The American newspaper the Washington Post's PostEverything Op-Ed One way to end violence against women? Married dads. (previously titled "One way to end violence against women? Stop taking lovers and get married") claims that marriage plays a causative role in preventing violence against women, and their children.

This social media outpouring makes it clear that some men pose a real threat to the physical and psychic welfare of women and girls. But obscured in the public conversation about the violence against women is the fact that some other men are more likely to protect women, directly and indirectly, from the threat of male violence: married biological fathers. The bottom line is this: Married women are notably safer than their unmarried peers, and girls raised in a home with their married father are markedly less likely to be abused or assaulted than children living without their own father.

This includes domestic violence:

Women are also safer in married homes. As the figure above (derived from a recent Department of Justice study) indicates, married women are the least likely to be victimized by an intimate partner. They are also less likely to be the victims of violent crime in general. Overall, another U.S. Department of Justice study found that never-married women are nearly four times more likely to be victims of violent crime, compared to married women. The bottom line is that married women are less likely to be raped, assaulted, or robbed than their unmarried peers.

Does getting married play a causative role (and not just a correlation) against violence against women and children?

With regards to whether the op-ed is talking about mere correlation, or actual correlation.

The title implies causation. It has a question of "One way to end violence against women?", and then the answer "Married dads."

The blurb underneath talks about causation, not correlation: "The data show that #yesallwomen would be safer with fewer boyfriends around their kids."

And from the body itself:

  • some other men are more likely to protect women, directly and indirectly, from the threat of male violence: married biological fathers
  • But marriage also seems to cause men to behave better.
  • the research tells us that marriage provides a measure of stability and commitment to the adults’ relationship
  • So, women: if you’re the product of a good marriage, and feel safer as a consequence

In fairness to the article, there is some mention of factors that cause marriage causing a reduction of violence.

For women, part of the story is about what social scientists call a “selection effect,” namely, women in healthy, safe relationships are more likely to select into marriage, and women in unhealthy, unsafe relationships often lack the power to demand marriage or the desire to marry. Of course, women in high conflict marriages are more likely to select into divorce.

However, just because it mentions correlation in some parts doesn't mean that it isn't implying causation in the rest of the Op-Ed.

In addition, would it make sense to write an entire Op-Ed that is purely about a non-causative correlation?

  • 13
    Sounds like selection bias
    – HappySpoon
    Jun 11, 2014 at 0:53
  • 12
    For women, part of the story is about what social scientists call a “selection effect,” namely, women in healthy, safe relationships are more likely to select into marriage, and women in unhealthy, unsafe relationships often lack the power to demand marriage or the desire to marry. As I said above, sounds like selection bias.
    – HappySpoon
    Jun 11, 2014 at 6:28
  • 2
    @AndrewGrimm Pretty sure. From WP self-selection bias arises in any situation in which individuals select themselves into a group, causing a biased sample with nonprobability sampling. which is what is suspected here.
    – HappySpoon
    Jun 11, 2014 at 11:14
  • 3
    Off hand, I suspect that this might be one of those complicated topics that are going to be hard to really come up with a solid explanation for. Off the top of my head another reason why the figures could be lower for married women is because they aren't actively dating and expose themselves to more strangers who could be potential threats.
    – rjzii
    Jun 11, 2014 at 12:40
  • 3
    It seems implicit that the claim relates to the United States of America, but the answer might very well depend on socio-cultural region of the world.
    – gerrit
    Jun 11, 2014 at 15:34

2 Answers 2


No, there is no conclusive evidence that getting married prevents violence against women and children.

It is true that there is a correlation between cohabitation (living together without being married) and violence in relationships. This correlation has been shown in a number of studies. See for example Extent, Nature, and Consequences of Intimate Partner Violence.

However, there is no conclusive evidence that one is the cause of the other. Conclusive statistical evidence will of course be hard to find, because randomization (let alone blinding) is impossible, and without that the direction of the cause-effect relation is impossible to tell - does getting married help to prevent violence, or does not being violent help getting married? Also, there is no convincing explaination why marriage could prevent violence.

While some claim that:

Marriage also seems to cause men to behave better. That’s because men tend to settle down after they marry, to be more attentive to the expectations of friends and kin, to be more faithful, and to be more committed to their partners—factors that minimize the risk of violence.

(from the WP article in the question)

there are also opposing claims:

[...] differences in selection out of cohabitation and marriage, including selection of the least-violent cohabiting couples into marriage and the most-violent married couples into divorce, lead to higher observed rates of violence among cohabiting couples in cross-sectional samples.

Our results suggest that researchers should be cautious when making comparisons between married and cohabiting couples in which the dependent variable of interest is related to selection into and out of relationship status.

(Why are cohabiting relationships more violent than marriages?, PMID: 16579211, emphasis mine).


Yes, married parents lowers the risk of violence against women and children.

From the Bureau of Justice Statistics on which the study was based, a correlation in family status and intimate partner violence was indicated. Married females had a much lower victimization rate.

In 2010, the rate of intimate partner violence against females living in households comprised of married adults with children was lower than those of households with one female only. The rate of female intimate partner violence in 2010 among households comprised of one female adult with children (31.7 victimizations per 1,000 females age 12 or older) was more than 10 times higher than the rate for females in households with married adults with children (2.5 per 1,000), and more than 6 times higher than the rate for those in households with one female adult only (4.6 per 1,000).

This particular study did not perform any sort of regression analysis, so causal inferences shouldn't be made from those particular results. Luckily, the study’s author, Shannon Catalano, did perform such analysis. After the logistical regression, their analysis found that:

Results of the logistic regression for women, but not men, support previous research that shows unmarried couples are at greater risk of intimate partner violence than married couples, and African-American couples are at greater risk than white couples.

Similar findings are found in other studies. The fourth release of the National Incidence Study of Child Abuse and Neglect (NIS-4), 2004-2009 had similar findings. In the NIS-4 marital status is highly correlated with abuse.

Abuse. The rate of Harm Standard abuse for children living with two married biological parents (2.9 children per 1,000) is significantly lower than the rate for children living in all other conditions of family structure and living arrangement (10.2 or more children per 1,000). Again, the highest rate was among children living with just one parent and that parent’s unmarried partner (33.6 per 1,000 children). The rates in the highest and lowest risk groups differ by more than a factor of 11.

In their supplementary analysis, they examine the independent variables and their effect on abuse.

The finding that family structure relates to risk of child maltreatment is not new. Nearly two decades ago, Wilson, Daly, and Weghorst (1980) reported increased risk for children living in households with a parent and a surrogate parent (whether stepparent or cohabiting partner) compared to children in mother-only households.

White children had a notably higher probability of maltreatment when they lived with married parents who were not both biologically related to them and a slightly higher maltreatment rate when they lived with a single parent who had no cohabiting partner, whereas Black children had a considerably higher maltreatment probabilities when they lived with their unmarried parents and when they lived with a single parent living with a partner. [...]

  • 19
    The phrasing of your first sentence makes it sound like you are claiming a causal relationship, which the sources you cite explicitly do not do. Jun 13, 2014 at 16:31
  • 9
    I don't see how regression analysis can defeat the self-selection problem here. It's not like smoking where choosing to smoke has nothing to do with your natural propensity towards lung cancer. Jun 13, 2014 at 23:51
  • 12
    @user1873 The self-selection is from whether she chooses to marry him or not. Jun 14, 2014 at 3:28
  • 4
    @user1873 Correlation is not, and has never been causation. You can't claim a causal effect just because you don't have a better answer. There are serious negative moral implications to controlled tests that deliberately create an environment to dis/encourage domestic violence. And you're likely right. Marriage is a big deal to people so asking for volunteers probably does have strong selection bias, but it's the only way to mitigate the morality concern. Studying this is not an easy thing. And you STILL don't get declare that one causes the other until you do. Welcome to Skeptics. Aug 5, 2014 at 14:27
  • 9
    @user1873 Yes but it doesn't demonstrate cause, it demonstrates a relationship. That's the point I was getting at. Aug 5, 2014 at 14:57

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .