Self-reference effect

The self-reference effect is the finding that people remember information better when they relate it to themselves than when they process it in other ways. A person who decides whether the word "generous" describes them, for example, is later more likely to recall that word than a person who decides whether it means the same as "kind".^[1:1] The effect is one of the most replicated results in the study of memory and the self-concept, and it is often used as a tool for studying how the brain represents the self.

A meta-analysis of the research literature found that self-referent encoding beat both semantic encoding (thinking about a word's meaning) and other-referent encoding (relating the word to another person).^[2:1] Researchers still disagree about why the effect occurs, with elaboration and organization of information as the leading explanations.^[3:1]

Original experiments

The effect was reported in 1977 by Timothy Rogers, Nicholas Kuiper, and William Kirker. Participants rated lists of adjectives while performing one of four tasks designed to force different kinds of processing: a structural task (about the appearance of the word), a phonemic task (about how it sounds), a semantic task (about its meaning), and a self-reference task (whether it describes the participant). Participants were not told that their memory would be tested. In two experiments, they later recalled the most words from the self-reference task.^[1:1]

The design compares self-reference with the processing levels in the levels-of-processing framework, in which meaning-based encoding normally beats surface-level encoding. Rogers and colleagues concluded that self-reference is a rich encoding process that acts like a superordinate schema in the processing and memory of personal information.^[1]

Size and robustness

A 1997 meta-analysis by Cynthia Symons and Blair Johnson confirmed that self-referent encoding yields better memory than semantic encoding and than other-referent encoding.^[2:1] A later paper describes that analysis as covering 129 studies.^[3:1]

The effect also holds outside the laboratory setting of the original work. In four online studies with 658 participants, Sarah Bentley, Katharine Greenaway, and Alexander Haslam found that words encoded in relation to the self were recalled more often than words encoded semantically (a standardized difference of d = 0.63), which is comparable to laboratory results.^[3:1]

The size of the advantage depends on the comparison. It is smaller against an intimate other, such as a close friend, than against a non-close other, and the meta-analysis concluded that closeness, not mere familiarity, influences the effect.^[4]

Explanations

Two explanations have dominated the literature.

• Elaboration. Because people know a great deal about themselves, relating a word to the self links it to many existing memories and facts, which gives more routes for retrieving it later.
• Organization. The self may work as a well-structured set of knowledge into which new items fit, so that items are stored and retrieved together.

Symons and Johnson attributed the effect mainly to the self being a well-developed and frequently used construct that promotes both elaboration and organization.^[2:1] Stanley Klein and Judith Loftus had earlier examined the separate contributions of elaborative and organizational processing to self-referent encoding.^[5:1] More recent work still describes the relative roles of elaboration and schematic organization as an open debate.^[3:1]

Neural basis

Brain imaging and lesion studies point to the medial prefrontal cortex (mPFC), a region along the midline of the frontal lobes.

In a 2002 functional magnetic resonance imaging study, William Kelley and colleagues had participants judge trait adjectives for self-relevance, for relevance to another person, or for letter case. A region of medial prefrontal cortex was selectively engaged by the self-relevance judgments.^[6:1] A follow-up study found that this region responded to judgments about the self but did not respond comparably to judgments about a close friend, suggesting that the brain keeps the self distinct from even intimate others.^[4:1]

Imaging shows only that a region is active during a task. A stronger test came from patients with focal mPFC damage: in a trait judgment task, the damage abolished the self-reference effect, indicating that this region is necessary for it.^[7:1]

Open questions

The elaboration and organization accounts make different predictions about which materials and populations should show the largest advantage, and the literature has not settled between them.^[3:1] The studies cited here mostly use trait adjectives as stimuli, so how far the findings extend to other material, such as facts, images, or study content, is a question this article does not resolve. Settling it would require experiments that manipulate elaboration and organization independently while holding the self-relevance of the material constant.