Three studies demonstrate that anchoring effects-the assimilation of a nume
ric estimate to a previously considered standard-depend on judges' knowledg
e about the target object. Manipulating judges' knowledge by varying the co
ntext in which a fictitious target is presented, Study 1 demonstrates that
anchor plausibility determines how a given anchor is processed and depends
on what knowledge base is used to judge the anchor. Studies 2 and 3 show th
at the less judges know about the target object, the more they assimilate t
heir estimate to the anchor. These differences in the magnitude of the anch
oring effect appear to be mediated by differences in the perceived plausibi
lity of the anchors. Furthermore, Study 3 demonstrates that these effects e
ven hold when the selection of the anchor values is under participants' con
trol. (C) 2000 Academic Press.