Drivers of Customers ’ Reactions to Service Failures : The Israeli Experience

This paper develops and tests a model of cultural, personality, and demographic drivers of three customer reactions to service failures, namely voice, switching, and negative word-of-mouth (N-WOM). Data from Israeli customers mostly support the model. Some drivers (uncertainty avoidance, self-efficacy and Machiavellianism) enhance the probability of all three reactions whereas others increase the probability of only one (e.g., perceived-control switching; self-confidenceNWOM). The theoretical and managerial implications of these findings are discussed and directions for future research are provided. The theoretical and managerial implications of these findings are discussed and directions for future research that integrates the Big Five personality traits and Hofsete’s five cultural dimensions are provided.


Introduction
Post-purchase behavior (product ownership, usage, and disposal) has been discussed widely in the marketing literature (Allison, 1978;Anderson, 1984;Bearden & Mason, 1984;Hernandez & Fugate, 2004;Nyer, 2000;Singh, 1988Singh, , 1989Singh, , 1990;;Westbrook, 1987).Such behaviors might lead to cognitive and affective responses including positive or negative satisfaction appraisal, seller-directed reactions, and word-of-mouth transmissions (Day, 1984).Customers' post-purchase behaviors are crucial for several reasons.First, customer satisfaction ratings are good indicators of firms' future profits (Anderson & Sullivan, 1993).Second, firms increasingly use customer satisfaction as a criterion for diagnosing product performance (Anderson & Sullivan, 1993).Third, understanding and responding more effectively to customers' dissatisfaction and complaints can improve firms' ability to attract new customers and sell more to current customers (Fornell, 1992).
Notably, attracting new customers is expensive whereas the lifetime value of future cash flows generated by repeat customers is high (Liu & McClure, 2001).Such a tradeoff suggests that firms can save resources by adopting a strategy of handling unsatisfied customers instead of attracting new customers.Yet, while the number of litigation cases and complaints to customer associations has increased, especially for service-related problems, many dissatisfied customers are either unable to or do not seek redress directly from sellers (Singh, 1989).Such tendencies do not allow marketers to capture information from unsatisfied customers, which is critical in responding to dissatisfaction.In addition, defensive marketing (e.g., complaint management) could reduce the cost of offensive marketing (e.g., advertising), thus lowering total marketing expenditures (Fornell & Westbrook, 1979).Hirschman (1970) suggested three customer dissatisfaction responses.Exit refers to customers' termination of the exchange relationship by switching to another supplier.Voice refers to customers' possible attempt to change their dissatisfied by talking to the supplier, which ultimately diminishes negative word-of-mouth (N-WOM), reduces dissonance, and increases satisfaction and loyalty (Chelminski & Robin, 2007).Finally, loyalty refers to customers, who refrain from reacting to dissatisfaction and repurchase the firm's product.
The literature on complaining behavior has examined drivers such as personality traits, attitudes toward complaining, and behavioral propensity to complain (Bearden & Mason, 1984;Richins, 1980).For example, Bodey and Grace (2007) examined the impacts of personality traits (self-efficacy, Machiavellianism, perceived control, and risk-taking) on customers' attitude toward and propensity to complain.Attitude toward complaining is defined as "the overall affect of goodness or badness of complaining to sellers" (Singh & Wilkes, 1996, p. 353).Propensity to complain is defined as "an individual's likelihood of seeking redress or expressing dissatisfaction to a service provider when he or she had an unsatisfactory service encounter" (Bodey & Grace 2007, p. 580).
While the post-purchase literature is vast, several gaps remain.First, the relationships between unsatisfied customers' post-purchase behaviors (voice, switching, and N-WOM) and personality traits (e.g., self-confidence and perceived control) have been under-researched.Thus, this study contributes to the literature by developing and testing an integrative model of individual-level concepts and the three dissatisfaction outcomes.
Second, culture might impact complaint behavior (Hernandez & Fugate, 2004) and influence patterns of responses to post-purchase dissatisfaction because cultural norms affect individuals' values, self-concepts, and perceptions of others (Liu & McClure, 2001).Since the literature has not addressed the impact of culture on dissatisfaction outcomes, this study adds to the literature by including uncertainty-avoidance as a possible driver of such outcomes.
Finally, most research on post-purchase behavior has been conducted in the US (for exceptions, see Bodey & Grace, 2006;Lam, Lee, & Mizerski, 2009).The current study contributes by replicating some previous findings in Israel, which is more uncertainty avoiding than the US (Hofstede, 2001).To the extent that similar findings emerge in Israel, confidence in the generalizability of previous findings would increase.
The remainder of the paper is organized as follows.First, a literature review is provided, leading to the research hypotheses and integrative model.Then, it describes the study designed to test them and report its findings.A discussion section with research and practical implications concludes the paper.

Literature Review and Research Hypotheses
Propensity to voice refers to "an individual difference variable that assesses a customer's inclination to engage in complaining behavior directly to a firm representative" (Chelminski & Robin, 2007, p. 96).Switching refers to individuals' tendency to abandon a supplier because of low quality, unfair price, or anger episode (Wetzer, Zeelenberg, & Pieters, 2007).Negative word-of-mouth (N-WOM) refers to informal negative information transmitted through person-to-person communication regarding a brand, product or service (Wetzer, Zeelenberg & Pieters, 2007).Below, we discuss the drivers of these three outcomes.
Self-Efficacy refers to "beliefs in one's capabilities to mobilize the motivation, cognitive resources, and courses of action needed to meet given situational demands" (Wood & Bandura 1989, p. 408).Customers with high self-efficacy tend to believe in their ability to adopt specific behavior and thinking in order to achieve their goals (Gibson, 2001).
Applied to our context, high-self-efficacy customers would believe in their ability to react effectively to negative consumption experiences.In other words, their propensity to complain, to switching behavior, and to N-WOM should increase because they will believe that such reactions might lead to positive outcomes (Bodey & Grace, 2007;McKee, Simmers, & Licata, 2006).Accordingly, Santos and Fernandes (2008) found positive relationship between self-efficacy and N-WOM and self-efficacy discriminated between complainers and non-complainers (Bodey & Grace, 2006).Finally, Luthans and Peterson (2002) argued that the higher the person's self-efficacy, the more persistent he or she is when faced with failure.They noted that as a result it would be expected that such individuals would view complaining as a way in which to overcome their problems and achieve their goals.

Thus:
H 1a : There is a positive relationship between self-efficacy and voice.
H 1b : There is a positive relationship between self-efficacy and switching.
H 1c : There is a positive relationship between self-efficacy and N-WOM.
Perceived Control refers to the "expectation of having the power to participate in making decisions in order to obtain desirable consequences and a sense of personal competence in a given situation" (Rodin, 1990, p .4).Studies on interpersonal communication have established that goals for WOM include revenge, warning, straightening social relationships, and advice (Alicke, 1992;Henning, Gwinner, Walsh & Gremler, 2004;Luminet, Bouts, Delie, Manstead, & Rime, 2000;Sundaram & Webster, 1998).Notably, people with high-perceived control are typically motivated to warn and advise, including in situations of dissatisfaction with companies' products and services.Accordingly, the propensity to complain depends on perceived control, among other personal traits (Bodey & Grace, 2007).Skinner (1996) distinguished between high-and low-control individuals.Individuals perceiving high levels of control over their environment tend to exercise greater effort to achieve goals by taking action and displaying attributes of competence.As a result, "individuals who maintain a high level of perceived control would be more likely to complain than those who do not" (Bodey & Grace 2006, p. 179).Thus: H 2a : There is a positive relationship between perceived control and voice.
H 2b : There is a positive relationship between perceived control and switching.
H 2c : There is a positive relationship between perceived control and N-WOM.
Self-confidence refers to the "extent to which an individual feels capable and assured with respect to his or her marketplace decisions and behaviors" (Bearden, Hardesty, & Randle, 2001, p. 122).Such confidence might affect several post-purchase behaviors.Low self-confidence people might feel resigned and believe that events are beyond their control; hence, lacking belief in achieving desired outcomes, they might refrain from acting, including via complaints (Chelminski & Robin, 2007).However, high self-confidence people tend to demonstrate goal-directed behavior (Tafarodi & Swann, 1996) and should be more likely to voice.Thus, people with high levels of self-confidence tended to demonstrate high levels of complaining, N-WOM, and switching behaviors compared to people with low levels of self-confidence (Gronhaug & Zaltman, 1981).Similarly, Bearden and Teel (1980) argued that customer self-confidence might predict customers' post-purchase voice behavior.Thus: H 3a : There is a positive relationship between self-confidence and voice.
H 3b : There is a positive relationship between self-confidence and switching.
H 3c : There is a positive relationship between self-confidence and N-WOM.
Machiavellianism refers to a "negative epithet, indicating at least an amoral way of manipulating others to accomplish one's objectives" (Hunt & Chonoko, 1984, p. 24).Accordingly, it might lead people to voice, switch, and complain to others.In line with this expectation, Bodey and Grace (2007) found a positive relationship between Machiavellianism and the propensity to complain.Additionally, in the event of service failure, high-Machiavellianism individuals might use power of persuasion to rectify the situation aggressively in order to win.As a result, they may view complaining as a mechanism to be utilized in order to get what they want (Bodey & Grace, 2006;Corzine & Buntzmanl, 1999).Moreover, people with strong propensity to complain also tend to switching behaviors compared to those with weak propensity to complain (Bolton & Bronkhurst, 1995).Thus: H 4a : There is a positive relationship between Machiavellianism and voice.
H 4b : There is a positive relationship between Machiavellianism and switching.
H 4c : There is a positive relationship between Machiavellianism and N-WOM.
Negative Word-of-Mouth and Switching.N-WOM refers to informal negative information about products or services, which is transmitted through person-to-person communication (Wetzer, Zeelenberg, & Pieters, 2007).Switching refers to individuals' tendency to abandon suppliers due to poor quality, low commitment, unfair prices, or anger episodes (Wetzer, Zeelenberg, & Pieters, 2007).N-WOM and switching should be related positively to avoid cognitive dissonance.People electing to switch supplier should demonstrate high levels of N-WOM consistent with their decision to change previous choices.Chelminski and Robin (2007) provided support for these consistency-based arguments vis-à-vis dissonance reduction.Thus: H 5 : There is a positive relationship between Switching and N-WOM.
Uncertainty -Avoidance.Hofstede (1991) identified four cultural dimensions (masculinity/ femininity, individualism/ collectivism, power distance, and uncertainty avoidance).Of the four, uncertainty avoidance pertains to the topic of this paper.According to him, uncertainty-avoidance refers to the "extent of feeling threatened by uncertain or unknown situations" (p.113).People from uncertainty-avoidance cultures, who experienced a problem, tended to exhibit lower intention to switch, engage in N-WOM, or complain, hence voice (Liu & McClure, 2001).Likewise, Hofstede (2001) found that the uncertainty-avoidance dimension was correlated positively with brand loyalty and non-switching policy.Thus: H 6a : There is a negative relationship between uncertainty-avoidance and voice.
H 6b : There is a negative relationship between uncertainty-avoidance and switching.
H 6c : There is a negative relationship between uncertainty-avoidance and N-WOM.
Age. Notably, all three outcomes entail risks.For example, voicing (complaining) could meet with a negative or unfriendly response from service personnel.Switching into a new and untested alternative might not lead to higher satisfaction if this alternative also fails to meet expectations.Finally, N-WOM might entail negative reactions from satisfied customers to whom N-WOM is communicated.However, switching is the riskiest because the probability that service personnel will be unfriendly is rather low since their job is to handle complaints.Likewise, negative reactions should not be expected from one's friends.Age might affect risk-taking positively or negatively.Arguably, young people tend to take higher risks than older people do because they have less to lose.Alternatively, older people, mired in routine day-to-day schedules, might seek risks to alleviate boredom.Similar arguments were made in the context of risky sports (Shoham, Rose, & Kahle, 1998, 2000).Thus, this study includes age as a control variable in the analysis for switching behavior.
The model depicted in Figure 1 summarizes the research hypotheses.This model guided the empirical study discussed below.
Insert Figure 1 Here

Sample
Data were collected in Israeli trains.Students, instructed thoroughly in research methodology, distributed 528 questionnaires to passengers.However, 89 questionnaires were disqualified for various reasons (20 left the train before completing the questionnaires; 31 said that they do not participate in surveys; 33 were too busy; and five provided no explanation for not participating).Thus, the final sample size included 439 individuals (83% response rate).Given this high response rate, non-response does not appear to be a problem in this research.
The students were instructed to aim for a representative sample of the adult Israeli population (≥ 18).They achieved a gender-balanced sample (221 males; 218 females).As for religion, 402 were Jewish (92.2%), 34 were Arabs (7.8%), and three were not classified.Israel's adult population (defined as ages 14+) includes 78.1% Jewish, 17.6% Arabs, and 4.3% without religion classification (Central Bureau of Statistics, 2008).These differences will be discussed in the "Limitations" section.Table 1 provides descriptive statistics for the sample.
Insert Table 1 Here

Measures
Data were collected through a structured, self-report questionnaire.Other than demographics, the questionnaire included twelve scales.Since all scales were originally in English, back-translation was used.One bilingual individual translated the original to Hebrew.Another bilingual individual, blind to the original questionnaire, back-translated the Hebrew version to English.Then, the two individuals and one of the authors evaluated the translations for wording, content, and local applicability and equivalence.Minor disagreements were resolved in this stage leading to the final version.
Propensity to voice is "an individual difference variable that assesses a customer's inclination to engage in complaining behavior directly to a firm representative" (Chelminski & Robin, 2007, p. 96).The original scale was developed by Bearden, Hardesty and Randle (2001) and included five items with a reported α reliability of .88.In the current study, seven items were used -the four items that worked well for Chelminski and Coulter (2007) and three items from Singh (1988).The reliability in this study was acceptable (α = .80).
Switching refers to the tendency of an individual to abandon the firm or the supplier mostly because of poor quality, low commitment, unfair price, or anger episode (Wetzer, Zeelenberg & Pieters, 2007).Ping's (1995) original scale included three items with a reported α reliability of .91.Striving to avoid using only three items, which might or might not work in a first use in Israel, qualitative interviews were used to develop five additional items patterned after the original three were developed.In the current study, the 8-item scale was reliable (α = .90).
N-WOM refers to informal negative information which transmission by person-to-person communication regarding a brand, product or service (Wetzer, Zeelenberg, & Pieters, 2007).The original scale was developed by Price et.al (1995) with a reported α reliability of .95.The reliability in the current study was acceptable (α = .87).
Self-Confidence refers to the "extent to which an individual feels capable and assured with respect to his or her marketplace decisions and behaviors" (Bearden, Hardesty & Randle, 2001, p. 122).The original scale was developed by Tafarodi and Swann (1996) and included seven items with a reported α reliability of .84.The reliability in the current study was acceptable (α = .90).
Self-Efficacy refers to "beliefs in one's capabilities to mobilize the motivation, cognitive resources, and courses of action needed to meet given situational demands" (Wood & Bandura., 1989, p. 408).The original 8-item scale was developed by Schwarzer and Jerusalem (1995;α = .84).The reliability in the current study was acceptable (α = .87).
Machiavellianism refers to a "negative epithet, indicating at least an amoral way of manipulating others to accomplish one's objectives" (Hunt & Chonoko, 1984, p. 24).The original 7-item scale was developed by Christie and Dacin (1970) and had an α reliability of .73.The reliability in the current study was acceptable (α = .79).
Perceived-Control refers to the "expectation of having the power to participate in making decisions in order to obtain desirable consequences and a sense of personal competence in a given situation" (Rodin, 1990, p. 4).The original 5-item scale was developed by James (1957;α = .79).The reliability in the current study was acceptable (α = .81).

Findings
Table 2 presents the means, standard deviations, and inter-scale correlations.Table 3 presents the results of the regression analyses (standardized coefficients).

A Test of Switching as Mediator
A four-step approach tested for mediation (Baron and Kenny, 1986).First, most independent variables (self-efficacy, p= .07;perceived control, p = .07;Machiavellianism, p =.06; and uncertainty-avoidance, p = .00)affected switching (the mediator), except for self-confidence.Second, switching had a significant effect on N-WOM (p = .00).Third, the independent variables (self-efficacy, p =.07; self-confidence, p = .07;Machiavellianism, p = .08;uncertainty-avoidance, p = .00)affected N-WOM in the absence of switching (the mediator), except for perceived control.Finally, the effects of several independent variables (self-efficacy, p = .16;Machiavellianism, p = .19;uncertainty-avoidance, p = 0.04) on N-WOM (the dependent variable) were reduced with the addition of the mediator, switching.Hence, switching acts as a mediator in the model.
Insert Table 4 Here

Theoretical Implications
Briefly reiterating the major findings, this study examined and documented the relationships between personality, cultural, and demographic and customers' dissatisfaction outcomes (voice, switching, and N-WOM).In general, personality factors, especially self-efficacy and Machiavellianism were related with customers' tendency to complain (voice), switch, or engage in N-WOM.The cultural driver (uncertainty-avoidance) and age affected all three dissatisfaction outcomes, with the former being opposite expectations.
At the personality level, self-efficacy and Machiavellianism were the strongest predictors of customers' likelihood to engage in the three dissatisfaction behaviors.Although some of these relationships (i.e., self-efficacy; Machiavellianism  switching; N-WOM) were weaker than others, they were all in the predicted direction.These results are in line with studies that showed that personality variables such as assertiveness, self confidence, and self-control to increased the probability of complaining (Fornell & Westbrook, 1979;Keng, Richmond & Han, 1995;Richins, 1987).Likewise, the role of self-efficacy as a voice predictor paralleled the finding that powerlessness was related to the likelihood of customers' voice behavior (Robertson & Shaw, 2009).As such, customers who perceive themselves as potent in achieving their desired outcomes will be more willing to engage in complaint behavior.Finding that Machiavellianism predicted voice, and, to a lesser extent, switching and N-WOM is important as it was not related with complaining tendencies in previous research (Bodey & Grace, 2006).
In addition, for the most part, non-significant results were found with respect to customers' perceived control and self-confidence and the three dissatisfaction behaviors.These findings contradict previous studies that found "complainers" to perceive high levels of control (Bodey & Grace, 2006).At the same time there was a marginally significant relationship between perceived control and switching, implying that for Israeli customers, a sense of control over the environment is achieved by a concrete act of change (switching) rather than communicating dissatisfaction that might or might not lead to the desired outcome.Again, the differences in the national setting and the cultural characteristics of customers across the different studies might explain the contradicting results.
The cultural dimension uncertainty-avoidance predicted the three dissatisfaction behaviors contrary to expectations.Uncertainty avoidance was positively associated with switching and N-WOM and, marginally so, with voice.These results indicate that customers perceive the situation that causes their dissatisfaction as a source of ambiguity that needs to be reduced.As such, complaining, switching and disseminating N-WOM can be used to reduce this state of uncertainty and doubt.The findings are consistent with the observation that high-uncertainty-avoidance customers are more active in seeking remedies to undesired situations.
Finally, in this study switching behavior played a mediating role between the personality and cultural predictors and N-WOM, so that customers tended to exhibit more N-WOM following switching their patronage.This negative communication can be viewed as customers' attempt to justify their action and reduce the risk or uncertainty involved in the change.This relationship is in line with dissonance theories, which suggest that after making a buying decision, customers experience post-purchase dissonance and express their concerns of having made the wrong choice (Festinger, 1957).In this context, dissonance which customers strive to reduce arises from the switching behavior.WOM has been shown to be one of the strategies customers take to reduce post-purchase dissonance (Richins & Bloch, 1986).As for post-switching N-WOM, it has been shown that dissatisfied switchers engage in such behavior when switching (dissatisfaction) followed negative experiences and this negative action was taken to reduce cognitive dissonance (Wangenheim, 2005).
At the demographic level, age predicted the three dissatisfaction behaviors, voice, switching and N-WOM, such that older customers were more likely to engage in all three behaviors.These findings go beyond the results reported by Solvang (2008).Solvang (2008) found age to be positively associated with the propensity to voice but not with N-WOM or switching.However, this relationship warrants further examination as most studies have found customer complaint behavior to be inversely related to age (Bearden & Mason, 1984;Day & Landon, 1977;Grønhaug & Zaltman, 1981;Singh, 1990).

Managerial Implications
Customers' reactions to dissatisfaction require special attention from firms as it could be a source of irritation or time-consuming interactions.Moreover, it can potentially cause considerable harm to firms' reputation.Especially today, online customer forums and social media provide an effective tool for instantaneous communication of dissatisfaction that can reach huge audiences with only a mouse-click.Social media can provide a credible and reliable forum for N-WOM, which can hurt companies' reputation.At the same time, providing a supportive environment for customers who seek redress can grant companies with the loyalty of satisfied customers that might disseminate positive WOM.In short, managing dissatisfaction is crucial in today's environment.
This study viewed customers' dissatisfaction outcomes -voice, switching and N-WOM -as separate processes with customer characteristics predicting each in a different way.Implications for managers are discussed below for each of the three.Self-efficacy, Machiavellianism, and age were strong predictors of voice propensity.Companies should encourage customers' expressions of dissatisfaction and provide a supportive environment for redress.Furthermore, providing customers with constant feedback or updates on the status of their complaint and the company's measures to address could be used to give complainers a sense of confidence in their ability to react effectively and obtain results in a given situation.In addition, as voice was related to customers' self-efficacy, companies should communicate the likelihood of success when complaining to customers.This should lead to customers feeling empowered and able to react effectively.Encouraging younger customers to communicate their dissatisfaction can be done by simplifying the grievance handling process and offering a variety of channels to file complains.Effective tactics include online complain mechanisms both appeal to young audiences and can be done at any time from any place they choose especially in the era of converging communication technologies.On the other hand, as Machiavellianism was related to the propensity to complain, companies should create mechanisms to discourage unjustified complains by customers that use voice as a manipulation to achieve preferred outcomes.Such mechanisms include maintaining a database that facilitates the identification of illegitimate complaints (Reynolds & Harris, 2005) and preparing a set of standardized responses for such complaints.
This study implies that all three reactions to dissatisfaction are customers' attempts to cope and reduce the uncertainty inherent in the situation characterized by a discrepancy between customers' expectation and reality.As such, customers that cannot tolerate the uncertainty that accompanies the dissatisfaction tend to voice, switch and disseminate N-WOM.Therefore, marketers should eliminate sources of uncertainties, ensuring customers that any dissatisfaction will be handled both promptly and effectively.This should be conveyed to customers in advance or concurrently with the act of purchase, so sources of uncertainties are eliminated, ideally, prior to encountering a problem.
Finally, responsiveness to customers' dissatisfaction not only enhances the positive image and reputation of a company, but could also prevent N-WOM.This study showed that, to some extent, N-WOM is the "last resort" customers take and occurs following brand switching.Whether N-WOM is explained by customers' attempts to reduce dissonance or by a genuine attempt to warn other customers, this form of communication very detrimental to businesses reputation as N-WOM is more influential than positive WOM (Bone, 1995;Mizerski, 1982).Therefore, dissatisfaction should be contained early to prevent customers from switching and engaging in N-WOM.Companies should emphasize how worthwhile it is to complain to them directly (Lau & Ng, 2001).In addition, companies should identify customers who are more likely to use N-WOM (i.e., those with high self-efficacy, self confidence, and Machiavellianism), especially in product categories that are more susceptible to WOM such as high-social-visibility products (Richins, 1987).Special relationship marketing programs and policies should be developed for such customers, including training of customer service personnel to effectively address early signs of dissatisfaction.

Limitations and Further Research
This study is subject to a number of limitations.First, data for the study were collected on Israeli trains.While the sample was balanced gender-wise, it deviated from national averages for several variables.Regarding religion, the sample included 92.2% Jewish respondents (78.1% in Israel adult population) and 7.8% Arab respondents (17.6% in Israel adult population).Regarding marital status, 5.5% of the respondents were divorced, widowed, or separate (versus 14% divorced in the Jewish population.Finally, regarding income, the sample included mostly average-or lower-than-average income respondents (65%).This is probably a result of collecting information in a public transportation.While the deviating demographics should not affect the findings, additional studies with representative samples would enhance the generalizability of the findings reported here.
In this paper we tested a model that included several predictor variables for three customer dissatisfaction outcomes.Predictor variables included four individual traits (i.e., self-efficacy, perceived control, self-confidence and Machiavellianism) and one cultural characteristic (i.e., uncertainty-avoidance).This study provides a launching pad for further research that will develop a comprehensive model that incorporates more personality traits as predictors of consumers' reaction to service failure.Such predictors might include the five domains of personality (The Big Five): Openness, conscientiousness, extraversion, agreeableness, and neuroticism, using Hogan's (1986) Personality Inventory.Thus, one can postulate that extraversion will be positively associated with self-efficacy, perceived control and self-confidence as measured in our model that will predict N-WOM and voice.This is further supported by previous studies that found complainers to be aggressive and extrovert in nature compared to non-complainers who were low-esteem, low confidence and reserved in terms of personality (Bodey & Grace, 2006).In contrast, agreeableness that entails friendliness, tolerance, altruism and modesty (Costa & McCrae 1992), less aggressiveness, rudeness and thoughtless (Tosi et al., 2000) will be negatively related to Machiavellianism, and as a result to Voice and N-WOM.However, it is likely that this personality trait will be associated with switching behavior that does not require verbal confrontation and conflict.
Additionally, this research included Hofstede's uncertainty-avoidance dimension.First, it would be illuminating to examine the role of the other four Hofstede dimensions, operationalized at the individual level (individualism, masculinity, power distance, and long-term orientation).Secondly, would the findings reported here generalize to other cultures, which differ from Israel at the macro level?Hofstede (2001) suggested four cultural dimensions: power-distance, individualism, uncertainty-avoidance and masculinity.Regarding power-distance, Israel is low (PDI = 13) the US is medium (PDI = 40) and Philippines is high (PDI = 94).Regarding individualism, Israel is medium (IDV = 54), the US is high (IDV = 91), and Taiwan is low in (IDV = 17).Regarding uncertainty-avoidance, Israel is high (UAI = 81), Norway is medium (UAI = 50) and Singapore is low (UAI = 8).Finally, regarding masculinity, Norway is low (MAS = 8) Israel is medium (MAS = 47), and Japan is high (MAS = 95).
Power-Distance (PDI).Power-distance is defined as "the potential to determine or direct the behavior of another person or other persons more so than the other way round" (Hofstede, 2001, p. 82).Individuals in low-power distance societies are less likely to be obedient than ones in high power-distance societies.Hence, the propensity to complain should be stronger in less power-distance countries compared to high power-distance ones.
Individualism (IDV).Individualism is defined as "the degree to which people in a country prefer to act as individuals rather than as members of groups" (Hofstede, 1993, p. 89).According to Hofstede (2001, p. 215), Israel is less individualistic (IDV = 54) compared to the US (IDV = 91).People in high-IDV countries are self-orientated, live in guilt-based cultures, have to take care of themselves, and tend to make decisions based on individual needs whereas people in low-IDV countries tend to belong to shame-based cultures and demonstrate absolute loyalty to groups (Hofstede, 2011a).Since high-IDV countries create a guilt culture, customers in such countries should demonstrate high levels of complaining behavior compared to customers in low IDV.
Uncertainty-Avoidance (UAI).UAI is defined as "the degree to which people in a country prefer structured over unstructured situations" (Hofstede, 1993, p. 90).According to Hofstede (2001, p. 151), Israel is more uncertainty-avoiding (UAI = 81) than the US (UAI = 46).People in high-UAI countries tend to conservative behavior whereas people in low-UAI ones demonstrate openness to change and innovation (Hofstede, 2011b).Thus, the propensity to complain should be greater in high-UAI countries compared to low-UAI countries.
Masculinity (MAS).Hofstede (2001, p. 297) argued that "masculinity stands for a society in which social gender roles are clearly distinct: Men are supposed to be assertive, tough, and focused on material success; women are supposed to be more modest, tender, and concerned with the quality of life.Femininity stands for a society in which social gender roles overlap; Both men and women are supposed to be more modest, tender, and concerned with the quality of life".In addition, people in high-masculinity cultures show more confidence in advertising than ones from low-masculinity cultures (Hofstede, 2001).Hence, people in high-masculinity countries should demonstrate more trust in firms and lower propensity to complain compared to individuals from low-masculinity countries.
In sum, several fruitful directions were proposed above.Some arose from addressing the limitations of this study.Others followed from a discussion of potential generalizability of the findings.Hopefully, these directions will suggest a map and compass for further research on this important topic.Baron & Kenny, 1986) 1. Testing the significance between the independent variables and the mediator, switching.
2. Testing the significance between the mediator, switching and the dependent variable, N-WOM.
3. Testing the significance between the independent variables and N-WOM, in the absence of the mediator (switching).
4. Testing the significance between the independent variables and N-WOM, including the mediator, switching.Some of the relationships between the independent variables and the dependent variable were reduced upon the addition of the mediator, Hence, according to Baron & Kenny (1986), the variable switching, acts as a mediator in our model.

Table 2 .
Scale means, standard deviations, reliability levels (α), and correlations * All correlations are significant at the p≤0.05 level ** All correlations are significant at the p≤0.01 level

Table 3 .
Regression models for predicting voice, switching and N-WOM a

Table 4 .
Testing switching as a mediator (based on