This article explores the links between implicit self-esteem and the automa
tic self (D. L. Paulbus, 1993). Across 4 studies, name letter evaluations w
ere positively biased, confirming that implicit self-esteem is generally po
sitive (A. G. Greenwald & M. R. Banaji, 1995). Study I found that this name
letter bias was stable over a 4-week period. Study 2 found that positive b
ias for name letters and positive bias for birth date numbers were correlat
ed and that both biases became inhibited when participants were induced to
respond in a deliberative manner. Studies 3-4 found that implicit self-eval
uations corresponded with self-reported self-evaluations, but only when par
ticipants were evaluating themselves very quickly (Study 3) or under cognit
ive load (Study 4). Together, these findings support the notion that implic
it self-esteem phenomena are driven by self-evaluations that are activated
automatically and without conscious self-reflection.