Many studies of the power-and-prestige rank ordering in groups that ar
e heterogeneous racially, ethnically, or culturally may have inadverte
ntly confounded their results by sampling across socioeconomic class l
ines. This study was designed to eliminate that confound and to determ
ine which of three theories (expectation states, social role, tokenism
) could most accurately account for the power-and-prestige rank orderi
ng in heterogeneous groups of Caucasian and Chinese males. In all grou
ps, a Caucasian became the leader even if he was the token, whereas Ch
inese group members participated less and had less influence than thei
r Caucasian counterparts. These results are most accurately explained
by revising tokenism theory to include the status characteristics conc
epts from expectation states theory. Another finding war that the pers
on who participated the most was always judged to be the leader This f
inding has practical implications: Managers might promote talkative pe
ople ahead of more qualified bur less talkative, individuals.