Because of its simplicity, the EC procedure has been regularly used for testing the automaticity of attitude formation. By “automaticity”, we refer to four operating conditions under which one might expect EC effects to be observed: (1) unawareness (of the CS-US pairings), (2) efficiency at encoding, (3) uncontrollability and (4) goal-independence. In this project, we focus on the last one: goal-independence. We are interested in whether EC effects are sensitive to the activation of processing goals prior to attitude acquisition, and if so why. Two papers directly addressed the impact of processing goals on EC: - Gast and Rothermund (2011) observed that focusing on the valence of the stimuli rather than on an evaluatively irrelevant dimension is necessary for EC to be observed. The focus on valence increased the explicit memory of CS-US pairings but it did not mediate the effect of processing goals on EC effects. The same effects were observed when they used indirect evaluative measure. - Corneille, Yzerbyt, Pleyers, and Mussweiler (2009) observed that activating a similarity versus difference processing goal prior to learning increases EC effects on a direct evaluative measure, while leaving unaffected the explicit memory of CS-US pairings. The present research aims at advancing our understanding of the similarity-difference effect. Processing goals may have impacted (1) the expression rather than the formation of attitudes, and (2) the implicit memory of the pairings rather than the explicit memory of the pairings (for relevant papers, see Hütter, Sweldens, Stahl, Unkelbach & Klauer, 2012; Hütter & Sweldens, 2013, Hütter & De Houwer, 2017; Mierop, Hütter & Corneille, 2017; Mierop, Hütter, Stahl, & Corneille, 2018). In a preregistered experiment, we test these two hypotheses by adding to the paradigm of Corneille and colleagues (2009): (1) an indirect evaluative measure that minimizes deliberate CS evaluations and (2) a multinomial processing tree (MPT) model that allows isolating and weighing the respective contribution of explicit and implicit memory processes in EC (Hütter et al., 2012).