In this age where serotonin seems to be the chic neurotransmitter linked to violence, we often forget the well studied and established biological link to violent behavior: testosterone. Following up on earlier compelling research, James Dabbs of Georgia State University in Atlanta studied how testosterone levels, both alone and interacting with age, were associated with criminal behavior and institutional behavior in a female maximum security prison population of 87 (Psychosomatic Medicine 59:477-480 (1997)). He found that testosterone is related to criminal violence and aggressive dominance in women, as has been reported among men.

Testosterone was assayed from saliva samples. Criminal violence was scored from court records of crimes for which subjects were incarcerated. Crimes were assigned a point value of 1 to 3: (1) nonviolent, (2) defensive-violent (responding to abuse), and (3) violent (homicide, assault, or robbery with no history of having been abused by the victim).

Aggressive dominant institutional behavior was scored from prison records and staff reports on a 3-point scale: (1) passive, (2) neutral or (3) aggressive dominant (staff identifying the subject as "always in trouble," physically aggressive or threatening others, dominating informal prison social groups, or acting as if "they think they run the place," and records showing repeated rule infractions).

A double scoring procedure was used. Aggressive dominance was first scored by the study investigator, based on written prison reports of misbehavior. These scores were then reviewed by three staff members, one of whom was the sergeant in charge of security. When staff members disagreed with the investigator's scoring, her scoring was changed. Both investigator and staff members were blind as to subjects' testosterone levels.

Analysis of variance showed testosterone significantly related to behavior in prison: the highest concentrations of the hormone were found in the aggressive dominant group. First-order product-moment correlations among age, testosterone, criminal violence, and aggressive dominance showed that testosterone decreased significantly with age; that aggressive dominance decreased significantly with age and increased significantly with testosterone; and that no relationship existed between aggressive dominance and criminal violence.

A LISREL structural equation causal model showed that age leads to lower testosterone, which in turn leads to less violent crime and less aggressive dominance in prison. These study findings underscore the potential for antiandrogenic interventions in treating violence irrespective of sex offenses.