TESTING AN EXPANDED ATTITUDE MODEL OF GOAL-DIRECTED BEHAVIOR IN...
TESTING AN EXPANDED ATTITUDE MODEL OF GOAL-DIRECTED BEHAVIOR IN A LOYALTY CONTEXT
Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior
Catatan: INI HANYA sebuah Catatan, JANGAN dijadikan rujukan.Silahkan merujuk KE Sumber aslinya
Marketers have known for some time that satisfaction is closely linked to loyalty intentions (Oliver 1997, 1999). However, Johnson et al. (2006) argue that the drivers of customer loyalty intentions are complex and dynamic, changing and evolving over time. The current study therefore approaches the explanation of loyalty intentions in a different way. Specifically, the argument is made and a model is tested and supported of loyalty intentions based upon attitudinal, Goal-Directed conceptualizations. Specifically, we adapt the Model of Goal-Directed Behavior (MGB) posited by Perugini and Bagozzi (2001) and apply the model to loyalty in a B2B service context. First, the goal-directed, attitudinal explanatory model represented by the MGB conceptualization is supported by the current research. Second, two proposed extensions of the MGB are supported involving unique forms of attitude and perceived behavioral control. Finally, a multi-stage loyalty conceptualization is generally supported by the data. The managerial and research implications of the study are presented and discussed. [PUBLICATION ABSTRACT]
The quest for loyalty is one of the most frequently sought after strategic marketing objectives today (Evanschitzky and Wunderlich 2006; Oliver 1999). Indeed, there appears little controversy in the marketing literature surrounding the general belief that customer loyalty can differentiate firms and generate sustained profits (Evanschitzky and Wunderlich 2006; Keiningham, Vavra, Aksoy, and Wallard 2005; Oliver 1999; Rust, Lemon, and Narayandas 2005; Vavra 1992). Oliver (1999) argues for the determined study of loyalty with the same enthusiasm researchers have devoted to a better understanding of customer satisfaction.
None-the-less, there are still a number of important gaps in our understanding of the construct. First, there continues to exist no agreement on the definition of loyalty (Evanschitzky and Wunderlich 2006; Oliver 1999; Uncles et al. 2003). Second, much remains to be learned of the nature of the relationship between loyalty and antecedent influences such as satisfaction, value and trust (Agustin and Singh 2005), and why they appear so inconsistent in fostering loyalty. This observation represents somewhat of a "loyalty riddle" for marketers today. Third, there appears little explicit consideration of the more general body of knowledge related to judgment and decision-making (J/DM) in marketing inquiries related to consumer loyalty. Fourth, the role of moderating impacts on loyalty are still poorly understood (Evanschitzky and Wunderlich 2006). Together, such gaps in our understanding (1) make it difficult to consistently define loyalty both constitutively and operationally across studies, (2) which subsequently hampers our understanding of how loyalty forms, and (3) further attenuates relating loyalty to other marketing theories and marketing performance outcomes.
The following study first calls for moving beyond strictly behavioral views of loyalty (i.e., that people act) toward models that help us to better understand how loyalty judgments form (i.e., how people form the motivation to act). Specifically, we call for commensurability in adopting Oliver's (1997, 1999) attitude-based constitutive definition of customer loyalty. We then propose an attitude-based model of how such loyalty decisions are made based on a special case of Perugini and Bagozzi's (2001) Model of GoalDirected Behaviors (MGB), which purports to well-reconcile with emerging more general models of judgment and decision making (J/DM). The contribution of the proposed model lies in its ability to help us solve the riddle of why loyalty seems so hard to engender in so many target audiences. Specifically, instead of relying solely on the often inconsistent influence of satisfaction on loyalty (Oliver 1999), marketers can instead identify the specific psychological antecedents (both cognitive and affective) motivating loyalty intentions and subsequent behaviors in marketing strategic management. Together, the results constitute a framework that may help begin the process of unifying loyalty research with the extant body of knowledge across social science disciplines. Thus, the presented study contributes to the body of knowledge for both marketing practitioners and theoreticians, and potentially beyond.
This study builds upon the argument that a cognitive and affective perspective of loyalty is consistent with the emerging marketing literature. For example, studies are emerging demonstrating the key role of affective commitment in strengthening loyalty to a brand (Fullerton 2003, Matilla 2006). Ratner and Herbst (2005) conduct four experiments and conclude that an emotional reaction to a negative outcome can affect switching behaviors. Olsen et al. (2005) conduct a study identifying the importance of taking ambivalence into consideration when measuring satisfaction and modeling satisfaction-loyalty relationships. Yu and Dean (2001) also suggest that the emotional component of satisfaction is a driver of loyalty. Consistent with the emerging literature, the following study builds upon the assertion that marketing models attempting to explain the evolution of loyalty intentions need to consider affective considerations. However, we argue that in order to consider affective considerations in loyalty models, loyalty theory must be reconciled with the underlying literatures related to both judgment and decision making (J/DM) and attitudes. Such reconciliation suggests modeling the formation of loyalty intentions in new ways.
The marketing literature most generally identifies the existence of two competing views of loyalty. The first view, which we will call the behavioral perspective, considers the domain of the concept of loyalty to essentially be constrained to retention of the brand (e.g., East et al. 2005). This perspective focuses almost exclusively on repeat purchase. The emphasis is entirely on behavioral measures. The gist of this perspective appears to be to focus on operationalizing the loyalty concept in as parsimonious a manner as possible. This perspective appears most popularly advocated by Reichheld's (2003) argument to simply measure word-of-mouth behaviors in order to measure customer loyalty.
That said, Uncles et al. (2003) suggests that loyalty represents a paradox. They explain the decision-making process underlying the behavioral perspective essentially involves passive acceptance of brands. An alternative perspective involves considering loyalty to be an attitudinal-based phenomenon, which we will refer to as the attitudinal perspective. While both perspectives appear to assume goal-directed behavior, the behavioral perspective identified above appears to be a relatively more (short-term) sales-directed perspective. The attitudinal perspective can arguably be characterized as a more defensive alternative that presupposes the efficacy of customer relationship management initiatives.
The attitudinal perspective is the approach taken in the current research based upon the Morgan and Rego (2006) study demonstrating the misguided nature of the behavioral perspective as identified above. Our review of the literature suggests that Oliver's (1999, p. 34) definition of loyalty represents a viable candidate for commensurable agreement as a constitutive definition for attitudinally-based loyalty: "a deeply held commitment to rebuy or repatronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior." We view this definition as adequately capturing both behaviorally- and attitudinally-based loyalty models identified above. Consistent with Jacoby and Chestnut's (1978) original expectancy-value based assertions, Oliver (1999) suggests that loyalty includes four hierarchical phases: cognitive, affective, conative (i.e., behavioral intention), and action. However, this conceptualization differs in the argument that consumers become "loyal" at each attitudinal phase relating to different elements of the attitude development structure.
Consumers are theorized to become loyal first in a cognitive sense, then later in an affective sense, and finally in a behavioral sense (i.e., action inertia). In addition, defining loyalty as a construct with cognitive, affective, conative, and action phases suggests that explanatory models should incorporate similar constructs. Not inconsistent with Oliver's (1997, 1999) conceptualization of hierarchical loyalty we have defined loyalty in the current research at two general levels: repurchase and fortitude. In addition, consistent with Reichheld (2003), we include word of mouth as a dependent variable in our model. Table 1 presents Oliver's (1997, 1999) theoretical model of loyalty, and how they relate to our conceptualization.
In summary, the current study investigates customer loyalty consistent with a goal-directed, attitudinal perspective that seeks to incorporate affective considerations in the decision-making model. However, prior to presenting the research model investigated herein, we first address the efficacy of our research strategy when compared to emerging theoretical considerations from the judgment and decision making (J/DM) and attitude literature.
Reconciling Loyalty Theory with the J/DM and Attitude Literatures
The vast majority of evidence related to choice under risk or uncertainty is based upon theories that are cognitive in nature (Lowenstein et al. 2001). That is, people assess choice alternatives and then integrate this information through some form of expectation-based calculus to arrive at a decision. As attitude theories have origins in multi-attribute utility theory (MAUT), attitude theory appears accurately described as having evolved based upon the consequentiality perspective as described by Lowenstein et al. (2001). Such a perspective suggests that consumers base decisions on the consequences of their actions. However, one of the problems associated with the consequentialist perspective has been its inability to easily incorporate affective considerations (Kahneman and Tversky 2000, Tversky and Kahneman 2000). Hsee et al. (2004), however, are proponents of a perspective that purports to reconcile affective considerations with the consequentialist perspective.
The Hsee et al. (2004) perspective allows for the contemplation of affective considerations within the context of attitude models (as a form of J/DM models). In fact, the MGB represents an effort to specifically accomplish this end. MGB differs from previous attitude models such as the Theory of Planned Behavior by (1) introducing emotions into traditional attitudinal explanations of goal-directed behavior through positive and negative emotions (As), and (2) accounting for motivation through the capture of desires as part of the model. These anticipated emotions influence volitional processes (i.e., cognitive planning), which in turn influence the development of desires (i.e., motivation, which subsequently influence instrumental behaviors). Figure 1 presents MGB specific to loyalty inquiries. However, the current research extends beyond simply applying MGB to loyalty considerations, but also includes extensions of the MGB which are appropriate for this context.
Proposed Extensions to the MGB Conceptualization & Research Hypotheses
A review of the literature identifies two potential extensions of the MGB conceptualization that also merit consideration within the current research. The first extension involves the appropriate conceptualization and operationalization of the Attitude^sub Act^ construct. Perugini and Bagozzi (2001, p. 81) define "attitudes" in the MGB conceptualization consistent with that typically associated with the Theory of Planned Behavior (see Eagly and Chaiken 1993): "Attitude is conceived as a 'psychological tendency' that is expressed by evaluating a particular entity with some degree of favor or disfavor." However, evidence is emerging supporting calls for specifically considering separate affective evaluations when measuring attitudes (Hagger and Chatzisarantis 2005, Okada 2005, Van Den Berg et al. 2005). Voss et al. (2003) report a multidimensional, parsimonious, reliable, and valid scale of consumer attitudes toward brands and product categories based on a two-dimensional conceptualization of the construct: hedonic and utilitarian attitude components. We suggest that the measures employed by Perugini and Bagozzi (2001) to measure Attitude^sub Act^ may obscure the unique contributions of hedonic and utilitarian consumer attitudes toward the act of consumption. This distinction is important given the previously identified linkages of the model to affective and cognitive considerations. Therefore, the current research first tests the validity of a two dimensional conceptualization of Attitude^sub Act^ based upon the scale developed by Voss et al. (2003), and then assesses whether utilitarian and hedonic attitudinal proclivities differentially influence respondents' motivations as desires. We further suspect that utilitarian attitude forms will be more influential in the current research setting based upon discussions with knowledgeable industry managers.
The second proposed extension concerns perceived behavioral control. Perugini and Bagozzi (2001) operationalize PBC using measures of both difficulty and control in a unidimensional scale. However, Trafimow et al (2002) present evidence that PBC is best conceptualized as a two-dimensional construct: perceived control and perceived difficulty, with perceived difficulty generally being a better predictor of behavioral intentions than perceived control. In the current study an assessment is performed to determine whether difficulty and control differentially influence the MGB as unique exogenous constructs. Figure 2 and Table 2 present the model assessed in the current research that incorporates both the original MGB conceptualization and the modifications proposed herein.
As discussed above, the current study includes utilitarian and hedonic dimensions of attitude. The utilitarian dimension of an attitude is defined by Voss et al. (2003) as being derived from the functions performed by products, being that the scale was developed in a product context. In the present context, the utilitarian dimension of attitude would be the function derived from the act of being loyal. Loyalty would be functional in that it would make the choice process easier. Information processing theorists suggest decision makers value effort reduction in the decision process (Bettman, Luce, and Payne 1998). An increasing utilitarian dimension should therefore be associated with increasing desire (i.e., motivation) to be loyal.
H1a: Utilitarian attitude will be positively associated with increasing desire to be loyal.
The hedonic dimension of an attitude is defined as the sensations derived from the experience of using the product (Voss et al. 2003). In the present context, the hedonic dimension would represent the sensation derived from the act of being loyal. Loyalty is associated with decreasing cognitive effort. Research offers evidence that increasing cognitive effort is associated with increasing negative affect (Garbarino and Edell 1997). Assuming a converse process, reductions in decision effort should be associated with increasing positive affect. Therefore:
H1b: Hedonic attitude will be positively associated with increasing desire to be loyal.
The MGB suggests that subjective norms will also be related to desires (Perugini and Bagozzi 2001). Subjective norms refer to consumers' overall perceptions of what relevant others (e.g., friends, family) think he or she should do (Evans, Christiansen, and Gill 1996). Evidence suggests that what we perceive others to think we should do will impact our desires (Perugini and Bagozzi 2001). These significant others would involve professional peers and colleagues within the context of the current research. Therefore:
H2: Subjective norms will be positively associated with increasing desire to be loyal.
As previously discussed, research offers evidence that negative and positive anticipated emotions impact desires (Perugini and Bagozzi 2001). Anticipated emotions are the emotions that a person expects to experience by achieving a sought after goal (Bagozzi, Baumgartner, and Pieters 1998). These are not the emotions experienced while being loyal, but are the anticipated emotions that consumers weigh when deciding whether to pursue a goal of being loyal. According to theory the process would work like this: first, consumers form a goal; second they consider the consequences of achieving the goal with the corresponding emotions arising termed anticipated emotions (Perugini and Bagozzi 2001). Research offers evidence that when a consumer considers a goal, anticipated emotions influence the level of motivation for that goal. In the MGB, anticipated emotions are positive when associated with achievement of a goal and negative when associated with not achieving a goal. Evidence suggests desire is influenced both by the expected emotions of achieving and not achieving a goal. Therefore:
H3: Positive anticipated emotions associated with achieving the goal of being loyal will be positively associated with an increasing desire to be loyal.
H4: Negative anticipated emotions associated with failure to achieve the goal of being loyal will be positively associated with an increasing desire to be loyal.
Perceived behavioral control is defined as the extent to which consumers consider the performance of a behavior to be under their voluntary control (Trafimow et al. 2002). In the context of loyalty, perceived behavioral control represents the degree to which a consumer believes loyalty to a firm is within their control. In the current B2B context, perceived behavioral control could be impacted by the degree to which the respondent believes they are able to make the decision. Desires can be generally defined as motivation. A consumer's belief that they have the control necessary to be loyal would seem to be a baseline necessity for them to desire to be loyal. Otherwise it would be a wasteful consideration. Perugini and Bagozzi (2001) offer evidence that perceived behavioral control has a positive association with desires. Therefore:
H5a: Perceived behavioral control of loyalty will be positively associated with desire to be loyal.
Perceived difficulty is defined by Trafimow et al. (2002) as the degree to which consumers consider a behavior to be easy or difficult to perform. In the current B2B context, the politics of the workplace might make it more difficult for a buyer to be loyal to a particular firm. If a consumer believes that being loyal is easier, the likelihood that they will be loyal should increase. Therefore:
H5b: Perceived difficulty of loyalty will be negatively associated with desire to be loyal.
Perceived behavioral control and perceived difficulty should be directly related to loyalty intentions. Both control and difficulty should lead directly to intentions because they are a baseline necessary for intentions to form. If consumers believe that they have control and that the task is easy there is the possibility of intention, even if there is no motivation. Therefore:
H6a: Perceived behavioral control of loyalty will be positively associated with loyalty repurchase intentions.
H6b: Perceived behavioral control of loyalty will be positively associated with loyalty fortitude intentions.
H6c: Perceived behavioral control of loyalty will be positively associated with word-of-mouth behavioral intentions.
H6d: Perceived difficulty of loyalty will be negatively associated with loyalty repurchase intentions.
H6e: Perceived difficulty of loyalty will be negatively associated with loyalty fortitude intentions.
H6f: Perceived difficulty of loyalty will be negatively associated with word-of-mouth behavioral intentions.
Recent research expands traditional models by suggesting that desire mediates the relationship between attitude and intention (Perugini and Bagozzi 2001). According to this view, intention does not encompass the motivation that is necessary for an intention to be formed. Desire represents this motivation. Research offers evidence that desire is a distinct construct from intention (Perugini and Bagozzi 2004), and does mediate the relationship between an attitude and behavior (Perugini and Bagozzi 2001). Desire is defined as "a state of mind whereby an agent has a personal motivation to perform an action or to achieve a goal" (Perugini andBagozzi 2004, p. 71). Desire is said to "represent the motivational state of mind wherein appraisals and reasons to act are transformed into a motivation to do so (Perugini and Bagozzi 2001, p. 84)."
A basic assumption underlying this research is that consumers have a goal to be loyal. Loyalty makes decision making more efficient. Once consumers determine an initial solution, they can reduce choice effort by remaining loyal. Research suggests that the reduction of effort is an underlying motivation for decision makers (Bettman, Luce, and Payne 1998). In today's marketplace which is characterized by hyperchoice (e.g., Mick, Broniarczlyk, and Haidt 2004), loyalty is one strategy for reducing effort. Thus loyalty can be modeled as a goal. The MGB suggests that desires will mediate the relationship between attitude and behavioral intentions. Therefore:
H7a: Desire to be loyal will be positively related to loyalty repurchase intentions.
H7b: Desire to be loyal will be positively related to loyalty fortitude intentions.
H7c: Desire to be loyal will be positively related to word-of-mouth behavioral intentions.
Some research suggests that different stages of loyalty have a hierarchical structure. Oliver (1999) suggests that cognitive loyalty precedes affective loyalty, affective loyalty precedes conative loyalty, and conative loyalty precedes action loyalty. Consistent with this view, we suggest that loyalty repurchase intentions will precede loyalty fortitude intentions, and loyalty fortitude intentions will precede word of mouth behavioral intentions. Consumers demonstrating brand repurchase loyalty buy because of compelling information (e.g., quality), a cognitive orientation. Consumers demonstrating brand fortitude loyalty buy because of a commitment to the brand with both cognitive and affective dimensions. Behavioral word of mouth intentions are an outcome of the other two stages of loyalty. Therefore:
H8: Loyalty repurchase intentions will have a positive relationship with loyalty fortitude intentions.
H9: Loyalty fortitude intentions will have a positive relationship with word-of-mouth behavioral intentions.
The Population and Obtained Sample
The population selected for empirically assessing the research model presented as Figure 2 involved key decision makers of accounting firms concerned with the purchase of professional liability insurance (PLI). The selection of this particular target population was made for several reasons. First, PLI and the accounting discipline both represent "pure" service considerations, a rigorous loyalty situation in which to test the research hypotheses. Second, the recent enactment of the Sarbanes-Oxley legislation underscores the need for PLI due to changes in the ethical perceptions of the accounting industry (Fletcher 2003). Consequently, PLI should be important and somewhat involving to respondents. A list of "Managing Partners" from the American Institute of Certified Public Accountants (AICPA) was purchased after discussions with industry managers and researchers to identify the best possible sampling frame. Two thousand surveys were sent to managing partners of accounting firms across all fifty states in the US, offering a $10 gift card as an incentive to encourage participation. All potential respondents received a follow-up postcard one week later.
A total of 210 completed surveys were returned, plus 79 that were nondeliverable for one reason or another. This yielded a working response rate of 2.8%. However, 20 surveys were discarded because respondents failed to completely answer the loyalty scales. The authors are encouraged by the response rate given generally declining survey response rates (Dillman 2007), as well as the correspondence of the study with the beginning of the 2004 tax season. Dillman (2007) identifies the challenges associated with meaningfully measuring nonresponse error, which we attempted to address by increasing our response rate through the reminder postcard. The original cover letter accompanying the survey instrument also encouraged recipients to personally complete the survey. In addition, prior to model testing, the obtained sample was scrutinized to compare it with the population of interest based upon demographic variables. Discussions were solicited with industry experts possessing extensive long-term experience in the PLI industry relative to accountants, with all agreeing that the obtained sample very closely corresponds to the known demographic characteristics of the population of interest (largely Caucasian, middle-aged males). The vast majority of respondents have extensive accounting experience (between 11 and 40 years), with more than half having held their current PLI policies for less than 11 years, signifying little reason to believe that non-targeted respondents completed any of the returned surveys. Thus, the conclusion is supported that the obtained sample appears representative of the population of interest.
The Measures Employed to Operationalize the Constructs
Appendix A presents the measures used in the reported study, as well as their constitutive definitions and sources. Those measures specifically developed for this study were based upon constitutive definitions found in the literature. In addition, all of the measures employed are of a global and reflective nature (see Jarvis, McKenzie, and Podsakoff 2003). That is, they assert to directly measure the underlying core of the domain in a relatively comprehensive manner. The measures used in the current research therefore possess a measure of face validity in that most of the scales rely on measures previously reported in this line of inquiry.
However, before moving on to a discussion of the reliability and validity of the measures, two issues merit discussion. First, loyalty intentions were divided into two general categories of repurchase intentions versus fortitude forms of loyalty in a manner not inconsistent with the Oliver's (1997, 1999) conceptualization (see Table 1). In addition, word-of-mouth behaviors were treated as an independent endogenous dependent variable consistent with the recommendations of Reichheld (2003) and Soderlund (2006). Second, Bagozzi et al. (2001) argue that it remains unclear whether satisfaction is phenomenologically distinct from other positive emotions. Specifically, they point to the fact that satisfaction is neither a basic emotion nor a central emotional category in leading theories of emotions. In other words, these authors suggest that satisfaction may just be another reflective indicator of positive emotions. Based on this assertion, we included positive anticipated satisfaction as a measure of positive anticipated emotions (see Appendix A). We therefore test the proposition that anticipated satisfaction may operate as a reflective indicator of positive anticipated emotions.
The reliability and variance extracted validity scores associated with the measures are reported in Appendix A. Hair et al. (1998) suggest calculating construct reliability estimates and variance-extracted measures. Raines-Eudy (2000) states that calculated share variance scores exceeding 50% are the recommended criteria for model constructs. Appendix A demonstrates that the scales used for the model constructs all exceed recommended criteria for reliability and validity in measurement except for previous behavior. Consequently, previous behavior was dropped from subsequent analyses. In addition, following Anderson and Gerbing (1988), we employed the two-step approach for demonstrating discriminant validity. Table 3 presents the PHI matrix of the confirmatory factor analysis of the measurement model (χ^sup 2^ = 1068.79, RMSEA = 0.042, CFI = .98, and SRMR = 0.055). Table 4 presents the χ^sup 2^ difference tests further supporting discriminant validity among the constructs most likely to lack discriminant validity in measurement. Thus, evidence is apparent for sufficient reliability and validity in measurement to proceed to hypothesis testing using SEM.
All of the hypotheses are tested using structural equation analyses (SEM) using LISREL 8.80. The data was normalized prior to analysis using the normalization algorithm in PRELIS consistent with need for normality in multivariate analyses (Hair et al. 1998). As is often inherent in survey-based research, a small number of missing values in responses were noted related to measures of the endogenous and exogenous constructs.1 This can be problematic in that LISREL provides full information and multiple imputation based on the assumption that data are missing at random. Missing values associated with the exogenous model variables were addressed prior to analyses by using the linear interpolation method within the SPSS Missing Values statistical software package.
McQuitty (2004) posits that it is possible to estimate the power associated with the test of an entire (SEM) model. He argues that addressing this matter is essential as statistical power directly affects the confidence with which test results can be interpreted. The obtained sample in the current research provides statistical power in excess of .9 based on the indices provided by McQuitty (2004).2 Further encouragement is provided by the recent results of Curren et al (2003) who assert that both sample estimates and confidence intervals of RMSEA's are accurate using SEM with sample sizes of ≥ 200. Thus, sufficient statistical power exists to proceed to analyses using SEM.
Table 5 presents results of hypothesis testing of the research model presented as Figure 2 and Table 2. The results support a number of conclusions:
1. Desires are shown to be most strongly influenced by utilitarian forms of attitude, followed by positive anticipated emotions (including anticipated satisfaction), subjective norms, and hedonic forms of attitude.
2. Initial repurchase forms of loyalty intentions are a function of the motivation to be loyal (i.e., desire) and the perceived difficulty associated with the act of repurchase.
3. Subsequent fortitude forms of loyalty intentions are a function of motivation as desires and repurchase forms of loyalty intentions, followed by perceptions of control. This result merits further discussion as the sign associated with the finding is negative, suggesting that the greater the perceived control the manager has over the behavior, the less likely the manager's intention to demonstrate fortitude forms of loyalty. Some insight into this finding is apparent by examining the PHI matrix presented as Table 3. As expected, control is negatively related to perceptions of difficulty. However, control is also negatively related to fortitude forms of loyalty, desires, and positive anticipated emotions, while positively associated with negative anticipated emotions. Discussions with industry managers with long-term experience speculate that this finding may be capturing the negative affect associated with the additional accountability that often accompanies greater control.
4. Word-of-mouth intentions are found to be solely driven by fortitude forms of loyalty in the current research.
In summary, the results offer some support for the major assertions of the current research. First, many of the components of the Goal-Directed, attitudinal explanatory model represented by the MGB conceptualization are supported by the current research (Please see Table 2). Second, the proposed MGB extensions involving unique forms of attitude and perceived behavioral control are supported herein. Finally, a multi-stage loyalty conceptualization is generally supported by the data.
In spite of the interesting nature of the results reported herein, we caution readers to recognize that a single study should never be sufficient evidence to create fully-accepted new theory. No research project is devoid of limitation, and the current study claims no exception. Identifiable limitations might first include the obtained sample size in spite of our arguments supporting conclusions of representativeness, reliability, validity, and sample power sufficiency.
Second, the study would have benefited from measures of behaviors, not just behavioral intentions. Behavior, in this context, is the act of remaining with a provider. Given that most do remain with the provider for an extended amount of time, it makes sense to measure loyalty intentions rather than behavior. Behavior would not necessarily represent loyalty as it could be due to several factors (e.g., inertia, switching costs) which have little to do with psychological attachment to the firm. For this reason, we use loyalty intentions, rather than behavior, as our dependent variable. The use of intentions as opposed to behavior may help to explain some of our weaker findings such as perceived behavioral control.
Third, our study did not measure the degree to which switching costs or unique knowledge may have explained loyalty behaviors. However, given our focus on attitudinal and behavioral measures of loyalty, we have some evidence that loyalty instead of switching costs explained intentions.
Fourth, we did not test frequency and recency effects given the selected research setting. Recency and frequency effects were omitted based on the recommendations of managers suggesting that the decision to continue a PLI policy was not universally consistent across time. In choosing a context, we wanted one that offered the opportunity for loyalty to develop; suggesting businesses need to be familiar with the service provider. However, the downside of this is that frequency of past behavior is limited, given that firms rarely change their insurance provider. In other words, many people make this decision once, and then revisit it only when there is a need.
Finally, given the dearth of such considerations within B2B contexts, much of the theory developed in the current research is generalized from consumer research. However, given the theoretical development of the research model from the literatures of a wide variety of social sciences, coupled with supported empirical efficacy of results, the finding associated with this study arguably provide a significant step forward in our understanding of the formation of loyalty intentions. We encourage efforts to replicate the results reported herein.
RESEARCH AND MANAGERIAL IMPLICATIONS
We began this study with the proposition that models of customer loyalty can and should incorporate affect based on emerging evidence from the J/DM and attitude literatures. We reject the notion that inquiries into loyalty should be grounded in purely behavioral considerations, and develop and test a model of loyalty based upon an arguably more valid attitudinal approach. So, what is gained based on the current research?
We begin by looking at the research implications. The preceding literature review makes clear that typical approaches to explain loyalty are initially attenuated by irregular constitutive and operational definitions. Significant advances in our understanding must begin with commensurable agreement as to the appropriate conceptualization and operationalization of the loyalty construct. The second typical practice that attenuates our understanding of customer loyalty concerns the common absence of reconciliation with underlying models of J/DM in marketing explanations of loyalty decisions. We suggest that this study provides a framework that can help overcome both of these impediments to the advancement of our understanding of loyalty. Academicians and/or practitioners who care to understand the motivational influences underlying loyalty judgments and decisions can replicate and extend the models proposed herein to inform their understanding of loyalty, and particularly the underlying motivations.
There are also a number of very interesting puzzles that remain to be answered specific to customer loyalty. The first puzzle involves whether or not a direct path should be modeled between affect and behavior, or whether affect always operates via an interaction with cognition. Such an assumption appears implicit in the prevailing models of loyalty based upon the argument that quality perceptions [arrow right] satisfaction judgments [arrow right] loyalty intention. Assertions concerning experienced utility aside, Loewenstein et al (2001) and Anderson (2003) present arguments that affective influences are difficult to reconcile with explanatory models of J/DM under conditions of risk. The gist of their argument is as follows: Objective or judged probabilities may be the same across two situations, yet feelings in the two situations may diverge, potentially yielding different preferences (i.e., those more affectively influenced). Prospect theory requires that they yield equivalent preferences. While interesting, our own review of the literature suggests that this issue remains unresolved and therefore constitutes an important area of future inquiry, particularly in light of experienced utility. We summarize that loyalty research must eventually be reconciled with both cognitive and affective explanations of J/DM.
This call is consistent with Rottenstreich and Shu (2004) who identify what is known about the connections between affect and decision making based upon the traditional expected utility-based paradigm and its derivatives. First, in terms of the deliberate, calculative versus affective, more automatic valuation processes: (1) the use of different valuation processes may contribute to preference pliability, (2) valuation by feelings are relatively scope-insensitive yielding step-function valuation functions, whereas calculative valuation is relatively scopeinsensitive yielding more linear value functions, and (3) valuation by feeling appears to yield greater loss aversion than calculative valuation. second, the weighting function can also be impacted by affect under uncertainty: (1) affectrichness potentially yields pronounced certainty and impossibility effects and extreme insensitivity to intermediate probability variations, (2) affectrichness may contribute to the elevation of w in models such as prospect theory, (3) the influence of affect is likely to depend upon the imagery evoked, and (4) affect, especially in the form of mood, appears to bias judgments of likelihood. Finally, the introduction of emotions to traditional explanatory models is complicated by the observation that emotional reactions can occur at the time the decision is made, after the decision but before the consequences are realized, or after the consequences are realized.
These conclusions support an expectation that virtually all expected utility-based J/DM explanatory models can be expected to vary across situations and circumstances. Therefore, it is not at all surprising that loyalty appears irrational when studied by models that do not explicitly account for affective considerations. This probable model inconsistency across settings and circumstances underscores our call for greater emphasis on replication in support of loyalty studies. Such a call is further supported by Norenzayan and Heine's (2005) compelling argument that generalization of results of psychological studies beyond one's sample is inherently risky due to an absence of psychological universels. A stronger emphasis on the identification of relevant psychological universals underlying loyalty judgments and decisions also appears as a worthwhile research endeavor supporting such calls.
The second obvious puzzle facing loyalty researchers concerns how to reconcile the plethora of empirically supported known antecedents (e.g., satisfaction, trust, switching costs, brand equity, value, etc.) to loyalty with attitude-based models such as are presented herein. We suggest that the surest path to reconciling the myriad of results reported related to loyalty is to agree upon an underlying J/DM model. Thus, we call upon future marketing research related to the loyalty construct to identify and defend results based upon a comparison to underlying models of J/DM. In order to answer such a call, we assert that some formal consideration of the cognitive and affective antecedents to loyalty intentions/behaviors is desirable.
This suggests including modeling perspectives related to attitudinally-based, goaldirected explanatory models. For example, the recent work of Hagger and Chatzisarantis (2005) suggests considering higher-order constructs in attitude models. Given the growing emphasis on the multidimensional nature of the constructs underlying attitudinal models of J/DM, such a modeling emphasis may prove to be particularly promising. Complementing this assertion, Perugini (2005) highlights the importance of predictive models of implicit versus explicit attitudes, which appears a promising means of extending the current research (also see Neumann et al. 2004). MacDonald and Nail's (2005) arguments related to attitude change and the public-private attitude distinction also merit consideration in future research. Bagozzi et al. (2002) point out that a number of individual and/or situational variables have the potential to influence motivation, including involvement's ability to influence elaboration and thereby influence the research model. The authors also recommend consideration of the relevant conditional influences impacting such models.
The final puzzle that we identify concerns how motivation operates in models of the formation of loyalty intentions/behaviors. Perugini and Bagozzi (2004) make a strong argument differentiating motivation as desire from intentions. We suspect that there is more to learn about the various unique and synergistic roles of desires, PBC, and/or intentions in terms of motivational content.
The current research also offers insight for managers who want to positively influence the customers' intentions to be loyal. While word-ofmouth has been identified as the important measure of customer loyalty/satisfaction (Reichheld 2003), it is also useful for managers to understand those factors impacting the customer's intention to want to use word-of-mouth to promote a specific manufacturer's market offering over that of a competitor. As predicted, we find that word-of-mouth intentions were influenced by loyalty fortitude intentions, which in turn, are influenced by loyalty repurchase intentions and the customer's desire to be loyal. Furthermore, a customer's loyalty repurchase intentions are influenced by the customer's desire to be loyal as well as the difficulty the customer experiences when trying to be loyal. Finally, the customer's desire to be loyal is influenced by others significant to the customer (subjective norm) the anticipated positive emotion associated with being loyal, as well as hedonic utilitarian and hedonic attitudes towards being loyal. As such, the organization must focus on factors that impact the customer's utility that they receive in the exchange, while simultaneously heightening the experience of customers so that the customer developed positive feelings toward being loyal to the organization. Finally, it may be beneficial to the organization to direct marketing actions toward the totality of targeted customers. By enhancing the image of the firm to all potential customers, word of mouth activities might enhance loyalty intentions through the identified influence of subjective norms. This effect should occur through influencing their desire to be loyal, and desire's impact on the formulation of repurchase and fortitude loyalty intentions. In much the same way, organizations that can enhance the customer's positive anticipated emotions about being loyal to the organization and hedonic attitude should also experience customers with higher levels of word of mouth intentions. Managers would need to focus on ensuring the customer has enjoyable, exciting, delightful exchanges when re-purchasing the market offering. Such exchanges should highlight to the customer the benefits of increasing loyalty to the organization.
Managerially, the results first suggest that at least some of the discussion related to the appropriate conceptualization and measurement of customer loyalty found in the consumer literature can be generalized to B2B settings. second, the results also generally support the recent arguments of Morgan and Rego (2006) calling into question the efficacy of Reichheld's (2003) behavioralist arguments. Third, the results suggest caution in using satisfaction measurements. If loyalty is a goal as is argued herein, then it remains unclear whether satisfaction is an anticipate emotion as envisioned by Bagozzi et al. (2003), or an outcome of a service interaction as traditionally envisioned, or perhaps even both. It appears that much remains to be learned about how to measure and incorporate satisfaction in managerial decision making. Fourth, if managerial perceptions of control are indeed related to negative affect, as found herein, then methods appear warranted to minimize such negative emotional associations with control in support of long-term marketing relationships. Finally, consistent with the general marketing literature, firms offering service products such as PLI insurance may best be served by seeking to foster fortitude-related forms of loyalty in support of relationship marketing initiatives.
1 Most variables were in the range of 2-8%.
2 McQuitty (2004) goes on to assert that recommendations for minimum sample sizes (in excess of 100) or 5-10 times the number of variables or estimated parameters may be outdated.
Agustin, Clara and Jagdip Singh (2005), "Curvilinear Effects of Consumer Loyalty Determinants in Relational Exchange," Journal of Marketing Research, 42 (February), 96-108.
Anderson, C. J. (2003), "The Psychology of Doing Nothing: Forms of Decision Avoidance Result from Reason and Emotion," Psychological Bulletin, 129(1), 139-167.
Andersen, James C. and David W. Gerbing (1988), "Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach," Psychological Bulletin, 103 (3), 411-423.
Bagozzi, R. P., Baumgartner, H., and Pieters, R. (1998), "Goal-directed Emotions," Cognition and Emotion, 12, 1-26.
Bagozzi, R. P.,Z. Gurhan-Canli, and J. R. Priester (2002), The Social Psychology of Consumer Behavior, Buckingham: Open University Press.
Bettman, James R., Mary Frances Luce, and John W. Payne (1998) "Constructive Consumer Choice Processes," Journal of Consumer Research, 25 (December), 187-217.
Buck, R., E. Anderson, A. Chaudhuri, and I. Ray (2004), "Emotion and Reason in Persuasion: Applying the ARI Model and the CASC Scale," Journal of Business Research, 57, 647-656.
Chaudhuri, A. (2006), Emotion and Reason in Consumer Behavior, Oxford: Elseveir, Butterworth-Heinemann.
Curran, Patrick, Kenneth A. Bollen, Feinian Chen, Pamela Paxton, and James B. Kirby (2003), "Finite Sampling Properties of the Point Estimates and Confidence Intervals of the RMSEA," Sociological Methods & Research, 32 (2, November), 208-52.
Dillman, Don A. (2007), Mail and Internet Surveys: The Tailored Design Method, 2nd Edition. New York: John Wiley & Sons, Inc.
Eagly, A. H. and S. Chaiken (1993), The Psychology of Attitudes, Fort Worth: Harcourt Brace Jovanovich College Publishers.
East, Robert, Philip Gendall, Kathy Hammond, and Wendy Lomax (2005), "Consumer Loyalty: Singular, Additive, or Interactive?" Australasian Marketing Journal, 13 (2), 10-26.
Evans, K. R., Christiansen, T., and Gill, J. D. (1996), "The Impact of Social Influence and Role Expectations on Shopping Center Patronage Intentions," Journal of the Academy of Marketing Science, 24 (3), 208-218.
Evanschitzky, Heiner and Maren Wunderlich (2006), "An Examination of Moderator Effects in the Four-Stage Loyalty Model," Journal of Service Research, VOL 6, 1-16.
Fletcher, Greg (2003), "Sarbanes-Oxley Leaves Heavy Footprints on Corporate America," AFP Exchange, 23 (4), 56-61.
Fullerton, Gordon (2003), "When Does Commitment Lead to Loyalty?" Journal of Service Research, 5 (4), 333-344.
Garbarino, Ellen C. and Julie A. Edell (1997), "Cognitive Effort, Affect, and Choice," Journal of Consumer Research, 24 (September), 147-158.
Hagger, M. S. and N. L. D. Chatzisarantis (2005), "First- and Higher-Order Models of Attitudes, Normative Influences, and Perceived Behavioral Control in the Theory of Planned Behavior," British Journal of Social Psychology, 44, 513-535.
Hair, J. F. Jr., R. E. Anderson, R. L. Tatham, and W. C. Black (1998), Multivariate Data Analysis, 5th Edition. Upper Saddle River, NJ: Prentice Hall.
Hsee, C. K., J. Zhang, & J. Chen (2004), "Internal and Substantive Inconsistencies in Decision Making," in D. J. Koehler & N. Harvy (Eds), Blackwell Handbook of Judgment & Decision Making, Oxford: Blackwell Publishing, 360-378.
Jacoby, Jacob and Robert W. Chestnut (1978), Brand Loyalty: Measurement and Management, New York: Wiley.
Johnson, Michael D., Andreas Hermann, and Frank Huber (2006), "The Evolution of Loyalty Intentions," Journal of Marketing, 70 (April 2005), 122-132.
Kahneman, D. & A. Tversky (2000), "Prospect Theory: An Analysis of Decision Under Risk" in D. Kahneman and A. Tversky (Eds), Choices, Values, and Frames, Cambridge: Cambridge University Press, 17-43.
Loewenstein, G., Elke, F. Weber, C. K. Hsee, and N. Welch (2001), "Risk as Feelings," Psychological Bulletin, 127 (2), 267-286.
MacDonald, G. and P. R. Nail (2005), "Attitude Change and the Public-Private Attitude Distinction," British Journal of Social Psychology, 44, 15-28.
Mattila, Anna S. (2006), "How Affective Commitment Boosts Guest Loyalty (and promotes Frequent Guest Programs)," Cornell Hotel and Restaurant Administration Quarterly, 47 (2), 174-181.
McQuitty, S. (2004), "Statistical Power and Structural Equation Models in Business Research," Journal of Business Research, 57, 175-183.
Mick, D. G., Broniarczyk, S. M., and Haidt, J. (2004), "Choose, Choose, Choose, Choose, Choose, Choose, Choose: Emerging and Prospective Research on the Deleterious Effects of Living in Consumer Hyperchoice," Journal of Business Ethics, 52 (June), 207-211.
Mittal, Vikas, Pamkaj Kumar, and Michael Tsiros (1999), "Attribute-Level Performance: Satisfaction and Behavioral Intentions Over Time: A Consumption System Approach," Journal of Marketing, 63 (April), 88-101.
Morgan, Neil A., and Lupo Leotte Rego (2006), "The Value of Different Customer Satisfaction and Loyalty Metrics in predicting Business Performance," Marketing Science, 25 (5), 426-39.
Norenzayan, A. and S. J. Heine (2005), "Psychological Universals: What are They and How Can We Know?," Psychological Bulletin, 131 (5), 763-784.
Okada, E. M. (2005), "Justification Effects on Consumer Choice of Hedonic and Utilitarian Goods," Journal of Marketing Research, Vol. XLII (February), 43-53.
Oliver, Richard L. (1997), Satisfaction: A Behavioral Perspective on the Consumer, New York: The McGraw-Hill Companies, Inc.
____ (1999), "Whence Consumer Loyalty?," Journal of Marketing, 63 (Special Issue 1999), 33-44.
Olsen, Svein Ottar, James Wilcox, and UIf Olsson (2005), "Consequences of Ambivalence on Satisfaction and Loyalty," Psychology and Marketing, 22 (3), 247-269.
Perugini, M. (2005), "Predictive models of implicit and explicit attitudes," British Journal of Social Psychology, 44,29-45.
_____ and R. P. Bagozzi (2001), 'The Role of Desires and Anticipated Emotions in Goal-Directed Behaviors: Broadening and Deepening the Theory of Planned Behaviors," British Journal of Social Psychology, 40,79-98.
_____ and _____ (2004), "The Distinction Between Desires and Emotions," European Journal of Social Psychology, 34, 69-84.
Raines-Eudy, R. (2000), "Teacher's Corner: Using Structural Equation Modeling to Test for Differential Reliability and Validity - An Empirical Demonstration," Structural Equation Modeling, 7 (1), 124-41.
Ratner, Rebecca K. and Kenneth C. Herbst (2005), "When Good Decisions Have Bad Outcomes: The Impact of Affect on Switching Behavior," Organizational Behavior and Human Decision Processes, 96, 23-37.
Reichheld, Frederick F. (2003), "The One Number You Need to Grow," Harvard Business Review, December 2003, 46-54.
Rhodes, R. E. and K. S. Courneya (2003a), "Investigating Multiple Components of Attitude, Subjective Norm, and Perceived Control: An Examination of the Theory of Planned Behavior in the Exercise Domain," British Journal of Social Psychology, 42, 129-146.
_____ & _____ (2003b), "Self-Efficacy, Controllability, and Intention in the Theory of Planned Behavior: Measurement Redundancy or Causal Independence?," Psychology and Health, 18 (1), 79-91.
Rust, Roland T., Katherine N. Lemon, and Das Narayandas (2005), Customer Equity Management, Upper Saddle River, NJ: Pearson/Prentice Hall.
Soderlund, Magnus (2006), "Measuring Customer Loyalty with Multi-Item Scales," International Journal of Service Industry Management, 17, 76-98.
Trafimow, D., P. Sheeran, M. Conner, and K. A. Finlay (2002), "Evidence That Perceived Behavioral Control is a Multidimensional Construct: Perceived Control and Perceived Difficulty," British Journal of Social Psychology, 41,101-121.
Tsang, E. W. K. and K. Kwan (1999), "Replication and Theory Development in Organizational Science: A Critical Realist Perspective," Academy of Management Review, 24 (4), 759-780.
Uncles, Mark D., Grahame R. Dowling and Kathy Hammond (2003), "Customer Loyalty and Customer Loyalty Programs," Journal of Consumer Marketing, 20 (4), 294-316.
Van Den Berg, H., A. S. R. Manstead, and J. van der Plight (2005), "The Role of Affect in Attitudes Toward Organ Donation and Donor-Relevant Decisions," Psychology and Health, 20 (6), 789-802.
Vavra, Terry G. (1992), After-Marketing: How to Keep Customers for Life Through Relationship Marketing, New York: Irwin Professional Publishing.
Voss, K. E., E. R. Spangenberg, and B. Grohmann (2003), "Measuring the Hedonic and Utilitarian Dimensions of Consumer Attitudes," Journal of Marketing Research, Vol. XL (August), 310-320.
Yu, Y. and A. Dean (2001), "The Contribution of Emotional Satisfaction to Consumer Loyalty," International Journal of Service Industry Management, 12 (3), 234-250.