“Hate begets hate; violence begets violence; toughness begets a greater toughness,” Martin Luther King wrote in 1958. He meant to explain his nonviolent tactics in the fight for civil rights, of course, but his words have been applied far beyond that context, including to the sphere of violence within families. Research confirms the intuitively plausible theory that children who experience and/or witness violence within their families are more likely to perpetrate and suffer violence in their romantic relationships in adulthood.

However, a new meta-analysis finds that the magnitude of the effect is not as large as one might assume. In last month’s Journal of Family Theory and Review, Erika N. Smith-Marek, Bryan Cafferky, and others examined the results of 124 studies published between 1980 and 2013 to investigate this question. The underlying studies compared adults’ retrospective reports of whether they witnessed violence between their parents and suffered physical violence at their parents’ hands as children to their reports of whether they had been victims or perpetrators of intimate partner violence (IPV) as adults.

Smith-Marek and her colleagues reported that overall, their meta-analysis revealed only modest links between family-of-origin violence and committing and suffering IPV in adulthood (r = .25 and r = .21, respectively).

The nature of the family-of-origin violence—whether it was a parent using physical violence against the child’s other parent or against the child him- or herself—made no significant difference for subsequent IPV risk. However, some studies failed to distinguish between respondents who had seen and suffered family violence and those who had only experienced one or the other, which limited the precision of the comparison. And most studies did not differentiate between types of interparental violence, such as situational couple violence versus intimate terrorism.

In contrast with the nature of family violence in childhood, the sex of the respondent did turn out to matter: While family violence in childhood increased each sex’s likelihood of both perpetrating and experiencing intimate partner violence in adulthood, the link between family-of-origin violence and later perpetrating IPV was significantly stronger for men (r = .25 vs. r = .19), and the link to IPV victimization was significantly stronger for women (r = .22 vs. r = .16).

Smith-Marek et al. consider their results consistent with the social learning theory of family violence, which “posits that children raised in a violent home, through processes of observational learning, modeling, and direct behavioral conditioning, will come to see violence as an appropriate response to conflict.” Yet they believe that the small effect sizes imply that other risk factors, like the aggression of people’s peers and the history of their romantic partners, must also influence individuals’ risk of relationship violence throughout life. Indeed, they see their findings as bolstering the argument that violence in one’s family of origin, “contrary to popular belief, does not play a central role in the development of adult IPV” (emphasis mine). Violence can beget violence, but it doesn’t have to.

If you or someone you know is in a violent or unsafe relationship, contact the National Domestic Violence Hotline at 800-799-7233 or visit their website at thehotline.org.