Testosterone levels were examined in prisoners convicted of violent crimes
(n = 13), in men previously convicted of violent crimes but currently not i
n prison (n = 15), in nonviolent alcoholics (n = 15), and in randomly selec
ted control males (n = 16). Morning, afternoon, and evening testosterone le
vels were assessed after a minimum alcohol abstinence period of 24 hr, Viol
ent and nonviolent men did not differ in plasma total testosterone level on
any sampling occasion. In violent men, however, testosterone levels were s
ignificantly correlated with hostility, as measured by the Derogatis Sympto
m Check List. Most violent men were diagnosed with Antisocial Personality D
isorder (ASP) [DSM-III-R; 301,70], and the unweighted ASP symptom count als
o correlated significantly with testosterone levels in these subjects. We s
uggest that individuals whose life histories involve numerous antisocial be
haviors tend to have high testosterone levels even when interpersonal viole
nce is excluded. This, however, does not eliminate the possibility that mal
es who are characterized by high hostility may also have elevated testoster
one levels. Violent predisposition and antisocial conduct beginning in earl
y adolescence predict adult aggressive behaviors, which are augmented by po
wer-related alcohol expectancies and alcohol abuse. (C) 1999 Wiley-Liss, In
c.