At the first interview in 1995, study participants constituted a multistage random probability sample that was representative of married and cohabiting couples in the 48 contiguous United States. A total of 1,925 eligible couples whose ages were 18 years old and older were randomly selected for the study. A total of 1,635 couples (3,270 individuals) completed the interview (85-percent response rate). Included in the sample were Black and Hispanic couples. In 2000, 1,620 couples were available to be reinterviewed, with interviews successfully completed with 1,392 of these couples (72 percent of the original sample). Of these couples, 1,136 were still married or cohabiting with the same partner as at the baseline survey. The study findings were limited to these couples. Most couples that reported violence in their interactions engaged in MPV. After controlling for other factors, Blacks were three times more likely than Whites to report MPV at follow-up; and Hispanics were nine times more likely than Whites to report MFPV. Findings indicate that ethnic minorities were at greater risk of MPV. In addition, the predictors of partner violence varied according to the type of partner violence. Further, the findings suggest that MPV and MFPV or FMPV alone are potentially discrete phenomenon with distinct sets of risk factors and potentially unique etiologies that require varying treatment and intervention approaches. The methodological strengths and weaknesses of the study are discussed. 4 tables and 30 references