The self-reference effect (SRE) refers to a memory advantage for information processed in relation to the self. Self-referential encoding seems to especially enhance recollection, the subjective “feeling of remembering” and retrieval of episodic detail. Improved memory, and a correspondingly smaller SRE, are also typically observed after reference to a close, compared to a distant other person. We investigated whether the SRE and the modulating effect of social distance would occur regardless of factorial design, that is, whether encoding was manipulated within or between subjects. Participants judged the descriptiveness of trait words regarding either themselves, someone close to them, or a distant acquaintance, followed by a recognition memory test and Remember/Know/Guess-judgments. In the within-subjects design (N=152), a third of trials was assigned to each reference task (self, close or distant other) in a mixed block. In the between-subjects design (N=155), the reference task always remained the same, with each participant only performing one type of reference. We replicated the SRE in both design types: Recognition performance and recollection estimates were higher after self- than either close or distant other-reference. However, we only observed a memory advantage for close- over distant other-reference when encoding was manipulated within participants, with no effect of social distance in the between-subjects design. We consider potential roles for sampling error and context effects on social cognition in producing these differing patterns. Our findings suggest that practical choices such as the type of experimental design may have important implications for the SRE.