People consistently over-eat when served a large compared with a small (appropriate) portion of food. However, the mechanism underlying this so-called portion size effect is not well understood. We argue that the process of anchoring and adjustment naturally describes this effect, such that the size of a presented portion works as an anchor that strongly influences consumption. The classical anchoring and adjustment paradigm was applied to six hypothetical eating situations. Participants were asked to imagine being served either a small or a large portion of food (i.e., low and high anchor) and to indicate whether they would consume more or less than this amount. Then, they indicated how much they would eat. These estimates were compared with a no-anchor condition where participants did not imagine a specific portion size but only indicated how much they would eat. In addition, half of participants in the anchoring conditions received a discounting instruction stating that the portion size they had been asked to imagine was randomly selected and thus not informative for their consumption estimate. As expected, participants who imagined to be served larger portions estimated to consume significantly more food than participants in the no-anchor condition, and participants who imagined to be served smaller portions estimated to consume significantly less food than participants in the no-anchor condition. The discounting manipulation did not reduce this effect of the anchors. We suggest that the process of anchoring and adjustment may provide a useful framework to understand the portion size effect and we discuss implications of this perspective.