For 22 years, domestic violence experts have estimated that a staggering number of children — between 3 million and 10 million — live in homes where violence occurs.
A new study has updated — and increased — that number, to 15.5 million children, including 7 million who live in homes with "severe" violence between adult partners.
Experts in domestic violence say these new numbers are an overdue recalibration of the tragic reality they see.
But the study, published in the Journal of Family Psychology, may also reopen a long-festering ideological argument about whether men or women are the most violent in the home.
The study found that, contrary to public perception, women committed more acts of violence than their male partners in 11 overall categories of violence. Specifically, women were more likely than men to throw something, push, grab, shove, slap, kick, bite, hit or threaten a partner with a knife or gun.
However, men were more likely than women to commit "severe" acts of violence, such as beating, choking, burning, forcing sex or actually using a knife or gun on their partners.
When minor and major acts of violence were tallied, female-to-male violence accounted for 18.2 percent of overall violence and 7.5 percent of severe violence. Male-to-female violence accounted for 13.7 percent of overall violence and 8.6 percent of severe violence.
The study, which is based on an interviews with 1,615 married or cohabiting couples and extrapolated nationally using census data, found that 21 percent of couples reported domestic violence. Around 60 percent of these homes contained minor children, which allowed researchers to estimate the national numbers of children living in homes with some violence to severe violence, updating much lower earlier estimates.
The new, larger estimate "has serious implications for policy and practice," said Steven Marans, director of the National Center for Children Exposed to Violence at the Yale University Child Study Center in New Haven, Conn.
Children exposed to domestic violence often develop antisocial behaviors; have problems with social skills, learning and "emotional regulation; and are more likely to be involved in domestic violence as adults, said Mr. Marans.
"The policy implications are absolutely clear," he said. If the nation doesn't confront the problem of family violence when children are young, "we pay through the teeth in multiple other ways" when they are grown.
Jackie Warrilow of the National Resource Center on Domestic Violence in Harrisburg, Pa., said the updated estimate of children living in violent homes is "far more representative" of the problem. She agreed with the study's authors that even 15.5 million is an undercount because the study didn't include single parents, homosexual couples or separated couples. "We know that violence escalates when the victim separates from her perpetrator," she said.
But Miss Warrilow's biggest concerns were the study's findings on who was committing the violence.
"Without understanding domestic violence, somebody looking at this [data] would assume that women are more violent than men in intimate relationships," Miss Warrilow said. But that's not borne out in much of research, government reports, victim service agencies and law enforcement, she said.
She also regretted the study's use of the Conflict Tactics Scale and its 11 categories of violence. The scale has been "refuted for many years," she said, because it doesn't take into account the frequency of violence, its intensity, intent or context.
The study's home interviews also could have resulted in skewed numbers, she added. Batterers typically "do not own their violence," but blame others, she explained. "So if you are interviewing the perpetrator, you're not going to get a clear vision of what happened, of what that person is doing." Victims, on the other hand, "tend to place responsibility on themselves because that is what they have been hearing," she said. "Without context, the numbers are very concerning."
However, other experts say the study confirms what has been widely known: Both men and women are fully capable of family violence.
Richard Gelles, dean of the School of Social Work at the University of Pennsylvania, said male-to-female violence is far more common at the most severe levels of violence, he said. Intimate-partner homicide data, for instance, show that in 2004, there were 1,200 females killed by their male partners, compared with 388 males killed by their female partner.
But domestic violence is not a "gender crime" against women, he said.
"The science is just overwhelming" that, in most kinds of violence, female-to-male rates of violence are equal to or exceed male-to-female violence, Mr. Gelles said.
Instead, he categorizes it as a "family violence" crime, in which both men and women can be culpable, he said. "I am hoping to see people stand up and say, 'Enough of the ideology. Enough of the advocacy.' "
Renee McDonald, one of the five authors of the federally funded study, said the findings on higher rates of female-to-male violence were not surprising because they were consistent with other studies.
"But we don't know the context" of the violence, she cautioned. "We don't want to minimize [female-to-male violence], but on the other hand we don't want to forget the fact that men can be much more harmful to women."
Mrs. McDonald, who teaches psychology at Southern Methodist University, called for more research to understand how family violence affects children and adults.
"What is the threshold" for when seeing violence harms a child, she asked. "Is it one hit? Two pushes? Five shoves? A beating? We don't know. Is it once a year? Twice a year? Ten times a year?
"We don't know very much about when [living with violence] begins to be a problem for children; we just know it doesn't take much of it for there to be an elevated risk for problems," she said.