cultural borders.

Appendix A. CVSCALE — Test of invariance Step

1 2 3 4 5 6 7 8 9 10

Level of invariance test

Model

Equality of sigma and mu Equality of sigma Equality of mu Initial configural invariance Full matrix invariance Final partial matrix invariance Initial partial scalar invariance Initial partial factor variance–covariance invariance Initial partial error variance invariance Final partial error variance invariance

Step 6–step 4 Step 7–step 6 Step 8–step 7 Step 10–step 8

Fit indexes

GFI

Chi-square value

df

RMSEA

Test of close fit (p-value)

CAIC

CFI

TLI

311.79 277.53 310.92 178.88 222.80 198.92 205.01 205.49 268.99 218.73

110 107 107 86 98 95 104 107 119 113

0.0870 0.0811 0.0887 0.0667 0.0725 0.0672 0.0633 0.0616 0.0721 0.0621

0.0464 0.1385 0.0318 0.6121 0.4165 0.5943 0.7383 0.7807 0.4170 0.7742

814.97 802.27 835.66 682.06 639.72 637.40 751.32 730.23 707.47 700.34

N/A 0.923 N/A 0.958 0.944 0.954 0.955 0.956 0.937 0.954

N/A 0.905 N/A 0.936 0.925 0.936 0.943 0.946 0.930 0.946

0.906 0.931 0.920 0.926 0.926 0.926 0.912 0.923

Δχ2

Δdf

χ2α=.05

Decision

χ2α=.025

Decision

χ2α=.01

Decision

20.04 6.10 0.47 13.24

9 9 3 6

16.5190 16.5190 7.8147 12.5916

Sig. Not sig. Not sig. Sig.

19.0228 19.0228 9.3584 14.4494

Sig. Not sig. Not sig. Not sig.

21.666 21.666 11.3449 16.8119

Not sig. Not sig. Not sig. Not sig.

Item

Factor loadings Aus

Thai

Aus

Thai

Aus

Thai

Aus

Thai

KA = IN

P3

0.57 (9.58) 0.59 (12.13) 0.44 (7.10) 0.55 (12.65) 0.80 (16.91) 0.64 (12.10) 0.92 (20.98) 0.64 (15.44) 0.90 (24.94)

0.93 (14.62) 0.59 (12.13) 0.73 (10.22) 0.55 (12.65) 0.80 (16.91) 0.82 (14.36) 0.92 (20.98) 0.64 (15.44) 0.90 (24.94)

0.41 (8.36) 0.53 (7.92) 0.53 (8.98) 0.82 (10.81) 0.36 (6.99) 0.46 (11.45) 0.01 (0.23) 0.60 (14.61) 0.13 (4.07)

0.41 (8.36) 0.77 (9.80) 0.74 (8.67) 0.58 (10.81) 0.36 (6.99) 0.46 (11.45) 0.29 (4.46) 0.60 (14.61) 0.23 (6.14)

1.17 (26.70) 2.01 (31.92) 2.15 (37.31) 3.25 (63.75) 3.75 (64.58) 4.08 (73.31) 4.45 (69.27) 3.32 (61.74) 3.13 (50.41)

1.17 (26.70) 2.01 (31.92) 2.15 (37.31) 3.25 (63.75) 3.75 (64.58) 4.08 (73.31) 4.45 (69.27) 3.32 (61.74) 3.13 (50.41)

3.69 (12.82)

2.61 (14.87)

3.04 (19.93)

6.13 (18.85)

4.52 (19.15)

5.19 (26.66)

3.88 (19.61)

4.11 (18.98)

3.97 (27.09)

P4 P5 U1 U2 U4 U5 C2 C3

Error variances

Item intercepts

Latent means

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277 C4

0.84 (22.72) 0.73 (18.42) 0.64 (15.17)

C5 C6

0.84 (22.72) 0.73 (18.42) 0.64 (15.17)

0.23 (8.16) 0.46 (14.26) 0.59 (13.85)

0.34 (8.46) 0.46 (14.26) 0.59 (13.85)

3.08 (51.67) 2.92 (51.92) 2.97 (54.44)

275

3.08 (51.67) 2.92 (51.92) 2.97 (54.44)

P = Power Distance; U = Uncertainty Avoidance; C = Collectivism. Justice dimensions — text of invariance

Step Level of invariance test

Model

Fit indexes

GFI

Chi-square df RMSEA Test of close CAIC value fit (p-value) 1 2 3 4 5 6 7 8 9 10 11 12

Equality of sigma and mu Equality of sigma Equality of mu Final configural invariance Full metric invariance Final Partial Metric Invariance Full scalar invariance Final partial scalar invariance Full factor variance–covariance invariance Final partial factor variance–covariance invariance Full error variance invariance Final partial error variance invariance (diagonal)

228.33 110.03 226.01 108.49 LX = IN 127.01 Free LX(4,1) 116.90 TX = IN 133.13 TX = IN, free LX(4,1) 123.28 PH = IN 134.67 PH = IN, free LX(4,1) 124.93 ΘAus = ΘThai 143.41 ΘAus = ΘThai and free LX(4,1) 137.55

70 53 67 50 60 59 67 66 70 69 80 79

0.0966 0.0666 0.0989 0.0695 0.0679 0.0636 0.0638 0.0598 0.0617 0.0578 0.0572 0.0553

0.015 0.566 0.010 0.505 0.555 0.653 0.663 0.768 0.732 0.805 0.846 0.869

659.63 519.76 678.87 539.79 486.42 483.51 585.99 583.33 565.96 563.41 502.84 504.15

CFI

TLI

N/A 0.972 N/A 0.972 0.969 0.973 0.969 0.973 0.969 0.973 0.969 0.971

N/A 0.953 N/A 0.950 0.953 0.958 0.958 0.963 0.960 0.965 0.965 0.967

Independence CAIC Saturated CAIC

1 2 3 4 5

0.9263 0.9571 0.9258 0.9580 0.9514 0.9543 0.9514 0.9542 0.9510 0.9537 0.9469 0.9482 2387.69 790.71

Different chi-square values

Δχ2

Δdf

p-value (α = .05)

Decision

Step 5 (full metric)–step 4 (configural) Step 6 (partial metric)–step 4 (configural) Step 8 (partial scalar)–step 6 (partial metric) Step 10 (partial factor Var/Cov)–step 8 (partial scalar) Step 12 (partial error variance)–step 10 (partial factor Var/Cov)

18.52 8.41 6.38 1.65 12.62

10 9 7 3 10

18.31 16.92 14.07 7.81 18.31

Sig. Not sig. Not sig. Not sig. Not sig.

Item

Factor loadings Aus

Thai

LX = IN

Aus

Thai

TD = IN

Aus

Thai

TX = IN

Aus

Thai

D1

0.57 9.25 0.70 11.48 0.55 8.54 0.62 10.35 0.71 10.75 0.81 11.85 0.77 13.15 0.72 11.69 0.77 13.27 0.90 16.93

0.66 10.62 0.69 11.57 0.64 10.22 0.85 14.28 0.72 9.55 0.71 9.85 0.66 11.06 0.68 11.97 0.62 10.12 0.83 14.84

0.62 14.04 0.69 16.24 0.60 13.38 free N/A 0.71 14.36 0.76 15.40 0.71 17.19 0.70 16.71 0.70 16.75 0.87 22.67

0.61 9.89 0.51 8.89 0.66 9.80 0.52 8.28 0.48 7.13 0.40 5.45 0.45 9.12 0.57 9.78 0.43 8.94 0.20 5.47

0.63 9.93 0.53 9.59 0.62 9.47 0.37 7.16 0.49 5.59 0.44 5.39 0.53 9.61 0.45 9.17 0.60 9.71 0.30 6.64

0.62 13.94 0.52 13.05 0.64 13.66 free N/A 0.49 9.19 0.42 7.58 0.50 13.34 0.51 13.42 0.52 13.33 0.24 8.33

3.74 44.47 2.55 29.84 3.70 42.59 4.02 48.73 2.53 39.52 4.30 64.81 3.50 53.14 3.15 46.76 3.08 47.39 2.90 44.65

3.70 56.05 2.62 40.88 3.69 56.77 3.88 57.97 2.62 40.71 4.32 69.55 3.49 55.78 3.16 51.86 3.04 47.82 2.90 45.65

3.73 66.06 2.59 44.04 3.69 66.55 3.93 66.94 2.61 44.07 4.35 71.69 3.51 61.60 3.18 56.89 3.07 54.65 2.92 46.81

5.57 (16.21)

4.63 (17.85)

4.67 (14.52)

4.93 (14.05)

3.90 (17.49)

4.41 (16.66)

D2 D3 D4 P1 P2 I1 I2 I3 I4

Error variances

Item intercepts

D = Distributive justice; P = Procedural justice; I = Interactional justice.

Latent means

276

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277

References Adams, J. S. (1965, October). Towards an understanding of inequity. Journal of Abnormal and Social Psychology, 67, 422−436. Andreassen, T. W. (2000). Antecedents to satisfaction with service recovery. European Journal of Marketing, 34(1/2), 156−175. Bies, R. J., & Shapiro, D. L. (1987). Interactional fairness judgements: The influence of casual accounts. Social Justice Research, 1(2), 199−218. Bitner, M. J. (1990, April). Evaluating service encounters: The effects of physical surroundings and employee responses. Journal of Marketing, 54, 69−82. Blodgett, G. J., Hill, J. D., & Tax, S. S. (1997). The effects of distributive, procedural and interactional justice on postcomplaint behaviour. Journal of Retailing, 73(2), 185−210. Bloemer, J. M., & Kasper, H. (1995). The complex relationship between consumer satisfaction and loyalty. Journal of Economic Psychology, 16, 311−329. Boshoff, C. (1997). An experimental study of service recovery options. International Journal of Service Industry Management, 8(2), 110−130. Boshoff, C., & Leong, J. (1998). Empowerment, attribution and apologising as dimensions of service recovery: An experimental study. International Journal of Service Industry Management, 9(1), 24−47. Bowen, D. E., & Johnston, R. (1999). Internal service recovery: Developing a new construct. International Journal of Service Industry Management, 10 (2), 118−131. Bowman, D., & Narayandas, D. (2001, August). Managing customer-initiated contacts with manufacturers: The impact on share of category requirements and word-of-mouth behaviour. Journal of Marketing Research, 38, 281−297. Brady, D. (2000, October 23). Why service stinks. Business Week. Brislin, R. W., Lonner, W. J., & Thorndike, R. M. (1973). Cross-cultural research methods. New York: Wiley. Clemmer, E. C. (1993). An investigation into the relationship of fairness and customer satisfaction with services. In R. Cropanzano (Ed.), Justice in the workplace: Approaching fairness in human resource management (pp. 83−92). Hillsdale, NY: Lawrence Erlbaum Associates. Clemmer, E. C., & Schneider, B. (1996). Fair service. In T. A. Swartz, D. E. Bowen, & S. W. Brown (Eds.), Advances in services marketing and management, vol. 5 (pp. 213–229). Greenwich, CT: JAI Press. Choi, J., Markus, H. (1998). Implicit theories and casual attribution east and west. Unpublished manuscript, University of Michigan. Craig, C. S., & Douglas, S. (2000). International marketing research. West Sussex: John Wiley and Sons. Czepiel, J. A. (1990). Service encounters and service relationships: Implication for research. Journal of Business Research, 20, 13−21. Danaher, P. J., & Mattsson, J. (1994). Customer satisfaction during the service delivery process. European Journal of Marketing, 28, 5−16. Dawar, N., & Parker, P. (1994, April). Marketing universals: consumers use of brand name, price, physical appearance, and retail reputation as signals of product quality. Journal of Marketing, 58, 81−95. de Ruyter, J. C., & Wetzels, M. (2000). Customer equity considerations in service recovery: A cross-industry perspective. International Journal of Service Industry Management, 11(1), 91−108. Deutsch, M. (1975). Equity, equality, and need: What determines which value will be used as the basis of distributive justice? Journal of Social Issues, 31 (3), 137−149. Donthu, N., & Yoo, B. (1998). Cultural influences on service quality expectations. Journal of Service Research, 1(2), 178−186. Dutta-Bergman, M. J., & Wells, D. (2002). The values and lifestyles of Idiocentrics and Allocentrics in an individualist culture: A descriptive approach. Journal of Consumer Psychology, 12(3), 231−242. Fields, D., Pang, M., & Chiu, C. (2000). Distributive and procedural justice as predictors of employee outcomes in Hong Kong. Journal of Organizational Behaviour, 21(5), 547−562. Fiske, A., Kitayama, S., Markus, H., & Nesbitt, R. (1998). Cultural matrix of social psychology. In D. Gilbers, S. Fiske, & G. Lindrey (Eds.), The handbook of social psychology, vol. 2 (Fourth Edition) (pp. 915–981). New York: Guildford Press.

Fornell, C., & Larker, D. F. (1981, February). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39−50. Gronroos, C. (1988, Winter). Service quality: The six criteria of good perceived service quality. Review of Business, 9, 3−32. Hambleton, R. K. (1993). Translating achievement tests for use in crossnational studies. European Journal of Psychological Assessment, 9, 57−68. Hoffman, D. K., & Kelly, S. W. (2000). Perceived justice needs and recovery evaluation: A contingency approach. European Journal of Marketing, 34(3/ 4), 418−428. Hoffman, D. K., Kelly, S. W., & Rotalsky, H. M. (1995). Tracking service failures and employee recovery efforts. The Journal of Services Marketing, 9(2), 49−52. Hofstede, G. (1980). Culture's consequence: International differences in workrelated values. Beverly Hills, CA: Sage. Hofstede, G. (1991). Cultures and organization: Software of the mind. Berkshire, UK: McGraw-Hill. Homans, G. C. (1958). Social behavior as exchange. American Journal of Sociology, 63, 597−606. Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: A guide to the program and applications. : Scientific Software International, Inc. Keaveney, S. M. (1995, April). Customer switching behaviour in service industries: An exploratory study. Journal of Marketing, 59, 71−82. Keillor, B., D'Amico, M., & Horton, V. (2001). Global consumer tendencies. Psychology and Marketing, 18(1), 1−19. Kelly, S. W., & Davis, M. A. (1994). Antecedents to customer expectations for service recovery. Journal of the Academy of Marketing Science, 22(1), 52−61. Kwok, S., & Uncles, M. (2005). Sales promotion effectiveness: The impact of consumer differences at an ethnic-group level. Journal of Product and Brand Management, 14(3), 170−186. Leventhal, G. S. (1980). What should be done with equity theory? New approaches to the study of fairness in social relationships. In K. J. Gergen, M. S. Greenberg, & R. H. Willis (Eds.), Social exchange: Advances in theory and research, vol. 27-55. NY: Plenum Press. Levesque, T. J., & McDougall, G. H. G. (2000). Service problems and recovery strategies: An experiment. Canadian Journal of Administrative Sciences, 17 (1), 20−37. Leung, K., Bhagat, R., Buchan, N., Erez, M., & Gibson, C. B. (2005). Culture and international business: Recent advances and their implications for future research. Journal of International Business Studies, 36, 357−378. Leung, K., & Bond, M. H. (1989, June). On the empirical identification of dimensions for cross-cultural comparisons. Journal of Cross-Cultural Psychology, 20, 133−151. Liljander, V., & Strandvik, T. (1994). Estimating zones of tolerances in perceived service quality. International Journal of Service Industry Management, 4, 6−28. Lind, A. E., & Tyler, T. R. (1988). The social psychology of procedural justice. New York: Plenum Press. Maheswaran, D., & Shavitt, S. (2000). Issues and new directions in global consumer psychology. Journal of Consumer Psychology, 9(2), 59−66. Mattila, A. S. (1999). The role of culture in the service evaluation process. Journal of Service Research, 1(3), 250−261. Michel, S. (2001). Analyzing service failures and recoveries: A process approach. International Journal of Service Industry Management, 12(1), 20−33. Money, B. R., & Crotts, J. C. (2003, April). The effect of uncertainty avoidance on information search, planning, and purchases of international travel vacations. Tourism Management, 24, 191−202. Money, B. R., Gilly, M. C., & Graham, J. L. (1998, October). Explorations of national culture and word-of-mouth referral behavior in the purchase of industrial services in the United States and Japan. Journal of Marketing, 62, 76−87. Mullen, M. R. (1995). Diagnosing measurement equivalence in cross-national research. Journal of International Business Studies, 26(3), 573−596. Oliver, R. L. (1997). Satisfaction. A behavioural perspective on the consumer. McGraw-Hill.

P.G. Patterson et al. / Intern. J. of Research in Marketing 23 (2006) 263–277 Oliver, R. L., & Swan, J. E. (1989, December). Equity and disconfirmation perceptions as influences on merchant and product satisfaction. Journal of Consumer Research, 16, 372−383. Patterson, P. G. (1993). Expectations and product performance as determinants of satisfaction with a high-involvement purchase. Psychology and Marketing, 10(5), 449−465. Patterson, P. G., & Smith, T. (2001). Modelling relationship strength across service types in an eastern culture. International Journal of Service Industry Management, 12(2), 90−113. Patterson, P. G., & Smith, T. (2003). A cross-cultural study of switching barriers and propensity to stay with service providers. Journal of Retailing, 70, 107−120. Schutte, H., & Ciarlante, D. (1998). Consumer behavior in Asia. Hampshire: Macmillan Business. Smith, A. K., & Bolton, R. N. (1998). An experimental investigation of customer reactions to service failure and recovery encounter. Journal of Service Research, 1(1), 65−81. Smith, A. K., Bolton, R. N., & Wagner, J. (1999, August). A model of customer satisfaction with service encounters involving failure and recovery. Journal of Marketing Research, 36, 356−372. Sparks, B. N., & Callan, V. J. (1996). Service breakdowns and service evaluations: The role of customer attributions. Journal of Hospitality and Leisure Research, 4, 3−24. Sparks, B. N., & McColl-Kennedy, J. R. (2001). Justice strategy options for increased customer satisfaction in a services recovery setting. Journal of Business Research, 54, 209−218. Steenkamp, J. B., & Baumgartner, H. (1998, June). Assessing measurement invariance in cross-cultural research. Journal of Consumer Research, 25, 78−90. Tax, S. S., Brown, S. W., & Chandrashekaran, M. (1998, April). Customer evaluations of service complaint experiences: Implications for relationship marketing. Journal of Marketing, 62, 60−76.

277

Triandis, H. C. (1995). Individualism and collectivism. Boulder, Co.: Westview Press. Triandis, H. C., & Suh, E. M. (2002). Cultural influences on personality. Annual Review of Psychology, 53, 133−160. Tyler, T. R. (1994). Psychological models of the justice motive: Antecedents of distributive and procedural justice. Journal of Personality and Social Psychology, 67, 850−863. Usunier, J. C. (1996). Marketing across culture. London: Prentice Hall. van Birgelen, M., de Ruyter, J. C., de Jong, A., & Wetzels, M. G. M. (2002). Customer evaluations of after-sales service contact modes: An empirical analysis of national culture's consequences. International Journal of Research in Marketing, 19, 43−64. White, M. M., Tansky, J. A., & Baik, K. (1995). Linking culture and perceptions of justice: A comparison of students in Virginia and South Korea. Psychological Reports, 77(3, Pt. 2), 1103−1112. Winsted, K. F. (1997). The service experience in two cultures: A behavioral perspective. Journal of Retailing, 73(3), 337−360. Winsted, K. F. (1999, Spring). Evaluating service encounters: A cross-cultural and cross-industry exploration. Journal of Marketing Theory and Practice, 7, 106−123. Yoo, B., & Donthu, N. (2002). The effect of marketing education and individual cultural values on marketing ethics of students. Journal of Marketing Education, 24(2), 92−103. Yoo, B., Donthu, N., & Lenartowicz, T. (2001). Measuring cultural values: Development and validation of CVSCALE. Unpublished manuscript. Zeithaml, V. A. (1981). How consumers' evaluation processes differ between goods and services. In J. H. Donnelly, & W. R. George (Eds.), Marketing of services (pp. 186−190). Chicago: American Marketing Association.