Consumer Trust and Satisfaction in the Formation of Consumer Loyalty Intentions in Transactional Exchange: The Case of a Mass Discount Retailer

Abstract

AbstractMarketing relationships can be placed on a continuum from short, discrete transactions to ongoing brand relationships. The majority of recent work has focused on relational exchanges, with some scholars even suggesting early on that the marketing discipline was undergoing a paradigm shift from a transaction-based marketing perspective toward a relational exchange perspective. However, there has been a growing recognition that not all customers seek relational exchanges. Consequently, the current research considers customer relationship management from the less studied, but oft seen, perspective of transactional exchange. A study is presented using recent advances in structural equation modeling analyses, including Bayesian estimation methods and mediation analyses. We further consider the psychological processes underlying the formation of consumer loyalty based on pre- and post-purchase measurements taken over multiple time periods. We specifically hypothesize that consumer satisfaction judgments will fully mediate any influences of post-purchase trust judgments on future loyalty intentions. With American consumers' trust in businesses at an all-time low, coupled with the recent trend that more and more brick-and-mortar retailers are at risk of "showrooming" for online retailers, there is an apparent need to also consider retail customers who see the value of relationship marketing only selectively.KEYWORDS: consumer relationshipsconsumer trustloyaltysatisfaction Notes1Prior beliefs can be noninformative or informative. Noninformative prior beliefs are a data-based function that contains ambiguous information about a parameter, whereas informative prior beliefs are a full generative function that contains specific information about a parameter. In the current study, informative priors are differentiated from noninformative priors by applying a small-variance prior of 0.01 that will allow the cross-loading variation to range between –.2 and.2.2Muller et al. (Citation2005) asserted that mediated moderation, or the occurrence of the magnitude of the overall treatment effect on the outcome depending on the moderator, can happen only when moderation occurs. Moderated mediation, alternatively, happens when the mediating process that is responsible for producing the effect of the treatment on the outcome depends on the value of the moderator variable. Readers will note that this definition does not imply any overall moderation of the treatment effect in our accepted model. What varies as a function of the moderator is not the magnitude of the overall treatment effect on the outcome, but the mediating process that produces it.3We further considered the potential threat of multilevel biasing of reported results based on the arguments of Preacher et al. (2010, 2011) and Muthén (2011). The gist of these arguments is that traditional methods for assessing mediation in the presence of multilevel phenomenon are inappropriate, primarily because the assumption of independence of observations is violated when clustered data are used. Specifically, traditional methods fail to appreciate differences surrounding between- and within-cluster components of variance. In such cases, Preacher et al. (2010) specifically recommended the use of Muthén and Asparouho's (2008) multilevel SEM framework. Fortunately, we can identify no reason to assume that multilevel considerations pose any threat in the current analyses given the research setting and population of interest.