Objective: The objective of the study was to determine how testosteron
e levels, both alone and interacting with age, were associated with cr
iminal behavior and institutional behavior among female prison inmates
. Method: Subjects were 87 female inmates in a maximum security state
prison. Criminal behavior was scored from court records. Institutional
behavior was scored from prison records and interviews with staff mem
bers. Testosterone levels were scored from radioimmunoassay of saliva
samples. Results: Product-moment correlations revealed first-order rel
ationships among age, testosterone, criminal behavior, and institution
al behavior. Structural equation analysis suggested a causal model in
which age leads to lower testosterone, which in turn leads to less vio
lent crime and less aggressive dominance in prison. Conclusion: Testos
terone is related to criminal violence and aggressive dominance in pri
son among women, as has been reported among men. Changes in these beha
viors with age are in part explained by a decline in testosterone leve
ls.