Actually, yes it does. The third bar chart only looks at those who are murdered by someone close and says that 64% of women who are murdered are murdered by someone close and that 36% of men who are murdered are murdered by someone close
I haven’t done the math, but glancing at the numbers, yes, I would expect that to be correct.
It’s not a question of two different ways of looking at it - your calculation is wrong, as I have explained. You are mistakenly taking the 36% figure in the article to mean the proportion of all murdered men who are murdered by someone close. It is not, 36% is the proportion of all people murdered by someone close who are men.
The 64% women / 36% men figure quoted in the article IS the figure that you are wrongly trying to calculate. Almost twice as many women as men are murdered by someone close.
No, it does not. You are misreading it. The third bar chart takes all people murdered by someone close, and shows the M/F proportions of those homicides.
Then you will need to glance more carefully, it is not correct.
The counterintuitive point of the UN report and the BBC article is that although only 20% of the victims of all homicides are women, 64% of the victims of homicides committed by a partner or family member are women.
Again, that last figure is no something we need to calculate. It is stated explicitly in the article at chart 3 and in the original UN document. It is the figure that Quartz is wrongly trying to calculate.
Ah, yes, yes, I see. My glancing did cause me to misread that. I read that as 64% of all homicides of women were killings by a partner or family member, where that second graph is saying that of the homicides that were killings by partner or family member, 64% of them were female, and 36% were male, which means that in raw numbers, there are more female victims.
Got it. Skimmed it too cursorily.
Exactly. If you made the same initial misreading as Quartz, perhaps the BBC could have described that third bar chart more clearly. But that’s definitely what it means, per the linked UN report.
ETA: In fact on reflection I do think the BBC presentation is pretty clear and can only be read one way. If chart 3 were the proportion of all female homicides committed by someone close & the proportion of all male homicides committed by someone close as Quartz supposed, then those would be two separate percentages that do not sum to 100 - so it would not be presented in a bar chart like that. And the former statistic is already given in chart 1 (58%).
No, it should have been immediately obvious to me that I was reading that wrong, as that percentage should seem high and,more importantly, it would only add up to 100% by sheer coincidence if it had the meaning was as I was initially reading it. (ETA: which I see now is a point you just made in your addendum. :))
I think it’s clear from the article you cite that they’re talking about the percentages of the genders killed by those close to them and not the actual numbers of those murdered.
Also, not covered by those statistics is the fact that men kill most of both genders.
Again, nope. You seem to be just paraphrasing Quartz’s error.
The key statistic is: of all people murdered by someone close, almost two thirds (64%) are women.
The fact that this is very surprising given than only 20% of all homicide victims are women is the point of the report and the article.
Quartz’s math is not wrong*, but he missed the point of the article. As I clearly stated above, the article is not about the numbers of those actually killed in any category, it’s about the percentages of men and women killed by those close to them and that the percentage of women killed is far greater. Please read more carefully.
And it’s not surprising that so many women are killed by those close to them if one is paying attention to what goes on in our world today.
*Assuming 100 deaths, then 20 are women and 80 are men. 36% of 80 is 28.8 and 64% of 20 is 12.8, so he was right: more men are murdered by those close to them.
Please read the thread more carefully, I’m getting sick of repeating myself. Quartz is wrong, you are wrong, that math is wrong, because 36% is not the correct input into that calculation.
The input into the calculation that you and Quartz are trying to do is:
Take all murders where the victim is male.
X% is the proportion of those murders where the perpetrator is someone close.
X is not stated explicitly in the article, but if you do the math it works out to around 10%, not 36%.
And you don’t need to do the calculation that you Quartz attempt, because the result of that calculation is the statistic that is provided explicitly.
Once again, the statistic provided in bar chart 3 in the article is this:
Take all murders where the perpetrator is someone close.
Of those murders, 64% of victims are female, 36% of victims are male.
But as stated above by Riemann said above, this is not the correct interpretation of the 36% and 64% numbers. It wasn’t that 36% of men who were killed were killed by someone they knew, it was that 36% of those who were killed by someone they knew were men.
Of 100% of people killed, 20% are women 80% are men. (second bar graph on page)
Of the 20% women killed 58% were killed by someone close to them (first bar graph on page). So of the people killed 11.6% (20%*0.58) are women killed by someone close
This 11.6 represents 64% of the people killed by someone close (3rd bar graph on page). So in total of 18.125% (11.6%/.64) of people killed were killed by someone close
Therefor 6.525% of those killed are males killed by someone close (18.125% - 11.6%)
Therefor 73.475% of those killed are males killed by someone not close (80%-6.525%)
So the final breakdown appears to be
73.475% of those killed are males killed by someone not close.
9.4% of those killed are females killed by someone not close.
6.525% of those killed are males killed by someone close.
11.6% of those killed are females killed by someone close.
Quartz is incorrect.
nm
I attempted the same calculation at post 15, via the cited “one in five of all murders are committed by some close”, which is probably only accurate to one sf. You did the calculation a better way using the more accurate 58% figure, so your results are probably more accurate (except for an arithmetic error, second number should be 8.4% I think).
So we have:
Out of all murders,
73.5% perpetrator is a stranger, victim is male
8.4% perpetrator is a stranger, victim is female
6.5% perpetrator is someone close, victim is male
11.6% perpetrator is someone close, victim is female
From these figures, various outputs to back check:
Of all murders, proportion where victim is female?
(8.4%+11.6%) = 20%
Of all murders where victim is female, proportion where perpetrator is close?
11.6/(8.4+11.6) = 58%
Of all murders where perpetrator is close, proportion where victim is female?
11.6/(6.5+11.6) = 64%
Of all murders where victim is male, proportion where perpetrator is close?
6.5/(73.5+6.5) = 8%
The last figure (8%) would have been the correct input into the calculation that Quartz was trying to do, where he was wrongly using 36%.
Buck and Riemann should both read more carefully and stop making unwarranted assumptions. No where in the article does it say the three graphs are talking about the same numbers, especially in light of the fact that graph one’s numbers don’t match graph three’s, and it’s only reasonable to suppose graph two and three, which appear together, are using the same numbers. So whatever math you’re trying to do here is flawed from the get go. Using your logic, you should squeeze graph four into your calculations too.
On the other hand, who cares? The point of the article is that over half of women murdered are killed by a partner or relative, and that a woman is much more likely to be than a man.
Oops, some of my reasoning is flawed as well, so just skip down to the part about who cares.
This part is easy. The OP (Quartz) cares, or else he wouldn’t have posted it. He thought the BBC may be presenting the data in a grossly misleading way, but I think he will agree that he’s wrong since he got the math wrong.
The whole point of this thread was to discuss that specifically. I’m not a mod, but coming into a thread to ask “who cares” about the main point of that thread is not a great habit.