Supplemental Material for How People (Fail to) Control the Influence of Affective Stimuli on Attitudes

This prediction was generally confirmed, although no significant changes were observed in the invalid condition of Experiment 3 with posterior validity information (applying one-tailed tests).The prediction is not discussed in detail in the main article, but the result is presented. Efficiency of Learning with Anterior vs. Posterior Validity Information(2) "Standard and reversed EC effects should be larger in the advanced-instructions condition than in the postponed-instructions condition."This hypothesis was formulated in anticipation of differences in the efficiency of learning in the two versions of the paradigm.As the posterior condition separates the US valence from the validity information in time, the joint encoding of CS, US valence, and validity information should be impaired.Additionally, the integration and interpretation at the judgment stage might take longer in the posterior condition.Due to this ambiguity, this hypothesis is not discussed in the main article, although the prediction was confirmed in both experiments as reported in the results sections.(3) "The c-parameter will be smaller in the postponed-instructions condition than in the advanced-instructions condition (both at Time 1 and Time 2)."The c-parameter reflects an evaluative judgment in line with the stated validity of a pairing.Therefore, the same prediction holds as for the evaluative ratings.As is reported in the main article, this hypothesis was generally confirmed with the exception that the difference between c-parameters did not reach significance at Time 2 in Experiment 3.(4) "The effect of the time interval on evaluative ratings should be stronger in the postponedinstructions condition than in the advanced-instructions condition."