In the Attentional Boost Effect (ABE), words or images encoded with to-be-detected target squares are later recognized better than words or images encoded with to-be-ignored distractor squares. The present study sought to determine whether the ABE enhanced the encoding of the item-specific and relational properties of the studied words by using the multiple recall paradigm. Previous evidence indicates that manipulations fostering item-specific encoding increased the number of item gains, whereas manipulations fostering relational encoding decreased item losses. Across three experiments, participants were presented with lists of semantically related or unrelated words paired with target (red) or distractor (green) squares, under the instructions to remember all the words and press the spacebar when the square was red. Immediately after the study phase, they were involved in four consecutive recall attempts. In all cases, the classical ABE was replicated, in that participants recalled more target- than distractor-paired words. Most importantly, the analyses converged in showing that item gains were significantly greater for target- than for distractor-paired words when participants studied lists of related words (but not when they studied unrelated lists); in contrast, item losses did not differ between the two types of words, irrespective of the nature of the studied list. Taken together, these data suggest that the ABE enhanced the encoding of item-specific information but had no effect on the encoding of relational information.