*3.2. Proposed Model and Testing*

Regarding the first construct—brand relationships—we found that it is formed by three dimensions: trust, commitment, and motivation. Trust was adapted from existing scales in the literature, but motivation and commitment (although based on the concepts of Hardwick and Ford (1986) and Wilson (1995)) were developed in this research, by using confirmatory factor analysis (CFA). The scales used to define the brand relationships construct were found to be valid and reliable.

To test the structural model, we used corporate brand identity (external part), which was developed in a previous study. The items used to characterize the physical dimension, the relation and the reflected consumer dimension were the result of previous research pursued by Barros et al. (2016). The authors used the external part of the brand identity prism to argue that the relationships among brands (brand relationships) influence the external part of corporate brand identity and reputation.

The result of a well-managed corporate brand identity is a positive reputation. Therefore, brand reputation is the expected result of an active corporate brand identity management under a relational approach. It is widely suggested in the literature that identity precedes reputation (Burmann et al. 2009; de Chernatony 1999; Kapferer 1986, 2008). Corporate brands should actively choose and select recognized brands with which to develop relationships, to bridge the gap between brand identity and reputation. The result of this management should be an increase in brand reputation.

We also used the reputation concept unidimensionality, developed by Vidaver-Cohen (2007), to connect with this research. Data were analyzed using CFA and structural equation modeling (SEM). A structural equation model was developed to test the brand relationships concept as an antecedent to corporate brand identity and reputation. According to Nachtigall et al. (2001), SEM represents the relationship between latent variables (brand relationships, corporate brand identity, and brand reputation in our model) and their manifest or observable indicators (the items that characterize the latent variables). The most prominent feature of SEM is the capability to deal with latent variables. These variables are connected to observable ones by a measurement model (Edwards and Bagozzi 2000).
