“Me” means more than “Good”: Stimuli’s self-relevance matters more than valence in shaping evaluative learning via the self.

Stimuli that relate to the self tend to be better liked. The Self-Referencing (SR) task is a paradigm whereby one target categorised through the same action as self-stimuli (i.e. possessive pronouns) is preferred over an alternative target categorised through the same action as other-stimuli. Past studies on the SR showed that valence could not fully account for the observed effect. Here we explored self-relevance as a possible explanation. Across four studies (N = 567), participants selected self-relevant and self-irrelevant adjectives to be used as source stimuli in a Personal-SR task. In that task, the two classes of stimuli were paired with two fictitious brands. We measured automatic (IAT) and self-reported preferences, and identification with the brands. Experiment 1 showed that the brand paired with positive self-relevant adjectives became more positive than the one paired with positive self-irrelevant adjectives. Experiment 2 confirmed this pattern with negative adjectives, and Experiment 3 ruled out the effect of a self-serving bias in the adjectives selection. Experiment 4 showed that the brand related to negative self-relevant adjectives was preferred over the brand related to positive self-irrelevant adjectives. We discussed the implications of our results and the potential mechanisms that might explain self-driven preferences.