The internet loves a shitty statistic, and the lifespan of those statistics is a fascinating thing.
What begins as a careful, considerate paragraph in a journal article becomes a guy with an AI-generated profile picture blasting off from his keyboard: “Actually, studies prove...” I would know, because I’ve been on the receiving end of one of those troubling stats too many times to even ballpark it.
Here’s an example from a video I posted in December of 2025, where I made a joke about lesbians loving women differently than men. As happens on the internet, my video left my circle of safety, and the comments got weird.
Let’s go to school for a minute.
The study that this commenter probably doesn’t even know they’re citing here is: The National Intimate Partner and Sexual Violence Survey: 2010 Findings on Victimization by Sexual Orientation.
And very simply, the data was interpreted as saying “Lesbians report the highest rates of domestic violence.”
I can’t stress enough how often I see some version of this parroted across social media platforms.
If I type “lesbian DV” into TikTok, here are some of the first comments that come up:
“People don’t like hearing it, but lesbians actually have the highest domestic violence rate of any relationship.”
“Statistically, women are more abusive than men in relationships. Just look at lesbian couples.”
“Everyone blames men for domestic violence, but lesbian relationships have the highest abuse rates. The data is public.”
“Women always say men are the problem, but if you remove men from the equation, domestic violence actually gets worse.”
“Google it. It’s 43.8%. This isn’t even controversial.”
Here’s what the study actually says:
Queer women are significantly more likely than straight women to have ever experienced intimate partner violence in their lifetime, reported by 61% of bisexual women and 44% of lesbian women, compared with 35% of straight women.
The part that is either misconstrued or conveniently left out is that the CDC also collected information about the sex of the perpetrator.
Among lesbian respondents who reported intimate partner violence, 67.4% reported only female perpetrators, while the remaining 32.6% reported at least one male perpetrator, meaning their lifetime experiences of abuse included violence by a man, or by both men and women.
When you look at the CDC’s follow-up report using data collected in 2016 and 2017, the agency changed how it reported perpetrator sex and broke out different categories of violence in much greater detail. For several forms of intimate partner violence, most lesbian victims identified exclusively male perpetrators.
The methodology isn’t identical to the 2010 report, so the numbers shouldn’t be compared directly, but they point to the same conclusion: the study was never evidence that women abuse women more than men do.
What should be: “Lesbian women experience high rates of intimate partner violence", became: "Lesbians have the highest rates of domestic violence."
Queer women experience high levels of male violence for concrete, structural reasons rooted in patriarchy, sexual entitlement, and the policing of women’s autonomy, not because there’s something inherently dysfunctional about queer relationships.
Some men don’t experience queerness as a boundary. They experience it as a challenge, a provocation, something to overcome. That can manifest as coercion, punishment, stalking, or, in its most extreme forms, what is often described as “corrective” sexual violence. These dynamics are especially relevant for bisexual women, who consistently report the highest rates of intimate partner violence and whose abusive experiences are disproportionately perpetrated by men.
Bisexual women are frequently framed as sexually adventurous and available. If someone believes your sexual orientation is negotiable, then it’s apparently not a huge leap to believing your consent is, too.
Compulsory heterosexuality also matters. Many lesbians have relationships with men before coming out, or while trying to reconcile their sexuality with social expectations. Those experiences are reflected in this data.
Even outside intimate relationships, queer women whose gender expression falls outside conventional expectations experience elevated rates of harassment and hate-motivated violence. Taken together, the pattern isn’t queer relationships producing violence, it’s queer women living in a society that punishes women for rejecting their expectations.
Data is evidence, but it is not truth.
A statistic doesn’t tell you what happened. It tells you how a question was answered by a particular group of people, in a particular way, at a particular moment in time.
Most research on intimate partner violence relies heavily on self-report surveys. Those surveys don’t measure the objective presence of abuse. They measure someone's ability to recognize what happened to them as abuse, whether they feel safe disclosing it, and whether the questions being asked capture their experience in the first place.
Researchers know this, which is why they don’t rely on a single source. They compare survey data with police reports, hospital records, shelter usage, hotline calls, court records, and protection orders. Even then, they’re careful about their conclusions because every source has its own blind spots. Police data underrepresents communities with lower trust in law enforcement. Shelter data misses people who never look for their services. Surveys miss people who don’t identify their experiences.
Research on queer relationships is also young. Sociology and family studies have spent nearly a century examining heterosexual relationships. Large-scale research on openly queer relationships has only really become possible in the last few decades, and in many places, only after same-sex marriage was legally recognized. We’re trying to compare bodies of research separated by generations of legal rights, social acceptance, and available participants.
So, now, let’s look at lesbian divorce rates. Another thing Simreaper1 wanted to talk about as evidence that lesbian relationships are uniquely unstable.
This is where there is some partial truth. Lesbian marriages tend to end in divorce more often than marriages between two men, and, in some datasets, more often than heterosexual marriages. But divorce rates don’t measure the quality of the relationship, they measure access to the exit door.
Lesbian couples tend to face fewer structural barriers to leaving unhappy relationships, and have higher expectations for emotional reciprocity. Having already distanced from social and traditional norms, long-term relationships between same-sex partners do not carry the same expectations. They have fewer children, they’re less likely to be constrained by traditional gender roles, less likely to have one financially dependent stay-at-home partner, and less likely to remain together solely because leaving is economically impossible.
And because same-sex marriage has only been legally observable for a relatively short period of time, the original data largely reflects marriages of couples who were waiting to marry. They tended to be older, more legally literate, and more willing to leave relationships that no longer meet their needs.
Higher divorce rates can reflect greater autonomy just as easily as greater dysfunction.
Now, let’s make one thing very clear: women can be abusive.
Lesbian relationships can be violent. Queer people don’t hold moral superiority because they’re queer. We are as capable of hurting each other as anyone else.
But if your understanding of domestic violence begins and ends with a statistic you saw in a TikTok comment, you’re not learning about abuse; you’re learning how misinformation spreads.
The people doing the hard work on domestic violence aren’t winning arguments in comment sections. They’re staffing shelters, answering crisis lines, helping people write safety plans, documenting injuries, navigating family court, and trying to keep victims alive.
A statistic is a piece of evidence, not a verdict. It exists within a study, which exists within a methodology, which exists within a social, historical, and cultural context. Strip away those layers, and almost any number can be made to support almost any argument.
We have to ask who was surveyed, who wasn’t, what was measured, what wasn’t, and whether the conclusion being shared is actually the conclusion the researchers reached. Then hold that information alongside everything else we know: history, sociology, psychology, lived experience, and the limitations of the data itself.