Have a Million American Women Been Shot at by an Intimate Partner?

Kamala Harris is using this claim to promote her gun-control plan:

I’m the first person to say that we could and should do more to disarm abusive men, roughly a thousand of whom killed their wives and girlfriends in 2015. But 1 million seemed rather high, considering that it implies nearly 1 percent of women have had a firearm discharged at them by an intimate partner and survived. So I did some digging.

The claim appears to come from this literature review, whose abstract indeed says that “the number of U.S. women alive today who have had an intimate partner use a gun against them is substantial: About 4.5 million have had an intimate partner threaten them with a gun and nearly 1 million have been shot or shot at by an intimate partner.”

But then I dug a little deeper. The review looks at a lot of different surveys, but most of them are of battered women only, or ask only about recent incidents. Only one survey — conducted in the mid 1990s and reported in 2000 — asked a representative sample about incidents covering their whole lives.

And that survey didn’t ask whether the women had been shot or shot at; it asked if anyone had “use[d] a gun” on them. Incredibly, the literature review itself notes the difference between this and being shot at: “Gun use — which was not defined in most studies and, therefore, might include gun threats as well as being pistol whipped and being shot at or shot — against an intimate partner was low in the general population (Tjaden & Thoennes, 2000).”

As it turns out, “Tjaden & Thoennes, 2000” is precisely the source of the 1 million claim: 0.7 percent of women in the survey said an intimate partner had “use[d] a gun on” them, which the authors of the literature review then multiplied by census numbers to arrive at a raw number (even though crime has fallen substantially since the 1990s). An earlier question in that survey asks if the respondent has been “threatened” with a gun, so this likely doesn’t include very many mere threats. But it does not suggest that this many women were “shot or shot at,” just that a gun was “used” in some way.

It’s also worth noting that trying to estimate very low numbers with surveys is tricky. If even a tiny percentage of respondents misunderstand the question or lie, the false positives can swamp the true ones. This is the same problem we encounter when trying to estimate defensive gun use, after all.

So again: Let’s do more to disarm abusers, no matter how they’re “using” their guns. But let’s build the case for that with better numbers.