Statistics taken from criminal reports, from shelters for battered women, and from therapy sessions for males convicted of violence towards their intimate partners show two clear trends: first, that men tend to be overwhelmingly more aggressive than their female intimate partners and that in those couples where the violence is mutual, women are overwhelmingly more likely to be injured in the violence. These statistics are often cited in support for continued funding for domestic violence programs targeted towards protecting women and, sometimes, even in the creation of conditions that are patently inimical to battered men.
What these statistics largely ignore is that there is a selection bias in how this data is collected that overstates effect sizes of male-on-female violence. Such statistics often consider the act of violence out of context, and largely ignore the consequences of the act. Numerous authors have showed that if one studies the evolution of the conflict up to the terminating act of violence, then any clear trend in the division of perpetrators and victims by gender disappears. Similarly, if violence were to be studied only in terms of the outcomes of the violence, the majority of the victims turn out to be women, yet, acts of physical violence are reported equally frequently by members of either gender.
These selection biases often reflect the personal biases of the people collecting and analyzing the data. There are primarily two groups of people who study such issues; family conflict researchers and feminist researchers. Family conflict researchers typically study representative samples of married, cohabiting or dating couples whereas feminist researchers typically sample from populations selected to reflect high level of violence from male partners.
In a study published in 2000, Archer conducted a meta analysis of a large number of studies done in the 1980s and the 1990s in the US. The samples for these studies were primarily young adults. In his meta-analysis, he observed some interesting phenomena.
- Based on self-reports, women were more likely than men to indulge in physical violence, but if one were to study partner reports, this difference disappears. A possible explanation for this is that women chronically report more violence (from themselves and their partners) than men do.
- The effect size of injury is larger for men than for women, and even removing outliers, the difference is statistically significant. This means that significantly more women are injured in domestic violence than men.
- If one measures violence based on specific acts, and not just on injury, then women are significantly more likely to use physical aggression towards their partners, and are likely to use it more often.
- One can measure violence by using two metrics: act-based metrics and meaning based metrics. The meaning based metric uses the qualifiers undefendable, intimidating and injurious (aggression form which partners cannot defend themselves and that which has a high chance of causing injury). Act based metrics have a significant effect size towards the female population (more women do it), while meaning based metrics have a significant effect size towards the male population (more men do it).
At the very least, Archer’s study brings up two important conclusions. Firstly, domestic violence does not fit the predominantly feminist position of involving a male perpetrator and a female victim. Second, the kind of violence that men and women indulge in is different, and needs to be dealt with differently.
This also helps me put into context Erin Pizzey’s statement that of the battered women she worked with, more than half were as or more violent than the partners they left.
The complete reference for Archer’s study is Archer, J. (2000). Sex differences in aggression between heterosexual partners: A meta-analytic review. Psychological Bulletin, 126 (5), 651-680.
A Google search also throws up a pdf that is not blocked by a journal pay-wall.