The robustness of implicit evaluation updating: Tests of time course and resistance to the return of negativity

Accumulating evidence suggests that even highly negative implicit (automatic) social evaluations are amenable to temporary shifts toward positivity; however, demonstrations of long-term change, especially in the negative-to-positive direction, are rare. Here we report 10 experiments (9 preregistered; total N = 3,103) probing whether negative-to-positive reversals in implicit evaluations can persist over time and are robust to reminders of initial negativity. To this end, we first identified two procedures involving diagnostic behavioral information (reinterpretation and negation + replacement) that, when measured immediately following the intervention, overturned experimentally created implicit negativity toward novel social targets, as measured by the Affect Misattribution Procedure (AMP; Exp. 1–2). Attesting to the durability of this effect, negative-to-positive reversals persisted without any decrement over a 2-day delay (Exp. 3). Although exposure to unrelated negative information following a 2-day delay caused implicit evaluations to return to negativity in the absolute sense (Exp. 4–5), the induced negativity was no stronger than negativity that emerged following univalent positive initial learning (Exp. 6A–6C) or in the absence of any prior learning (Exp. 7). Demonstrating the flexibility of automatic evaluative processing, these data suggest that negative-to-positive implicit evaluation change can be both durable and relatively robust to reminders of initial negativity. We discuss implications for social cognitive theory and the prospect of durably changing well-established implicit evaluations of known targets.