Assessing implicit alcohol associations with the implicit association test: fact or artifact?

Studies using bipolar Implicit Association Tests (IATs) found that heavy drinkers have negative and arousal associations with alcohol relative to soda. Study I examined whether these results were due to the label 'alcohol' and the choice of the contrast category 'soda'. Four unipolar IATs assessed alcohol associations with positive and negative valence, arousal, and sedation, while varying the target dimension: alcohol or beer versus soda or animals. Results showed that drinkers had the strongest associations between alcohol and negative valence with the exact strength depending on the choice of the target categories. They also showed associations between alcohol and positive valence, arousal, and to a lesser extent sedation, which were uninfluenced by composition of the target dimension. These findings indicate ambivalence in both the valence and arousal-sedation dimension, underscoring the importance of using unipolar alcohol-IATs. Further, study 2 showed that "figure-ground7 asymmetries could not account for these IAT results. These findings provide support that implicit alcohol associations are not merely IAT artifacts and that they can be assessed in a meaningful way with unipolar IATs.