After tuning (adapting) their message to an audience, communicators’ memory is often evaluatively biased towards the audience’s judgment about the topic. Imagine that team member A tunes her message about a newcomer’s behaviors toward the judgment of member B (the audience). The audience-tuning memory bias emerges when A’s later recall of the newcomer’s behavior is evaluatively biased towards the audience’s judgment. In the first funding period, we examined cognitive processes anchored in communicators’ motivation to create a shared reality with the audience about a target person. Consistent with a motivational account of accessibility (Eitam & Higgins 2010), participants responded more quickly to audience-congruent (vs. incongruent) trait representations in a new reaction-time task. The findings support the notion that the memory bias results from the enrichment of episodic memory traces with semantic trait information (Cheng & Werning, 2016). The accessibility bias also correlated with reported shared reality experience. Thus, spontaneous accessibility of information is an important facet of the memory effect, reflecting shared reality with the audience.

The present project investigates, for the first time, the role of communicators’ own, audience-independent judgments of the topic. In Part I, we will manipulate communicators’ epistemic confidence in their initial judgments, which should restrict potential shared reality and hence the audience-tuning memory bias. The manipulations range from individual-level processes (mere thinking and elaboration, Work Package/WP1) via dyadic processes (disclosure of one’s judgment to a prior audience, WP2) to social verification by a group of others with high (vs. low) epistemic authority (WP3), gradually approximating real-life communication. Because we found that free recall and cognitive accessibility interfere with one another, we will assess the two measures in separate studies. To enhance conceptual precision, we will also model the proposed processes in collaboration with P5.

Part II (WPs 4 & 5) extends our approach to memory for one’s own behaviors, building on the notion that self-related information is chronically accessible and highly relevant to people’s needs and motivations (Dings & Newen, 2021). When participants have high (vs. low) confidence in their own prior judgments, shared reality and audience-congruent memory biases are less likely, especially when need for self-protection is high. Conversely, when communicators have low confidence in their own judgments, shared reality and corresponding memory biases are more likely to emerge (also for negative audience evaluations or expectancy-inconsistent behaviors), especially when need for self-protection is low. The findings will have important implications, for instance, for eyewitness memory, the persistence of biased beliefs shared in social networks, and relations between social groups with conflicting memories of the past.