My theory is that any distortion of real data in fact causes more victims because you're not allocating resource efficiently when you model skews from reality. Resource allocation is only as good as your data.
A political example of that is the "wage gap". 77 cents per the dollar, but if you correct for hours worked etc, seniority etc then it's like a few cents to the dollar - and by some measures women are ahead at that level. Almost 100% of the wage gap is in fact related to women with children, so if you want to help, help women with children. Women without kids are on near parity with men. That actually means the pay gap for women with kids is much worse than the 77 cents, because this is merely the average pay gap. So not only are they misallocating effort, they are in fact severely underestimating the scale of the wage gap facing women with children, in order to "be good feminists".
So they pass laws aimed primarily at young childless Democrat-voting women promising to raise their paychecks by closely scrutinizing businesses to make sure they're not underpaying women for the same work as a man. But such a business practice would be extremely costly. If men cost 20% more than women for the same role, you wouldn't hire them. The wage gap exists between industries, not in single roles. And what is the new Equal Pay law going to do about industrial segregation? Probably nothing, and could even make things worse somehow. The model they're proposing is to make it easier to bring litigious cases against employers for wage discrepancies. And even if they're paying you properly, the mere hint of a lawsuit means they need to hire more lawyers just in case. People being "potentially litigious" makes their market value lower. It's a quantifiable risk, and businesses assess possible risk when hiring, and they deduct that from how much they're willing to offer you. That's not discrimination it's just sound business practice. So attaching more potential regulatory issues to hiring women might backfire.
I've mentioned the Duluth Model of domestic violence treatment before. It ticks all the right "ideological boxes" so they proponents promote it as the "one true method" of rehabilitating male abusers. However, when you do controlled trials of it's effectiveness at reducing domestic abuse recidivism, it approaches 0%, because as even the creators admit "we saw what we wanted to see", rather than tailoring the approach to real world data and situations. It's probably worse than 0% however, because proponents of this treatment are actively hostile to alternative treatments that teach anti-stress, life, coping, anger management or conflict de-escalation skills to abusers. I.e. they are actively hostile to stuff that might actually help, because they believe admitting that teaching coping skills could reduce abuse is admitting that abusers have actual issues at play other than being a male asshole. So guys who are abusers and are seeking treatment are being denied access to programs that could actually be effective at helping female victims, because of smug back-slapping Duluth people who are just happy they kept "the program" ideologically pure.
Another example where only focusing on only one type of abuse hurts the intend people you're trying to help is in child abuse. There was some British research where they looked at cyclic patterns in child sexual abuse. You know the idea that if you're abused you become an abuser. Contrary to the media's perception there was about zero% effect from male abusers. The biggest effect was on male childhood victims of female abusers. The abuser group had a 1 in 3 chance of having been abused by a female perpetrator, and the non-abuser group only had a 1 in 9 chance of that. If that connection checks out, then turning a blind eye to female child abusers / male victims in fact creates many more female victims a generation later. So I'd argue that when you pick and choose victims you care about, you make everyone a victim. Treating gender as a competition is making everyone a victim.