Experiment 1 unexpectedly found sex differences in evaluative gender stereo
types (only men associated male with potency and only women associated fema
le with warmth). Experiment 2 dramatically reduced sex differences in gende
r-potency judgments when measures were redesigned to avoid implying that po
tency was positive (the concepts, strong and weak, were represented by eval
uatively matched words; e.g., destroy vs. feeble, loud vs. quiet, and might
y vs. gentle). Experiment 3 tested the hypothesis that these sex difference
s were in the service of self-esteem but found no correlation between own-g
ender-favorable stereotyping and implicit self-esteem. Rather, participants
showed a correlation between linking self to the favorable potency trait a
nd linking own gender to that trait. Experiment 4 confirmed the correlation
between implicit self-concept and gender stereotype using the contrast bet
ween potency and warmth for the implicit stereotype measure. In concert, re
sults suggest that people possess implicit gender stereotypes in self-favor
able form because of the tendency to associate self with desirable traits.