Social-cognitive perspectives suggest that ingroup perceived similarity between a message recipient and communicator stimulates systematic (e.g., more effortful) information processing of the message, which is posited to occur through implicitly-conferred ingroup benefits (e.g., expecting/perceiving greater ease of processing messages with ingroup members). Yet, effects of perceived similarity on information processing have not been empirically tested in the context of learning, nor have their presumed underlying processes. I proposed a conceptual model postulating that a salient shared social identity will activate greater perceived similarity between the processor of information and authors of the respective information which, in turn, will enhance information processing via cognitive and motivational mechanisms. Specifically, I tested whether a shared social identity enhances learning serially via a) perceived similarity, b) cognitive synergy between the processor and authors (perceptions encompassing ease of processing and cognitive connection with one’s partner), and c) effort. College participants completed sample reading comprehension sections of the Graduate Record Examinations (GRE). Passages for each reading comprehension task were purported to be written by “experts” who have mastery of an esoteric topic. Experts were perceived to be either similar or dissimilar to the participant based on information processing (learning) styles. After making shared social identity salient, participants rated their perceptions (presumed cognitive and motivational mechanisms) vis-à-vis the expert and task both before and after the reading comprehension tasks. Results failed to support a direct relationship between shared social identity and learning. But, supporting the hypothesized mediational sequence, a shared social identity predicted greater perceived similarity which, in turn, predicted greater learning serially via perceived (i.e., post-assessment) cognitive synergy and effort. Interestingly, inconsistent mediation manifested with pre-assessment mediators in the model, namely that the effect of shared social identity on learning was accentuated via perceived similarity but also inhibited via expected cognitive synergy. Findings have both theoretical and practical implications for understanding the complexities of a salient shared social identity on how people process information.