LAUSR

Abstract Previous research demonstrated that mere instructions to approach one stimulus and avoid another stimulus result in an implicit preference for the to-be-approached over the to-be-avoided stimulus. To investigate the mechanisms underlying approach-avoidance (AA) instruction effects, we tested predictions of a propositional account and an associative self-anchoring account in a preregistered adversarial collaboration. Consistent with the propositional account, Experiment 1 showed that avoidance instructions had a negative effect on implicit evaluations over and above the positive effect of approach instructions. Consistent with the associative self-anchoring account, Experiment 2 showed that changes in implicit self-stimulus linking mediated AA instruction effects on implicit evaluations. However, mediation was only partial, in that AA instructions showed a significant effect on implicit evaluations after controlling for implicit self-stimulus linking. Together, the results support the contribution of propositional processes to AA instruction effects; the results remain ambiguous regarding an additional contribution of associative self-anchoring.