When do prototypes bias person memory? Differential effects of abstraction level

Three experiments investigated the operation of prototypical person categories, which were either at
a superordinate or at a subordinate level of abstraction, on person memory. In Experiment 1, a
recognition memory experiment, subjects received person attributes describing fictional target
persons, each description pertaining to one of the two abstraction levels. Distractors in
the recognition test varied in degree of relatedness to a prototypical category. Distractors that
were related to the corresponding category were falsely recognized, if the category was at a
superordinate level, but not for a subordinate category. This revealed a significant bias in recognition
toward conceptually related but nonpresented items only at the superordinate level. In Experiment 2
which used a conceptual priming procedure subjects responded faster to distractor items related to a
superordinate category than to those related to a subordinate category. Experiment 3 replicated the
difference between levels of abstraction when subjects’ task was to recall rather than to recognize the
presented attributes. The findings imply that person categories at different levels of abstraction
operate differentially on person memory and restrict assumptions from prototype theory to a
superordinate level.