Academic research has suggested that social media can improve corporate reputation, sales, return on investment, positive word-of-mouth and create value for consumers. Playing the central role in this relational process, consumer or customer engagement (used interchangeably in this context), emerged as one of the main constructs explaining the success and failure of brands on social media. Considering engagement as a multidimensional construct composed of cognitive, emotional and behavioral dimensions that varies according to the subject (e.g., customer or employee), focal object (e.g., brand, product or advertising), and context (e.g., retail or online), in this dissertation we make a specific field cutting in order to discuss consumer engagement (subject) with brands (object in Article 2 and 3) or brand communities (object in Article 1) on social media (context). As a general aim of the dissertation, we analyze consumer engagement in different nomological networks, prioritizing its antecedents and mediating role (Article 1) or consequences and its relationship with another measure of engagement (Article 3), in addition to perform a comparative analysis of different scales developed to measure it (Article 2). In the Article 1, we show that the Vivek, Beatty, Dalela, and Morgan (2014) consumer engagement scale, originally developed for application in multiple contexts, can also be applied to multiple objects. We validate this scale in a nomological network with perceived homophily and brand involvement as antecedents and corporate reputation as consequence. Our results show that besides having a direct positive effect on corporate reputation, consumer engagement has a strong mediating role. In the Article 2, our goal is to compare some of the main scales developed to measure consumer engagement: Dessart, Veloutsou, and Morgan-Thomas (2016), Hollebeek, Glynn, and Brodie (2014) and Vivek et al. (2014). To do this, we propose a procedure for the comparative analysis of scales that involves i) the selection, ii) standardization and iii) analysis of measures. We replicate three selected scales with the same object (brand) and context (social media), considering that most of the developed scales were applied to the same object and context. Based on the Classical Test Theory (CTT) and Item Response Theory (IRT) analysis, our results show similarities, differences, and advantages of applying each scale. In Article 3 we jointly analyze two streams of consumer engagement studies by considering brand engagement in self-concept (BESC) as a measure of engagement with brands in general (Sprott, Czellar, & Spangenberg, 2009) and consumer engagement with a specific brand Vivek et al., 2014), and how the latter impacts the perceived value of consumers and corporate reputation. By means of the multigroup analysis with samplings carried out in Brazil and the United States, our results showed that in addition to excellent adjustment, both at the model level and at the path level there was no difference between the groups, which indicates the cross-cultural invariance in the structural model. Finally, in Chapter 5 we synthesize the conclusions of the dissertation.