Although the judgmental consequences of anchoring have been frequently demonstrated, the processes underlying these effects are still insufficiently understood. Results of three studies support the notion that anchoring (defined as the assimilation of an absolute judgment toward a previous standard of comparison) is a special case of semantic priming in that the information that is activated to solve the comparative anchoring task will subsequently be more accessible when the absolute judgment is formed. Assimilation toward the anchor occurs when the comparative judgment task elicits a "positive test strategy", in which judges test the possibility that the anchor value characterizes the target. Anchor-consistent information will consequently be activated, and be more accessible when absolute judgments are generated. This hypothesis was tested in three studies using the logic of priming research. In the first experiment, we showed that the strength of the assimilation effect depended on the applicability or relevance of the activated information. In the second experiment, a contrast effect was obtained when the targets of the two judgment tasks were sufficiently different. In the third study, it was demonstrated that the comparative judgments required less and the absolute judgments more response time when the anchor values were entirely implausible, presumably because implausible standards of comparison do not require relevant target information to be activated. Subsequently, however, because relevant information is less accessible, absolute judgments will be more difficult and will require more response time. In sum, the results are consistent with mechanisms of semantic priming but are difficult to explain by invoking mechanisms of a numerical priming that are based on the numerical anchor value as the only determinant of the assimilation effect.