6

Training Institute on Strangulation Prevention - makes the claim:

A woman who has suffered a nonfatal strangulation incident with her intimate partner is 750% more likely to be killed by the same perpetrator…with a gun.

This 2021 Guardian article about strangulation lists TISP as a source:

According to one study, [... once] a woman has been strangled by her partner, the [...] likelihood that he will murder her rises nearly eightfold.

I am wondering if the original claim is true (supported by research), given that original source does not list any studies:

Is the likelihood of a woman being murdered by their partner 7.5 times greater after the partner strangles her non-lethally once?

Looft
  • 441
  • 4
  • 7
  • 2
    I stumbled upon [one study](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2573025/) that gives some numbers. I guess there will be many more. – Looft Jul 12 '22 at 15:27
  • 2
    I'll admit that I had to click through onto the site/article to verify that they were talking about non-consensual strangling rather than breathplay, although I've seen arguments that the latter is a red flag as well, especially in situations of questionable or post-fact consent. – Sean Duggan Jul 15 '22 at 15:15
  • 11
    Just seat-of-the-pants, off-the-cuff, I'd have to imagine that a huge precedent of violence "is bad"... whatever the numbers. I'm saddened to think that some people would feel trapped to stay with a partner who had repeatedly assaulted them in a potentially fatal fashion... – paul garrett Jul 16 '22 at 19:06
  • 2
    So I first read this title as "strangled ... 8 times" and had to read the body of the question before I understood the title accurately. Because if your partner (non-consensually) strangles you on 8 occasions I don't know if a study is required... – Mark Henderson Jul 18 '22 at 01:46
  • 2
    Just looking at it statistically... Both strangulation and murder are rare; I assume murder is more rare. Say 1 million women, 1000 strangulated, 100 murdered. Chances of being murdered are in in 10,000. Chances of being murdered after strangulation about one in 1,300. That is less than one woman murdered after being strangulated. Or the other way around, less than one of the 100 murderers strangulated their partner earlier. I would have thought that the chances that a murderer acted violently earlier would be bigger than that. – gnasher729 Jul 18 '22 at 14:39
  • The 'institute for strangulation prevention' has an article that says "... According to the Training Institute on Strangulation Prevention, 'A woman who has suffered a nonfatal strangulation incident with her intimate partner is 750% more likely to be killed by the same perpetrator' ..." so there is some circularity there - the papers i could find on strangulation in ipv were more concerned with the lethality of strangulation in itself, and the neurological damage that can occur, instead of the statistical implications – bukwyrm Aug 03 '22 at 12:24
  • The wording in the TISP quote makes sense. In contrast, the wording in the Guardian quote and your question do not make sense: It does not make sense to claim that before being strangled by her partner, a woman had probability *x* of being murdered, but right after being strangled once, the probability rose to 8 *x*. –  Apr 01 '23 at 04:41

1 Answers1

5

the original claims ends in a circular sourcing: The Training Institute on Strangulation Prevention says:

According to the Training Institute on Strangulation Prevention, “A woman who has suffered a nonfatal strangulation incident with her intimate partner is 750% more likely to be killed by the same perpetrator…with a gun.

This article by Glass, Laughon, Campbell, Wolf, Block, Hanson, Sharps, Taliaferro

"Non-fatal strangulation was reported in 10% of abused controls, 45% of attempted homicides and 43% of homicides"

This seems to uphold the claim somewhat, showing a far higher risk for homicide in an abusive relationship that includes strangulation, though not at the 7.5-times factor the claim puts forward.

Weirdly, the same article says:

"Wilbur and colleagues' in 2001 found that 68% of a convenience sample of 62 women presenting to a domestic violence advocacy program reported strangulation by their abuser"

i.e compared to that, the ~50% previous strangulation they found in their murder sample would suggest previous strangulation is actually negatively correlated with later murder(&-attempts). Their own numbers came from finding women that reported abuse on randomized cold calls, though, and the 68% figure relates to a sample of women (self?)presenting to an advocacy program, so possibly already suffering from far higher levels of abuse.

ibid:

"The Chicago Women’s Health Risk Study (CWHRS) found that 24.6% of 57 adult women killed by a male intimate partner in 1995 or 1996 in Chicago were killed by strangulation or smothering" , and later, ibid: "There was no difference between women who were not killed and the women who were killed in having experienced prior choking or strangulation.

so the actual predicitve quality of strangulation seems rather weak, overall. But: The claim was murder by the same perp, and with a gun, which might wildly sway the statistics. Also, it is never explained against which baseline this would work? Strangulated women vs all women? Strangulated vs all IPV victims?

Strange addendum: there is a "twin" study from 2003 that mirrors Glass et al 2009 on the time frame of the murders, the number of cities, even down to the size of the random control base and their verbatim description

"A total of 4746 women met the age and relationship criteria and were read the consent statement." Nancy Glass et al 2009 and Jacquelyn C. Campbell et al 2003

The 2003 study does not go into strangulation, though. honi soit... (Research Committee of TISP is led by Dr. 'Jackie' Campbell)

bukwyrm
  • 1,585
  • 1
  • 20
  • 31
  • 1
    I am tired, so forgive me if I am missing the point, but how do you conclude the rate is "not at the 7.5-times factor the claim puts forward." from the data you provide? – Oddthinking Aug 15 '22 at 14:13
  • @Oddthinking :To me, the claim about the 750% is very quiet on the comparison for that 750% - but if 43% of homicides have non-fatal strangulation (by any perp) in their background, and 10% of abused controls have non-fatal strangulation in their history - and as a strangled person is definitely abused, there are no non-abused but strangled persons. The homicides are a supergroup including gun-homicides by the same perp - thus gun-homicides by the same perp are at the highest at 43% (though likely much lower) and non-fatal strangulation is at 10% thus the ratio is at most 430% – bukwyrm Mar 23 '23 at 06:09