Objective: To determine the relationship between length of soldier deployme
nt and self-reports of moderate and severe spousal violence. Methods: The C
onflict Tactics Scale was used to measure self-reports of behaviors exhibit
ed in marital conflict. Surveys were administered to a 15% random sample of
26,835 deployed and nondeployed married active duty U.S. Army men and wome
n in the 50 United States during the period 1990 to 1994, Multinomial logis
tic regression and ordered probit analysis were used to estimate the probab
ilities of moderate and severe violence by length of deployment. Results: A
fter controlling for demographic variables, the probability of severe aggre
ssion was significantly greater for soldiers who had deployed in the past y
ear compared with soldiers who had not deployed. Conclusions: Deployment co
ntributes a significant but small increase to the probability of self-repor
ted spousal aggression during a 1-year period. Although deployment is a mil
itary operation, similar effects may be observed in certain civilian occupa
tions.