Journal of Retailing 85 (3, 2009) 321–335

     Dysfunctional Customer Behavior Severity: An Empirical Examination
322                                    K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335

    In assu...
K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335                             323

the term “psycholo...
324                                      K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335

K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335                              325

Influence of psych...
326                                    K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335

physical de...
K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335                             327

Swan’s (1989) four...
328                                             K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335

K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335                                           329

330                                           K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335

K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335                             331

through the design...
332                                    K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335

significant ...
K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335                                          333

2. (A...
Dysfunctional Customer Behaviour
Dysfunctional Customer Behaviour
Dysfunctional Customer Behaviour
Upcoming SlideShare
Loading in...5

Dysfunctional Customer Behaviour


Published on

Published in: Technology, Business
1 Comment
1 Like
No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Transcript of "Dysfunctional Customer Behaviour"

  1. 1. Journal of Retailing 85 (3, 2009) 321–335 Dysfunctional Customer Behavior Severity: An Empirical Examination Kate L. Reynolds a,∗ , Lloyd C. Harris b a Cardiff Business School, Cardiff University, United Kingdom b Warwick Business School, Coventry, United Kingdom Abstract Although many studies assume that customers monotonically act in both a functional and a good-mannered way during exchange, considerable anecdotal evidence suggests that customers routinely behave negatively and often disrupt otherwise functional encounters. However, to date, rigorous empirical evidence of this phenomenon is lacking. This study synthesizes extant literature from a broad range of areas and advances two alternative conceptions of the factors associated with dysfunctional customer behavior severity. That is, after controlling for a variety of factors, the authors suggest that psychological obstructionism, disaffection with service, and servicescape variables are significantly associated with the severity of deliberate dysfunctional customer acts. The results provide insights for researchers interested in the darker side of service dynamics and generate useful implications for services practitioners charged with reducing the severity and the frequency of episodes of deviant customer behavior. © 2009 New York University. Published by Elsevier Inc. All rights reserved. Keywords: Dysfunctional customer behaviour; Customer misbehaviour; Consumer deviance; Antecedent; Structural equation modelling The majority of research into customer–firm interactions several countries, including the Czech Republic, Japan, Iceland, is founded on the assumption that customers act in both and New Zealand. This leads Reynolds and Harris (2006) to a functional and a good-mannered way (e.g., Ringberg, argue that customer misbehavior is endemic within the service Odekerken-Schröder, and Christensen 2007). This contrasts industry. Fullerton and Punj (2004) suggest that norm-violating with practitioner-oriented research that repetitively alludes to behavior is pervasive and representative of everyday customer customers behaving badly (e.g., Dube 2003) and with intermit- behavior, rather than constituting a segregate faction of society tent scholarly studies that typically highlight the prevalence of a (see also Harris and Reynolds 2004). single form of customer misbehavior, such as shoplifting (Kallis The disparate focus on individual forms of dysfunctional cus- and Vanier 1985) and illegitimate complaining (Reynolds and tomer behavior is detrimental to a broader understanding of these Harris 2005). issues. Indeed, Fullerton and Punj (1993) argue that there is a Despite limited academic attention, the pervasiveness of cus- need to elucidate the range of antecedents and to provide empiri- tomer dysfunction appears global. Focusing on one individual cal insights into their dynamics. Thus, research that examines the form of customer misbehavior, Grandey, Dickter, and Sin (2004) factors associated with dysfunctional customer behavior is lit- reveal that, on average, service employees within the United tered with calls for future studies to examine its antecedents more States fall victim to episodes of customer aggression ten times thoroughly (e.g., Al-Rafee and Cronan 2006). Fullerton and Punj a day. These findings are comparable to that of a study con- (1993, 2004) stress the need for “better” data that captures a ducted in the United Kingdom (USDAW 2004), which reveals more inclusive investigative approach. Echoing this, Harris and that front-of-store assistants are subjected to verbal abuse once Reynolds (2003) call for research to examine the antecedents of every 3.75 days, to threatening behavior every 15 days, and to dysfunctional customer behavior more holistically. acts of violence every 31 days. Moreover, Bamfield (2006) pro- The current research responds to these calls by investigat- vides evidence of the ominous rise in thefts by consumers across ing the factors associated with dysfunctional customer behavior severity, that is the extent to which a customer deliberately behaves in a way that violates the norms and unwritten rules of ∗ Corresponding author. an individual service setting in a negative fashion. Our aim is to E-mail addresses: (K.L. Reynolds), amalgamate previous insights and to examine empirically actual (L.C. Harris). (as opposed to speculative) incidents of customer misbehavior. 0022-4359/$ – see front matter © 2009 New York University. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.jretai.2009.05.005
  2. 2. 322 K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 In assuming a norm-breaking perspective, we draw on lit- implications of customer deviance, research in this area is in its erature from diverse areas, including: sociology, psychology, infancy and tends to be exploratory. criminology, ethics, environmental psychology, marketing, and The third theme of research, and indeed, the focus of the employee deviance that offer insight into the constructs corre- current paper, pertains to the drivers of dysfunctional customer lated with dysfunctional behavior. Synthesis of these literatures behavior. Typically, existing studies are framed within a specific leads to the forwarding of three main constructs that associate academy, including sociology (Rosenbaum and Kuntze 2003), with customer misbehavior severity: psychological obstruction- psychology (Al-Rafee and Cronan 2006), and business ethics ism, disaffection with service, and servicescape variables. By (Fukukawa 2002). Here, the main research focus is on exploring integrating insights from wide-ranging literature streams and the antecedents of individual forms of misbehavior. In particular, drawing on multiple theoretical bases, we develop and test our shoplifting and consumer fraud have received prolific attention research model and a rival model. Specifically, our research over the past four decades (Harris 2008). However, although the model is founded upon the propositions of Bitner (1992) and majority of research into the antecedents of dysfunctional cus- Fullerton and Punj (1993). By contrast, our rival model is tomer behavior has centered on shoplifting, sporadic insights inspired by literature that approaches customer dysfunction in into other forms of customer misbehavior include consumer a linear and direct fashion (e.g., Phillips, Alexander, and Shaw resistance, vandalism, illegitimate complaining, and rage (see 2005). Further details of which are explored in the later discus- Grove, Fisk, and John 2004; Reynolds and Harris 2005). sion. The managerial relevance of this study is evident. Our model Factors associated with dysfunctional customer behavior provides insights into customer dysfunction that may help man- severity agers reduce such behaviors. That is, we reveal that managers might manipulate many of the factors that associate with cus- Before we model the antecedents of dysfunctional customer tomer misbehavior. Our study is also of interest to marketing behavior, the construct of dysfunctional customer behavior theorists. By drawing on diverse research areas, this paper makes severity requires further elaboration. In an attempt to assimi- a conceptual contribution in deepening the understanding of the late norm breaking into a conceptual structure and an actionable factors that relate to customer misbehavior severity. Further- dependent variable, several studies support the validity of more, this study makes a methodological contribution through researching the perceived severity of dysfunctional behavior (see the development and validation of multi-item scales. Finally, Lawrence and Robinson 2007; Vitell and Muncy 1992). Harris this study contributes to the theory of customer deviance by and Reynolds (2004) advocate the study of people who know- operationalizing and empirically examining conceptual frame- ingly break behavioral norms. Given these arguments, the focus works that depict the constructs associated with dysfunctional of our study centers on the severity of dysfunctional customer customer behavior holistically. behavior in terms of the extent to which a customer deliberately behaves in a way that violates the norms and unwritten rules of Literature review an individual service setting in a negative fashion. The disparate nature of existing research which offers insight Evidence pertaining to the prevalence of dysfunctional cus- into the factors associated with dysfunctional customer behavior tomer behavior has drawn the attention of a small but growing results in a multitude of possible constructs of interest, thus rep- number of academies that have supplied insights into this phe- resenting a challenge to the researchers to identify those most nomenon. Such studies often adopt the label “dysfunctional suitable for study within the consumer and services context. customer behavior,” which refers to behavior by consumers Consequently, a number of iterative processes were employed within the exchange setting that deliberately violates the gen- by the authors to identify the most relevant constructs. These erally accepted norms of conduct in such situations. We use the stages include reviewing the literature to ascertain the breadth term “dysfunctional customer behavior” because of its emphasis and depth of study of each construct, and the contextual and on the issues of intent and norm infringement. methodological applicability of each construct. For example, We divide the studies that generate insight into dysfunctional the role of environmental (servicescape) variables in episodes customer behavior into three themes. The first theme focuses of misbehavior is discussed across a wide range of disciplines on profiling the different forms of customer misbehavior. Pos- including: ethics, environmental psychology, criminology, soci- sibly the best-known categorization is from the anecdotal work ology, and strategic marketing. The cumulative outcome of this of Lovelock (1994), who identifies six service-based jaycus- process revealed three reflective factors: psychological obstruc- tomers. Contrasting typologies are also offered by Fullerton and tionism, disaffection with service, and servicescape, as those Punj (2004), Harris and Reynolds (2004), and most recently, most worthy of further examination. Berry and Seiders (2008). However, although these classifica- tions offer notable insights into the diverse varieties of customer Psychological obstructionism misbehavior, such studies lack empirical support. The second theme of research focuses on the consequences of dysfunctional In terms of the first of our three main associative factors, customer behavior. Specifically, the effects of customer misbe- Fullerton and Punj (1993) draw a link between consumers’ haviors are wide-ranging affecting employees, firms, and fellow personality traits and predispositions, and dysfunctional cus- customers (Harris and Reynolds 2003). Yet, despite the grave tomer behavior. Within the context of our research, we utilize
  3. 3. K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 323 the term “psychological obstructionism” to denote the enduring comprising four reflective sub-dimensions: layout and design, personality traits and predispositions, which impede individual atmospheric environment, behavior of fellow customers, and consumers from behaving in a normative fashion, and shape exterior environment. As with disaffection with service, eval- and constrain their interpretations of service encounters. In this uations regarding servicescape variables occur during service. regard, psychological obstructionism is present before, during, Although prior research has attempted to examine the effect of and after an exchange. Specifically, the dimensions of Machi- servicescape variables on consumers’ behavior (Morin, Dubé, avellianism, sensation seeking, aggressiveness, and consumer and Chebat 2007), only recent advancements within marketing alienation are conceptualized here as reflective dimensions of theory note a link between servicescape design and customer psychological obstructionism. misbehavior (e.g., Areni 2003). Nevertheless, support for the Within the fields of psychology and criminology, theorists relationship between environmental design and misbehavior is pay considerable attention to the study of personality and indi- evident within the environmental psychology and criminology vidual misbehavior. An early study in criminology was that of literature streams (Hopkins 2002; Wilson and Kelling 1982). Eysenck (1964), who proposed a relationship between hered- Therefore, as mentioned previously, this synthesis leads us itary personality traits and criminal behaviors. Subsequently, to advance the three constructs of psychological obstruction- numerous studies have correlated aspects of psychological ism, disaffection with service, and servicescape variables as obstructionism and acts of deviance (Romero, Luengo, and those that are most consistently linked to dysfunctional behav- Sobral 2001; Sutherland and Shepherd 2002). Studies focusing ior. However, there is little consensus as to how these constructs on exchange settings have also recognized the role of personality should be modeled. Indeed, different research traditions infer in driving acts of dysfunctional behavior. Fox and Spector (1999) contrasting theoretical models. Thus, we devise two competing and Harris and Ogbonna (2006) both advance personality vari- models and detail each in turn. Because of the degree of overlap ables to correlate with acts of workplace sabotage by employees, between the hypothesized paths across the two competing mod- while Rallapalli et al. (1994) and Ross and Robertson (2003) els, we present a summary of the theoretical foundation for each separately forward personality as key in ethical decision-making model. We follow this with a detailed discussion of each of the processes. With reference to dysfunctional customer behavior, five hypotheses. several studies draw links between personality dimensions and individual forms of misbehavior, including shoplifting (Kallis Theoretical models and Vanier 1985), cheating behavior (Wirtz and Kum 2004), and digital piracy (Al-Rafee and Cronan 2006). Our favored model, the research model (see Fig. 1), is grounded in Bitner’s (1992) and Fullerton and Punj’s (1993) Disaffection with service frameworks and research traditions and centers on the indirect relationship between servicescape variables and dysfunctional Consumers’ appraisals of the exchange encounter are also customer behavior severity. Drawing on environmental psychol- associated with acts of deviant behavior. Mills (1981) draws ogy, Bitner (1992) suggests that servicescapes do not directly a direct link between consumer disaffection and dysfunctional affect consumer behavior, but rather that a person’s internal customer behavior. Thus, we define “disaffection with service” responses mediate the relationship between the servicescape as customers’ negative cognitive-emotive evaluations of a ser- and the behavior. Thus, the research model hypothesizes a path vice provision. Specifically, the dimensions of dissatisfaction between servicescape variables and disaffection with service and and inequity are conceptualized as reflective indicators of dis- a relationship between disaffection with service and the severity affection with service. Disaffection with service occurs during of dysfunctional customer behavior. This indirect relationship service exchange and therefore implicitly comprises perceptions is investigated extensively in studies of “functional” consumer of front-line service employees who are intrinsically linked with behavior within environmental psychology and servicescape- service provision and the exchange experience (Bitner 1992). related literature streams (Lin 2004; Williams and Dargel 2004) This contrasts with psychological obstructionism, which is both pre-existing and enduring. Indeed, customers’ negative inter- pretations and reflections of a given situation are prominent antecedents of acts of undesirable behavior (Lee and Allen 2002). Specifically, judgments of dissatisfaction and inequity are key motives for incidents of dysfunctional behavior (Diamond 1997; Fullerton and Punj 2004). Servicescape Research also highlights the role of the characteristics of the exchange setting in influencing deviant customer behav- ior. Individual’s interpretation of the tangible and intangible characteristics of the exchange setting, which we label “ser- vicescape” for reasons of parsimony, is conceptualized as Fig. 1. Research model.
  4. 4. 324 K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 this. Similarly, Bitner (1992) argues that servicescapes do not directly cause consumers to behave in a certain fashion, but rather, behaviors are mediated by an individuals’ interpretation of the servicescape and their overall evaluations of the store. More specifically, Baker and Cameron (1996) argue that it is the combination of servicescape elements (for example, loud music, repellent decorative colors, uncomfortable furniture, and long queues), that antecede feelings of customer disaffection. Providing complementary findings within the context of the hospitality industry, Schmidt and Sapsford (1995) forward evi- dence of an empirical relationship between negative perceptions of servicescape and customer disaffection. Moreover, in utiliz- Fig. 2. Rival model. ing critical incident technique, Bäckström and Johansson (2006) reveal a combination of servicescape constructs to give rise to and is suggested within the domain of customer dysfunction poor service experiences, in that negative interpretations of the (Fullerton and Punj 1993; Mills 1981). Furthermore, Bitner servicescape put consumers in a negative frame of mind, which (1992) acknowledges the role of personality traits in influencing increases the likelihood that the consumers will be disaffected a person’s evaluation of his or her physical surroundings, a view with the overall service encounter. Complementary evidence is that is upheld within the sphere of deviant behavior by Eysenck also forwarded by a number of studies, which find an asso- and Eysenck (1970) and Fullerton and Punj (1993). Further ciation between negative interpretations of servicescapes and echoing Fullerton and Punj’s (1993) model of aberrant cus- increased levels of customer displeasure and thus, disaffection tomer behavior, the research model hypothesizes a link between (e.g., d’Astous 2000). Thus: psychological obstructionism and the severity of dysfunctional customer behavior. In their conceptual model, Fullerton and Punj H1 . The greater the negative interpretation of an outlet’s ser- (1993) propose a relationship between personality traits and vicescape, the higher is the level of perceived disaffection with aberrant customer behavior. Support for this hypothesis with service. specific reference to deviant behavior also derives from studies of business ethics, criminology, and psychology (Gottfredson Influence of disaffection with service on the severity of and Hirschi 1990; Rayburn and Rayburn 1996). Given the dysfunctional customer behavior strength of conceptual support for the research model, we favor it above the rival model. Commentators often position dissatisfaction and inequity The rival model (see Fig. 2) is distinct in that it depicts only (disaffection) as key drivers of misbehavior (see Harris and direct linear relationships between each of the three constructs Reynolds 2004). In addition to research that examines the link of interest and the severity of dysfunctional customer behavior. between dissatisfaction and inequity, and incidents of dysfunc- This model originates from and thus reflects the disparate nature tional behavior in a separate fashion (see Lee and Allen 2002; of prior research into deviant behavior, much of which focuses on Mills 1981), prolific support for the relationship between dis- the association between a single antecedent and a specific form affection and deviance is offered within both employee and of misbehavior. Therefore, in addition to the direct paths with consumer contexts. For example, Diamond (1997) forwards severity, as presented in our preferred research model, this model empirical results that link employee job disaffection with acts of suggests that there is a direct relationship between servicescape employee sabotage. Robinson and Bennett (1997) who presents and severity. This is in direct contrast to traditions that support evidence of a relationship between worker disaffection and a mediated model (Bitner 1992; Mehrabian and Russell 1974). employee misbehavior, also echo this finding. Focusing on the Support for this model comes from multiple research disciplines. deviant activities of consumers, Lovelock (1994), and Wirtz For example, Kuo and Sullivan (2001) suggest that the relation- and Kum (2004) separately note the presence of customer ship between physical environments and aggressive behavior is disaffection in increasing the intensity of acts of belligerent strong. Similarly, Phillips, Alexander, and Shaw (2005) high- and cheating behaviors. Offering corresponding findings, Yi light a link between the physical design of servicescapes and and Gong (2008), provide empirical support for the associa- incidents of consumer theft. tion between customer disaffection and dysfunctional customer behavior within a student context. Focusing specifically on acts Influence of interpretations of the servicescape on of consumer retaliation, Huefner and Hunt (2000) offer fascinat- disaffection with service ing insight that depicts how consumers engage in misbehaviors of ranging severities, including theft, vandalism and physical The view that an organization’s servicescape is associated violence, as a means of expressing their disaffection with an with consumer disaffection is widely debated and supported individual organization or employee. Thus: within services literatures (see Bitner 1992). Barnes, King, and Breen (2004) who finds that perceptions of environmental fac- H2 . The higher the level of disaffection with service, the greater tors contribute to feelings of customer disaffection demonstrate is the severity of dysfunctional customer behavior.
  5. 5. K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 325 Influence of psychological obstructionism on the evaluation aspects of psychological obstructionism play an important role of the servicescape in driving acts of criminal behaviors of ranging severities. To detail, evidence to support this association is forwarded in Support for a link between psychological obstructionism and an individual and holistic manner. Regarding individual facets individual’s evaluations of servicescape can be sourced from of psychological obstructionism; Machiavellianism, aggressive- multiple literatures including: criminology, environmental psy- ness, sensation seeking, and consumer alienation are repeatedly chology, consumer behavior, and organizational deviance (Fox considered significant drivers of ethically questionable behav- and Spector 1999; Lindsay and Anderson 2000). Here, it is iors (e.g., d’Acremont and Van der Linden 2005). Support for argued that an individual’s personality traits and predispositions the broader link between psychological obstructionism and the will affect and shape how they interpret the world around them. severity of dysfunctional customer behavior is also evident In particular, aspects of psychological obstructionism (Machi- within the deviance literature. For example, traits and predis- avellianism, aggressiveness, sensation seeking and consumer positions pertaining to psychological obstructionism are argued alienation) are commonly argued to foster negative interpreta- to antecede a variety of criminal behaviors, employee misbe- tions of environments (e.g., Mudrack 1993; Slater 2003). That is haviors, and unethical behaviors (Al-Rafee and Cronan 2006; to suggest, individuals’ who are high in psychological obstruc- Harris and Ogbonna 2006). Focusing explicitly on the drivers tionism inherently observe the environmental settings in which of customer deviance, Fullerton and Punj (1993), forward that they encounter in a negative light, in comparison to persons individual personality traits and predispositions are crucial to who exhibit low levels of these traits. Offering insight into the understanding the antecedents of aberrant customer behavior. mechanism of this relationship concerning normative consumer Personality traits and predispositions that are considered to behavior within service settings, Bitner (1992) maintains that obstruct normative behavior are also recognized as important an individual’s analysis of a servicescape environment is influ- within McGrath and Goulding’s (1996) contemplation of cus- enced by their personality. Also focusing on consumers, Aylott tomer misbehavior within public service settings. Thus: and Mitchell (1998) uncover evidence to suggest that the manner in which individuals respond to stressors within the environment H4 . The greater the level of psychological obstructionism, the is shaped by their personality and predispositions. greater is the severity of dysfunctional customer behavior. Within the context of unethical and deviant behaviors, in examining the drivers of criminal behaviors, Eysenck and Influence of servicescape variables on the severity of Eysenck (1970) suggest that an individual’s enduring traits dysfunctional customer behavior are linked to their interpretation of the environment. Focusing on the aberrant acts of consumers, Fullerton and Punj (1993) Research that offers insight into the relationship between argue that individuals who are high in psychological obstruc- the perceived design of servicescape and the severity of dys- tionism traits are more prone to viewing servicescapes in a functional customer behavior is varied. To illustrate, Phillips, negative fashion owing to their predominantly negative out- Alexander, and Shaw (2005) highlight a link between the phys- look. This, Fullerton and Punj (1993) suggest, may ultimately ical design of self-service servicescapes and varying severities result in acts of aberrant customer behavior. Aligned with this of customer theft. Focusing on more severe acts of misbehav- argument, Robinson and O’Leary-Kelly (1998) draw on the per- ior within inner cities, based on a review of past research, Kuo spective of attraction–selection–attrition (Schneider 1987). In and Sullivan (2001) argue that the relationship between phys- particular, Robinson and O’Leary-Kelly (1998) suggest that per- ical environments and aggressive behaviors is well established sons who possess anti-social tendencies or personality traits are and widely accepted. Furthermore, in exploring this relationship, instinctively drawn to certain types of environments that fos- Homel and Clark (1994) find organizations that are perceived to ter such behaviors. Providing an aligned argument, Kenrick et be overcrowded, poorly ventilated, unclean, and noisy, experi- al. (1990) suggest that within certain environmental settings, ence higher rates of physical violence, than establishments that specific personality characteristics may become more apparent. possess qualities of ‘good’ physical design. Comparable find- Concurrently, Williams and Dargel (2004) discuss the propo- ings are offered by Graham et al. (1980) who forward evidence sition that individual personality and predispositions affect the of a statistically significant relationship between poorly main- way in which individuals screen, and thus ultimately respond to, tained, dirty, and unattractive service environments and incidents environmental cues. Thus: of customer aggression. Also assuming a holistic view of the servicescape, Rose and H3 . The greater the level of psychological obstructionism, the Neidermeyer (1999), stress that the manipulatable components greater is the negative interpretation of the outlet’s servicescape of service outlets including levels of crowding, background environment. music, ambient temperatures, and color schemes, may influence the severity of aggressive behaviors by consumers. Lawrence Influence of psychological obstructionism on the severity of (2004) broadens this proposition in arguing that organizations dysfunctional customer behavior should consider the initial design of servicescape environments at the planning stage of construction, thus avoiding the need to Within the sphere of criminology and indeed, the general later ‘react’ to acts of misbehavior once the outlet has started to psychological study of deviance, it is widely accepted that trade. In agreement, Dotter and Roebuck (1988) suggest that the
  6. 6. 326 K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 physical design of organizational settings may induce an array Measure development of misbehaviors. Indeed, in examining episodes of vandalistic behavior, Allen and Greenberger (1978) propose that the envi- The scales used were newly created or modified from existing ronmental design of an organization plays a role in driving its scales. Following standard psychometric scale development pro- own mutilation. Thus: cedures, first, we completed an extensive review of the literature to gain insight into the underlying dimensions of each construct. H5 . The greater the negative interpretation of an outlet’s ser- Second, we consulted 12 consumers, eight frontline employees, vicescape, the greater is the severity of dysfunctional customer three service managers, and four academicians during in-depth behavior. interviews. In addition, we employed Q-sort procedures as a In summary, the research model depicts both a direct and an means to assess each measure, with a panel of 21 judges (10 indirect relationship between psychological obstructionism and consumers, three frontline employees, four service managers, the severity of dysfunctional customer behavior. In addition, it and four academicians). Thereafter, we followed pretesting pro- denotes an indirect relationship between servicescape variables cedures that entailed two separate pilot studies of the research and the severity of customer misbehavior mediated by disaffec- instrument. During the first pilot study (n = 50), particular atten- tion with service. By contrast, the rival model advances direct tion was devoted to the phrasing of the severity measure and relationships between all three constructs and severity. However, ensuring that only those cases of misbehavior that occurred as indicated previously, we favor the composition of the research within the service setting were eligible. The second pilot study model and posit that it yields a significantly better fit with the (n = 66) trialed the refined measures. We examined the results data than the rival model. for reliability and validity and found that they met the standard benchmark criteria. Method Measures of constructs The hospitality industry is an ideal context for this study In addition to demographic and control measures, we used because of the sector’s economic importance and features, such eleven scales, seven of which we adapted from existing mea- as extended and close customer contact (Reynolds and Harris sures. We adopted seven-point Likert-type scoring for all 2006). Specifically, we considered the bar, hotel, and restaurant items because seven-point scales increase the reliability of sectors potentially fruitful contexts of inquiry. Indeed, sev- data findings. To measure the severity of dysfunctional cus- eral studies suggest that this industry is a particularly “potent” tomer behavior, we developed a four-item scale specifically environment in which to study the dynamics of customer mis- for the study (see Appendix A). We designed this scale to behavior (Harris and Reynolds 2003; Jones and Groenenboom gauge the extent to which the measured behavior violated the 2002). norms of the service outlet. Thus, we developed and used a A total of 1300 customers were approached in a public space four-item scale underpinned by the concept of norm viola- (e.g., shopping malls) and asked a screening question to (i) ascer- tion. To assess the robustness of our dependent measure, we tain their suitability (regarding, having deliberately behaved in employed three additional measures. First, respondents were a dysfunctional manner within a bar, hotel, or restaurant during required to indicate which form of behavior they had per- the past three months), and (ii) so that the researcher could pro- petrated. Second, during screening and according to social vide confidentiality assurances. Of the customers approached, norms, the researcher recorded (1) the severity of the behavior 696 declined to participate and 220 indicated that they had mis- and (2) the form of behavior performed. Subsequent anal- behaved in the past but not within a hospitality-based outlet, or ysis revealed strong correlations between the four measures had misbehaved in a services setting but not in the role as a cus- (p < .01), indicating a high degree of consistency in terms tomer. Consequently, 384 questionnaires were completed (four of perceptions of severity and form, both across the sample of which were incomplete and removed from the sample). This and between respondents and the researcher, and support for yielded a response rate of nearly 30 percent. Of the respondents, the self-reported severity scale employed in subsequent analy- 53.9 percent were female, the median age was 51 and the largest ses. group comprised respondents who earned between $40,000 and We gauged psychological obstructionism using four separate $60,000 annually. measures. We measured Machiavellianism using four items from Before completing the survey instrument, respondents were Christie and Geis’s (1970) MACH IV scale. A refined version required to recall and describe an incident of dysfunctional of Buss and Perry’s (1992) measure was employed to gauge behavior that they had undertaken. This enabled us to gain a aggressiveness. We assessed sensation seeking propensity using better understanding of the episode and to record both the form four items adapted from Steenkamp and Baumgartner (1992). and the severity of the behavior. This first stage of data collection We gauged consumer alienation using six items from Singh’s assists in stimulating memory, helps respondents complete the (1990) scale. questionnaire in a more focused frame of mind, and engenders We gauged disaffection with service using two scales. First, trust (Podsakoff et al. 2003). Thereafter, respondents individ- we assessed dissatisfaction with a four-item scale based on the ually completed a structured questionnaire that focused on the themes and items in the studies of Bloemer and Odekerken- single incident (see Appendix A). Schröder (2002) and Pizam and Ellis (1999). We used Oliver and
  7. 7. K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 327 Swan’s (1989) four-item scale of fairness to measure perceptions ceptualization of psychological obstructionism, servicescape, of inequity with the service. and disaffection with service as higher-order factors and severity We gauged servicescape variables using four scales. To assess and intoxication as lower-order constructs. Indeed, in line with the atmospheric characteristics of the servicescape, we drew and Hair et al. (2006), in comparing the higher-order structure with refined four items from d’Astous’s (2000) classification of ambi- a lower-order factor model, the second-order structure shows ent variables. Using the scale development procedure outlined superior predictive validity. previously, we developed five-, five-, and six-item measures, In each measurement model, the CFA results indicate good respectively, to gauge consumers’ interpretations of the lay- psychometric properties for all constructs. All loadings and cor- out and design of a service outlet, an organization’s exterior responding t-values at both the lower- and higher-order level environment, and the perceived behavior of fellow customers. were significant (t > 2.58), thus indicating convergent validity. We measured four control variables—sex, age, income, and Furthermore, we scrutinized Cronbach’s alphas, composite reli- level of intoxication—because each have been shown to influ- ability (CR), and average variance extracted (AVE) for each ence dysfunctional customer behavior (Harris and Reynolds scale. The lowest Cronbach alpha value was .85, and the lowest 2003). We also developed a five-item intoxication scale to mea- CR value was .79; all measures exceeded acceptable thresh- sure the degree of alcohol and/or drug intoxication. Finally, we olds. Furthermore, each AVE exceeded Fornell and Larcker’s assessed social desirability using four items derived from the (1981) suggested minimum value of .50. To examine the relia- study of Reynolds (1982). Appendix A presents all the measures bility of each higher-order construct, we calculated Nunnally’s and their reliabilities. (1978) formula for the reliability of linear combinations. All three higher-order reliabilities exceeded the cutoff of .70. Scale assessment We used two separate forms of analysis to determine discrim- inant validity. First, we conducted a series of CFA tests in which We used CFA to assess our measurement model. Other than we analyzed each possible pair of constructs. For every pair, we the dependent construct severity and the multi-item control vari- supply evidence of discriminant validity through a statistically able intoxication, we consider all factors in our theoretical model significant chi-square difference between the constrained and second-order constructs. Here, each first-order factor represents unconstrained model. Second, we assessed discriminant validity a reflective indicator for the higher-order construct. In theo- using Fornell and Larcker’s (1981) test. We found that each con- rizing each of the three higher-order constructs, we followed struct’s AVE is greater than the squared correlation between the the criterion of Hair et al. (2006). Consequently, we conducted two constructs. We also compared the AVE of each second-order second-order CFA (see Table 1). Because of sample-size con- measure with the squared structural link with other constructs straints in relation to the number of parameters to be estimated within the model. In all cases, the AVE was greater than the (Bentler and Chou 1987), we ran three separate CFA models that squared structural link, thus providing further evidence that all contained subsets of the most theoretically aligned variables. factors exhibit discriminant validity. We then assessed each measurement model using the elliptical Given the possibility of social desirability bias, we followed reweighed least squares estimation procedure. Podsakoff et al.’s (2003) recommendations. Using a latent mea- To evaluate the fit of the second-order psychological sure, we assessed the effect of social desirability at both the obstructionism construct, Measurement Model 1 (see Table 1) measurement model and the structural model stage of analy- comprises Machiavellianism, aggressiveness, sensation seeking, sis using a pair test approach. The results suggest that social and consumer alienation. Analysis of the fit indexes suggested desirability did not bias the data (see Graziano and Tobin 2002). good model fit (χ2 /df = 1.82, comparative fit index [CFI] = .98, In investigating if common method variance biases the data, nonnormed fit index [NNFI] = .98, and root mean square Podsakoff et al. (2003) recommend conducting Harmann’s sin- error of approximation [RMSEA] = .05). Measurement Model 2 gle factor test. This analysis was subsequently conducted and evaluated the servicescape variables of layout and design, atmo- satisfied. Finally, in order to assess the scope of misbehaviors spherics, exterior environment, and fellow customer behavior. captured within the data set, the form perpetrated as indicated As Table 1 shows, this second-order CFA measurement model by the respondent was analyzed. As summarized in Table 2, the represents a satisfactory fit with the data (χ2 /df = 1.83, CFI = .99, data represents a wide range of dysfunctional customer behav- NNFI = .99, and RMSEA = .05). Finally, Measurement Model 3 iors, both of unethical legal (36.7 percent) and unethical illegal assessed situation-specific factors of disaffection with service, (63.3 percent) orientation. Furthermore, employees, fellow cus- severity, and intoxication. We consider disaffection with service tomers and organizations were the most likely victims of the a second-order construct reflected by the first-order dimensions reported misdemeanors. of dissatisfaction and inequity. We estimate the constructs of severity and intoxication as individual first-order constructs. The Hypotheses testing measure of disaffection with service breached the three-item rule and therefore is underidentified. To remedy this, an equal- We adopted a parsimonious approach to estimate our two ity constraint is added to the disturbance terms, thus satisfying competing structural models to satisfy the five-to-one stipula- identification stipulations. As Table 1 shows, the results indi- tion of sample size to parameters (Bentler and Chou 1987). For cate an acceptable fit (χ2 /df = 1.98, CFI = .98, NNFI = .98, and each first-order factor, we used weighted composites, which we RMSEA = .05). All three measurement models support our con- derived in part from the first-order scale’s alpha coefficient in
  8. 8. 328 K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 Table 1 Measurement models. Measurement Model 1 Measurement Model 2 Measurement Model 3 Psychological obstructionism Servicescape variables Situation-specific variables First-order factors Standardized loadingsa First-order factors Standardized loadingsa First-order factors Standardized loadingsa Machiavellianism Layout and design Severity of DCB MACH1 .70b LAY1 .64b SEV1 .79b MACH2 .83 (13.13) LAY2 .77 (11.22) SEV2 .84 (15.45) MACH3 .85 (13.29) LAY3 .82 (11.79) SEV3 .85 (15.48) MACH4 .71 (11.51) LAY4 .90 (12.56) SEV4 .89 (16.46) LAY5 .91 (12.72) Aggressiveness Atmospherics Level of intoxication AGG1 .70b ATM1 .79b TOX1 .85b AGG2 .85 (14.63) ATM2 .88 (16.70) TOX2 .90 (20.74) AGG3 .87 (15.01) ATM3 .87 (16.38) TOX3 .97 (23.85) AGG4 .81 (15.01) ATM4 .83 (15.47) TOX4 .68 (13.14) AGG5 .87 (14.99) TOX5 .86 (18.78) AGG6 .71 (12.38) Sensation seeking Exterior environment Dissatisfaction SNS1 .66b EXT1 .90b DIS1 .94b SNS2 .81 (12.60) EXT2 .91 (24.12) DIS2 .96 (35.07) SNS3 .85 (13.07) EXT3 .81 (18.68) DIS3 .92 (28.80) SNS4 .89 (13.44) EXT4 .91 (24.34) DIS4 .89 (25.81) EXT5 .86 (21.24) Consumer alienation Fellow customers Inequity ATB1 .83b CUS1 75b INE1 88b ATB2 .78 (16.42) CUS2 .81 (14.54) INE2 .91 (21.21) ATB3 .84 (18.17) CUS3 .91 (16.52) INE3 .77 (15.86) ATB4 .84 (18.42) CUS4 .88 (15.93) INE4 .82 (17.34) ATB5 .65 (12.93) CUS5 .92 (16.87) ATB6 .89 (19.92) CUS6 .89 (16.24) Goodness-of-fit statistics Goodness-of-fit statistics Goodness-of-fit statistics χ2 /df = 1.82 χ2 /df = 1.83 χ2 /df = 1.98 CFI = .98 CFI = .99 CFI = .98 NNFI = .98 NNFI = .99 NNFI = .98 RMSEA = .05 RMSEA = .05 RMSEA = .05 a The t-values from the unstandardized solution are in parentheses. b Fixed parameter. Appendix A. Furthermore, each of the three second-order factors tistical support for H2 (β = .15, t = 2.72, p < .01) suggests that employed first-order composites as indicators. the higher the level of disaffection with service, the greater We present the results of the two rival models in Table 3. is the severity of dysfunctional customer behavior. We also The results reflect those of trimmed models. That is, on initial found support for H3 (β = .54, t = 5.58, p < .001), which focuses analysis of each model, we found that the effects of two of the on the relationship between psychological obstructionism control variables (income and intoxication) were consistently and the servicescape. Finally, the link between psycho- not significant (p > .05). Subsequent removal of these two control logical obstructionism and customer misbehavior (β = .37, factors improved the overall model fit statistics for both rival t = 4.66, p < .001) provides support for H4 . Thus, H1 –H4 are models. accepted. Research model Rival model The fit statistics indicate that the research model provides The goodness-of-fit statistics indicate that the rival model a good fit with the data (χ2 /df = 3.02, CFI = .92, NNFI = .91, represents a poor fit with the data (χ2 /df = 4.71, CFI = .85, and RMSEA = .07) and supports the four hypotheses (H1 –H4 ). NNFI = .82, and RMSEA = .10; see Table 3). Consistent with Table 3 documents the results for H1 (β = .49, t = 7.92, p < .001) our previous findings, analysis of the individual path coeffi- and indicates a relationship between negative interpretations of cients in this model indicates support for both H2 (β = .25, the servicescape and customers’ evaluations of disaffection. Sta- t = 5.06, p < .001) and H4 (β = .50, t = 6.04, p < .001). However,
  9. 9. K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 329 Table 2 Illustrative forms of dysfunctional customer behavior reported. Illustrations of forms of dysfunctional Percentage of respondents reporting Indicative percentage of victims of customer behavior reported perpetrating such behavior dysfunctional customer behavior reporteda 1. Failing to tell an employee when a 20 Employees = 42 mistake had been made in the Fellow customers = 10 respondent’s favor Organization = 92 2. Complaining without genuine cause 13.2 Employees = 33 Fellow customers = 21 Organization = 84 3. Using/consuming the facilities of a 13.4 Employees = 47 service outlet without intending to pay Fellow customers = 32 Organization = 90 4. Knowingly stealing an item from a 16.3 Employees = 63 service outlet Fellow customers = 36 Organization = 100 5. Arguing with, or being openly rude to 17.4 Employees = 86 a service employee or fellow customer Fellow customers = 44 Organization = 65 6. Knowingly damaging or vandalizing a 12.4 Employees = 59 service outlet’s property Fellow customers = 52 Organization = 100 7. Physically touching/striking a service 7.4 Employees = 90 employee or fellow customer Fellow customers = 43 Organization = 61 Note. Respondents were requested to indicate the form of misbehavior that most closely characterized their performed behavior. Consequently, this table constitutes a reflection of these behaviors, rather than an absolute and exact description of the individual misbehaviors perpetrated. a In many cases the misbehavior performed resulted in perceived consequences for more than one ‘victim’, this is reflected within the statistics shown. we find no support for H5 (β = .07, t = 1.52, p > .05), which Discussion suggests a direct relationship between the servicescape and the severity of dysfunctional customer behavior. Therefore, we The aim of this research was to conceptually develop and reject H5 . empirically test a framework of the factors that associate with dysfunctional customer behavior severity. In synthesiz- ing literature from disparate research domains and advancing Model comparison survey-derived results pertaining to our two competing theo- retical models, we offer empirical insight into these issues and To compare the goodness-of-fit with the data between the provide significant implications for marketing academicians and two structural models, in addition to reviewing the standard- practitioners. ized coefficients for each of the five hypothesized paths, we drew on the Akaike information criterion fit index. In review- ing the Akaike information criterion statistic across the two Theoretical contributions models (see Table 3), the research model is favored, thus confirming that it represents the best fit with the data. The This study makes four main contributions. First, by garnering chi-square difference statistic is also commonly used to assess and synthesizing literature from diverse academies and per- rival models that are hierarchical in nature. As is depicted spectives, our study contributes to existing knowledge by high- in Table 3, the difference in the chi-square value between lighting the pivotal role of three core concepts—psychological the two models is greater than 3.84, thus exhibiting statisti- obstructionism, disaffection with service, and servicescape cal significance and favoring the research model. Furthermore, variables—as the primary factors associated with dysfunctional in order to assess the credence of the espoused mediated customer behavior severity. Although previous studies have relationship between servicescape and severity through disaffec- tended to include at least one of these dimensions, to date, tion, a mediation analysis was conducted. The results revealed empirically based and holistic analyses have been lacking. Our support for the mediated relationship with a statistically non- review of the literature uncovers diverse theoretical traditions significant chi-square difference (p > .05) and non-significant that differ radically in their modeling of the dynamics between t-value between servicescape and severity, as depicted within these antecedents. Critical analysis leads us to advance two dis- the rival model. tinct models founded within diverse conceptual academies. In
  10. 10. 330 K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 Table 3 tion, the research design we employed contributes insights into Structural model results. the real behavior of customers, which laboratory-based stud- Hypothesized paths Research model Rival model ies are ethically and morally constrained from inducing (e.g., β (S.E.) t-value β (S.E.) t-value acts of violence). The supply of grounded data regarding cus- tomer misbehavior constitutes an important step in developing H1 : Servicescape → disaffection .49 (7.92) – a greater understanding of the darker, less salubrious side of H2 : Disaffection → severity of .15 (2.72) .25 (5.06) DCB service encounters and highlights the need for a broader per- H3 : Psychological .54 (5.58) – spective on customer dynamics that extends beyond the currently obstructionism → servicescape dominant (albeit understandable) emphasis on managerially pre- H4 : Psychological .37 (4.66) .50 (6.04) scriptive issues. obstructionism → severity of Third, our conception of dysfunctional customer behavior DCB H5 : Servicescape → severity of – .07 (1.52) posits that such actions are centered on societal, cultural, and DCB contextual norm breaking. Our research reveals multiple fac- Goodness-of-fit statistics tors across diverse customer behaviors that vary considerably χ2 190.50 301.75 in severity. This suggests that the broadening of emphasis to df 63 64 norm-breaking issues is likely to be both more insightful and χ2 /df 3.02 4.71 Probability .001 .001 more generalizable than narrow forms. In this regard, norm- CFI .92 .85 breaking issues are important. We are not suggesting that people NNFI .91 .82 prone to less severe forms of dysfunctional customer behavior RMSEA .07 .10 are equally prone to severe acts. However, our findings indicate Akaike information criterion 64.50 173.75 that the antecedents of multiple forms of customer misbehavior Note. DCB: dysfunctional customer behavior. are commonly shared (albeit at differing degrees). Thus, peo- ple with extreme psychological obstructionism are associated with extreme forms of dysfunctional customer behavior, par- this sense, each of the two competing perspectives provides the- ticularly when faced with extreme contexts and situations. Our oretical insight into the dynamics of dysfunctional customer findings strongly indicate that dysfunctional customer behav- behavior severity. Data analysis indicates that our preferred ior severity cannot be reliably attributed to a single stimuli, but model constitutes a significantly better fit with the data and leads rather is triggered by individual, situational, and contextual fac- to the rejection of the less robust rival model. This leads us to tors that amalgamate to elicit episodes of misbehavior (for a question the orthodox apparent within research domains that stay similar conception, see Fullerton and Punj 1993). firmly and narrowly focused within their own literature base. If Fourth, our study offers methodological contributions by we want to generate novel insights into phenomena that have a developing, operationalizing, and testing several new scales. In rich but diverse research traditions, such as dysfunctional cus- particular, the development of a novel, robust, multi-item mea- tomer behavior, it is both prudent and enriching to delve into sure of the severity of dysfunctional customer behavior is worthy such literature and immerse ourselves in the varied perspectives of comment. Although researchers have previously alluded to and positions that exist. Thus, we believe that our amalgamated the notion of severity in conceptual treaties (e.g., Robinson and conceptual approach can be fruitfully applied to other areas of Bennett 1997; Vitell and Muncy 1992), to date, research that research interest within marketing. focuses on deviant customer behavior has overly and narrowly Second, by undertaking the first holistic, survey-based study concentrated on, individual forms of misbehavior. In addition, of the factors associated with dysfunctional customer behav- although other studies have previously emphasized the severity ior, our research also makes an empirical contribution. In this of customers’ misbehavior, they did not attempt to operationalize regard, our study was motivated in part by a desire to respond this construct. We contend that to advance our comprehension of to the plethora of calls for empirical research into these issues incidents of dysfunctional customer behavior, new scale devel- (see Bitner, Booms, and Mohr 1994; Fullerton and Punj 2004; opment is essential. In developing, evaluating, and validating a Harris and Reynolds 2003; Wirtz and Kum 2004). Moreover, measure, we provide the methodological tools for researchers in we make a contribution through the study of actual incidents this area to progress and extend the understanding of customer of customer misbehavior. Previous research is weakened by an deviance dynamics. overemphasis on the employment of experimentation techniques to study artificial scenarios, typically in relation to a single Managerial implications determining factor and a single form of customer misbehavior (e.g., propensity to shoplift). Although the study of hypotheti- Although many practitioner-oriented commentaries dismiss cal situations can garner useful insights, our view is that such incidents of dysfunctional customer acts as random, irrational research is complemented by concurrent studies that focus on events that are endemic to some services contexts, the find- real people in real situations. Paper-based or computer-generated ings of this study refute such notions by highlighting that some contexts are unable to replicate perfectly a multiplicity of crucial, factors are subject to managerial control and manipulation. In determining factors (e.g., the ambient conditions of a service this regard, the impact of customer evaluations of disaffec- setting) that are associated with customer misbehavior. In addi- tion and servicescapes is especially important to consider. First,
  11. 11. K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 331 through the design and monitoring of the quality of service pro- prevent problem customers from disrupting service encounters, vision, customer service, and complaint and feedback structures, potentially saving firms considerable costs in time and money. practitioners can actively manage customers’ evaluations of dis- Second, in the vein of Berry and Seiders (2008), managers might satisfaction and inequity. Indeed, our study provides evidence consider ‘firing’ reoffending misbehaving customers exhibiting to suggest that rather than being unreasonable and illogical (at obstructionist tendencies, in order to prevent them becoming least from the consumer’s perspective); the seemingly “rational” ingrained within the framework of the servicescape and conse- motive of disaffection with service directly affects the severity quently having a negative effect on other patrons. Thus, although of dysfunctional customer behavior performed. the complete eradiation of customer deviance seems unfeasi- Second, while the findings of our study do not support a direct ble, tactical maneuvers are likely to reduce the severity of the relationship between servicescape and the severity of dysfunc- misbehavior. tional customer behavior, an indirect relationship, mediated by For executives and senior managers, this study highlights the disaffection with service is championed. Further, psychological merit of a strategic approach to the phenomenon of dysfunctional obstructionism is shown to have an indirect effect on severity customer behavior. Given an increased understanding of the fac- through servicescape and disaffection with service. Given these tors linked to customer misbehavior, managers should be able results, through the careful design or redesign of servicescape, to shape and develop systems, structures, and design priorities managers should be able to reduce the severity of dysfunctional calculated to monitor, minimize, and manage misbehavior. For customer behavior by creating environments that are satisfac- example, customers’ evaluations of dissatisfaction and inequity tory and exhibit a degree of ‘fit’ with patrons. While potentially can be reduced through the implementation of effective service fraught with difficulties owing to diverse consumer segments delivery, service failure, and customer feedback mechanisms. possessing varied levels of psychological obstructionist traits, Furthermore, the development of databases to record, track, this suggests that those charged with service environment design and scrutinize such incidents should allow managers to ana- should contemplate the target audience of the servicescape with lyze patterns, trends, and the frequencies of different forms of deviance in mind during embryonic design stages. Further- deviant acts. The insights gained should feed into company pro- more, the conceivable compatibility between patrons should cedures and policies, as well as both redesigned systems and also be acknowledged. Indeed, ensuring a degree of congruency servicescapes. between consumers, who owing to the nature of service provi- sion may have to spend extended periods in close proximity to Limitations and avenues for further research one another, may reduce incidents and the severity of dysfunc- tional customer behavior brought on by inter-client conflict and The findings and contributions of our study are bounded by subsequent disaffection. In this sense, although dysfunctional limitations that, in turn, highlight potentially fruitful avenues customer behaviors are unlikely to cease entirely, a key man- for further research. In particular, four limitations are espe- agerial implication of this study is that persons responsible for cially worthy of further discussion. First, the context of our the physical design of service environments and customer care study limits the extent to which we can universally generalize strategies are far from powerless in the proactive management the results and implications. Although we deemed the hos- of such behaviors. pitality industry as an appropriate setting, its idiosyncrasies The finding of a direct link between psychological obstruc- (e.g., extended customer contact) are far from universal. Thus, tionism and dysfunctional customer behavior severity also research should explore dysfunctional customer behavior in dif- indicates interesting and important implications for practice. ferent and contrasting contexts (varying servicescapes being a While our finding that enduring traits are linked to the sever- potentially fruitful avenue for future studies). We believe that ity of dysfunctional customer behavior may be interpreted by such research could build on the conceptions and measures we managers as unhelpful (as such factors are not subject to manage- employed herein and not only gauge the reliability and valid- rial control), such an interpretation is imprudent. Although the ity of the developed measure of the severity of dysfunctional findings of this study suggest that dysfunctional customer behav- customer behavior but also further explore the critical role of ior cannot be completely eradicated by judicious servicescape customer disaffection. design and improvements to service standards, an understanding Second, as the first study to conceptualize and then empir- of the psychological factors linked to customer misbehavior can ically test a range of factors associated with the severity of provide insights into the reduction of such events. First, man- dysfunctional customer behavior, our focus was on the prin- agers can develop training schemes to improve the abilities of cipal linkages. However, it would be naive to claim that these frontline employees in recognizing obstructionist traits and man- factors constitute an exhaustive list. Therefore, further research aging their subsequent interactions with consumers exhibiting could identify additional variables and extend understanding of levels of psychological obstructionism. In this regard, induc- the dynamics between such factors. For example, future research tion and training efforts should stress that, while consumers is needed to develop an understanding of the motives for such with such traits may be difficult to recognize, identifying poten- behaviors and to explore further, how third parties influence tial offenders and adjusting interaction styles and surveillance these behaviors. practices may reduce (but not eliminate) incidents. Through edu- Third, discussions with practitioners revealed a widespread cating customer-contact employees about the prevalence and assumption that the level of intoxication contributes to customer triggers of customer deviance, firms can better position them to misbehavior. However, in our study, we found no statistically
  12. 12. 332 K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 significant link. Research should explore this issue further. For 3. (SNS3) I would like to try an ‘extreme’ sport such as bungee example, does intoxication link to the frequency rather than the jumping. severity of dysfunctional customer behaviors? Are certain per- 4. (SNS4) I like to have new and exciting experiences and sen- sonality types more inclined to become intoxicated than others? sations even if they are a little frightening, unconventional, Fourth, our approach focuses on the retrospective analysis of or illegal. customer interpretations of real events. As such, similar to most survey-based approaches, we base our study on the assumption Consumer alienationa (α = .92, CR = .87, AVE = .54) that consumers truthfully and accurately recall events. Fur- ther, our focus has been on actual behavior rather than future 1. (ATB1) In general, the customer is usually the least important intentions or motives. Further research should explore the demo- consideration to most companies. graphic and psychographic factors linked to both behavioral 2. (ATB2) In general, shopping is usually an unpleasant expe- intentions to misbehave as well as the motives for such acts. rience. 3. (ATB3) In general, people must be willing to tolerate poor Appendix A. Construct and measurement items service from most businesses. 4. (ATB4) In general, companies are dishonest in their dealings Severity of dysfunctional customer behaviora (α = .91, with customers. CR = .84, AVE = .57) 5. (ATB5) In general, businesses who offer product and service guarantees will honor them. (reverse scored) 1. (SEV1) If others had witnessed my behavior, they would have 6. (ATB6) In general, most companies care nothing about the thought it was inappropriate behavior within that specific customer. outlet. 2. (SEV2) In hindsight, I acknowledge that my behavior is not Disaffection with service what is expected of customers within that service outlet. 3. (SEV3) I believe that others would generally view my behav- Dissatisfactiona (α = .96, CR = .90, AVE = .70) ior as acceptable in today’s society. (reverse scored) 4. (SEV4) If others had witnessed my behavior, they would have 1. (DIS1) I was dissatisfied with the level of service that I thought it was acceptable behavior within that specific outlet. received from the outlet. (reverse scored) 2. (DIS2) My expectations were not met. 3. (DIS3) I was dissatisfied with the quality of service that I Psychological obstructionism received. 4. (DIS4) I was very satisfied with the outlet. (reverse scored) Machiavellianisma (α = .85, CR = .79, AVE = .52) Inequitya (α = .92, CR = .85, AVE = .59) 1. (MACH1) Honesty is always the best policy. (reverse scored) 2. (MACH2) The majority of people are basically good and 1. (INE1) The outlet treated me fairly. (reverse scored) kind. (reverse scored) 2. (INE2) I was not treated right by the outlet. 3. (MACH3) Most people who get ahead in the world lead good 3. (INE3) I felt that the outlet was taking advantage of me. and honest lives. (reverse scored) 4. (INE4) I felt that the outlet behaved in an unfair way towards 4. (MACH4) A white lie is often a good thing. me. Aggressivenessa (α = .92, CR = .88, AVE = .56) Servicescape variables 1. (AGG1) Given enough provocation, I may hit another person. Layout and designa (α = .91, CR = .85, AVE = .55) 2. (AGG2) I rarely find myself disagreeing with other people. (reverse scored) 1. (LAY1) The interior of the outlet was designed to my taste. 3. (AGG3) When people annoy me, I tell them what I think. (reverse scored) 4. (AGG4) When frustrated, I let my irritation show. 2. (LAY2) It was very crowded inside of the outlet. 5. (AGG5) Some of my friends think that I am hot-headed. 3. (LAY3) The interior design of the outlet was unpleasant. 6. (AGG6) When people are especially nice, I wonder what they 4. (LAY4) It was very cramped inside of the outlet. want. 5. (LAY5) It was easy to move around the outlet. (reverse scored) Sensation seekinga (α = .88, CR = .82, AVE = .54) Atmosphericsa (α = .91, CR = .84, AVE = .57) 1. (SNS1) I do not like to try new foods that I have never tasted before. (reverse scored) 1. (ATM1) The temperature inside of the outlet was pleasant. 2. (SNS2) I prefer friends who are exciting and unpredictable. (reverse scored)
  13. 13. K.L. Reynolds, L.C. Harris / Journal of Retailing 85 (3, 2009) 321–335 333 2. (ATM2) The music inside of the outlet was too loud. a Seven-point scale (1 = “strongly disagree,” and 7 = “strongly 3. (ATM3) The air quality inside of the outlet was poor. agree”). 4. (ATM4) The outlet was very clean. (reverse scored) Exterior environmenta (α = .94, CR = .89, AVE = .62) References Allen, Vernon L. and David B. Greenberger (1978), “An Aesthetic Theory of 1. (EXT1) The exterior of the outlet was unappealing. Vandalism,” Crime and Delinquency, 24 (2), 309–32. 2. (EXT2) The outlet was located in a nice area. (reverse scored) Al-Rafee, Sulaiman and Timothy Paul Cronan (2006), “Digital Piracy: Factors 3. (EXT3) The outside of the outlet did not look well main- that Influence Attitude Towards Behavior,” Journal of Business Ethics, 63 (3), 237–59. tained. Areni, Charles S. (2003), “Exploring Managers’ Implicit Theories of Atmo- 4. (EXT4) The exterior of the outlet looked run down. spheric Music: Comparing Academic Analysis to Industry Insight,” Journal 5. (EXT5) The exterior of the outlet looked attractive. (reverse of Services Marketing, 17 (2), 161–84. scored) Aylott, Russell and Vincent-Wayne Mitchell (1998), “An Exploratory Study of Grocery Shopping Stressors,” International Journal of Retail and Distribu- tion Management, 26 (9), 362–73. Fellow customersa (α = .94, CR = .90, AVE = .60) Bäckström, Kristina and Ulf Johansson (2006), “Creating and Consuming Experiences in Retail Store Environments: Comparing Retailer and Con- sumer Perspectives,” Journal of Retailing and Consumer Services, 13 (6), 1. (CUS1) Fellow customers behaved in a pleasant manner. 417–30. (reverse scored) Baker, Julie and Michaelle Cameron (1996), “The Effects of the Service Envi- 2. (CUS2) Fellow customers behaved in a way that I was not ronment on Affect and Consumer Perception of Waiting Time: An Integrative expecting. Review and Research Propositions,” Journal of the Academy of Marketing Science, 24 (4), 338–49. 3. (CUS3) I enjoyed being around the other customers in the Bamfield, Joshua (2006), European Retail Theft Barometer: Monitoring the outlet. (reverse scored) Costs of Shrinkage and Crime for Europe’s Retailers. Nottingham, UK: 4. (CUS4) Fellow customers conducted themselves in a manner Centre for Retail Research, []. that I did not find appropriate. Barnes, James G., Brian R. King and Gordon A. Breen (2004), “The Almost 5. (CUS5) Fellow customers behaved in a way that I found to Customer: A Missed Opportunity to Enhance Corporate Success,” Managing Service Quality, 14 (2–3), 134–46. be unpleasant. Bentler, Peter M. and Chih-Ping Chou (1987), “Practical Issues in Structural 6. (CUS6) Fellow customers behaved in a way that I did not Modeling,” Sociological Methods and Research, 16 (1), 78–117. agree with. Berry, Leonard L. and Kathleen Seiders (2008), “Serving Unfair Customers,” Business Horizons, 51, 29–37. Bitner, Mary Jo (1992), “Servicescapes: The Impact of Physical Surround- Multi-item control variables ings on Customers and Employees,” Journal of Marketing, 2 (April), 57– 71. Intoxicationa (α = .93, CR = .88, AVE = .60) Bitner, Mary Jo, Bernard H. Booms and Lois A. Mohr (1994), “Critical Ser- vice Encounters: The Employee’s Viewpoint,” Journal of Marketing, 58 (4), 95–106. 1. (TOX1) I had consumed an intoxicating substance prior to Bloemer, Josée and Gaby Odekerken-Schröder (2002), “Store Satisfaction and visiting the outlet. Store Loyalty Explained by Customer- and Store-Related Factors,” Journal 2. (TOX2) I believe that I was under the influence at the time of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 15, 68–80. of the incident. Buss, Arnold H. and Mark Perry (1992), “The Aggression Questionnaire,” Jour- 3. (TOX3) I consider myself to have been intoxicated at the time nal of Personality and Social Psychology, 63 (3), 452–9. of the incident. Christie, Richard and Florence L. Geis (1970), “Studies in Machiavellianism,” 4. (TOX4) I consumed an intoxicating substance during my London: Academic Press. time in the outlet. d’Acremont, Mathieu and Martial Van der Linden (2005), “Adolescent Impulsiv- ity: Findings from a Community Sample,” Journal of Youth and Adolescence, 5. (TOX5) I was not intoxicated at the time of the incident. 34 (5), 427–35. (reverse scored) d’Astous, Alain (2000), “Irritating Aspects of the Shopping Environment,” Jour- nal of Business Research, 49 (2), 149–56. Diamond, Michael A. (1997), “Administrative Assault: A Contemporary Psy- Social desirabilitya (α = .89, CR = .82, AVE = .54) choanalytic View of Violence and Aggression in the Workplace,” American Review of Public Administrative, 27 (3), 228–47. Dotter, Daniel L. and Julian B. Roebuck (1988), “The Labeling Approach 1. (SD1) It is sometimes hard for me to go on with my work if Re-Examined: Interactionism and the Components of Deviance,” Deviant I am not encouraged. Behavior, 9 (1), 19–32. 2. (SD2) There have been times when I was quite jealous of the Dube, Jonathan (2003), “Office Wars: Clients Attack Employees When They good fortune of others. Feel Wronged”, ABC News (April 16). 3. (SD3) No matter who I am talking to, I am always a good Eysenck, Hans J. (1964), “Crime and Personality,” London: Routledge and Kegan Paul. listener. Eysenck, Sybil B.G. and Hans J. Eysenck (1970), “Crime and Personality: An 4. (SD4) I am sometimes irritated by people who ask favors of Empirical Study of the Three-Factor Theory,” British Journal of Criminol- me. (reverse scored) ogy, 10 (3), 225–39.