Ballot position effects have been documented across a variety of political and electoral systems. In general, knowledge of the underlying mechanisms is limited. There is also little research on such effects in preferential-list PR systems, in which parties typically present ranked lists and thus signaling is important. This study addresses both gaps. Theoretically, we formalize four models of voter decision-making: pure appeal-based utility maximization, implying no position effects; rank-taking, where voters take cues from ballot position per se; satisficing, where choice is a function of appeal, but voters consider the options in the order of their appearance; and a hybrid “satisficing-with-rank-taking” variant. From these, we derive differential observable implications. Empirically, we exploit a quasi-experiment, created by the mixed-member electoral system that is used in the state of Bavaria, Germany. Particular electoral rules induce variation in both the observed rank and the set of competitors, and allow for estimating effects at all ranks. We find clear evidence for substantial position effects, which are strongest near the top, but discernible even for the 15th list position. In addition, a candidate’s vote increases when the average appeal of higher-placed (but not that of lower-placed) competitors is lower. Overall, the evidence is most compatible with the hybrid satisficing-with-rank-taking model. Ballot position thus affects both judgment and choice of candidates.