The UN's own study says boys face discriminatory treatment. The version sent to member states left that out.

What the study is, and the sentence in it that nobody quotes.

In December 2014 the General Assembly invited the Secretary-General to commission an in-depth study on children deprived of liberty. Manfred Nowak was appointed to lead it in October 2016, and the result was transmitted to member states on 11 July 2019 as A/74/136. It is the most complete count anyone has of the children the world locks up. The headline number is in paragraph 86: a minimum of between 1.3 and 1.5 million children deprived of liberty every year. Institutions hold the largest share, then the administration of justice, then migration detention, then armed conflict. That is the figure the coverage led on, and it deserved to.

Fifty-one paragraphs before it, on page nine, there is a section headed Gender dimension. It is four paragraphs long. This is the first of them, in full.

Ninety-four per cent. Not a campaign's figure. Not an advocacy group's estimate. A study commissioned by the General Assembly, delivered by an Independent Expert, transmitted by the Secretary-General, and sitting on page nine of a twenty-three page document that member states were sent.

The study does not only count. It explains, and one of its explanations is remarkable.

The next paragraph asks why. It gives a funnel: girls are attributed roughly a third of all criminal offences committed by children worldwide, and receive 6 per cent of prison sentences. It then lists reasons, and it lists them fairly. Girls commit less violent offences. Girls are more often accused of status offences. Girls are generally first-time offenders and, in the study's words, more receptive to the deterrent effect of incarceration. Those are real and they are not disputed here. But the study does not stop there.

Read that again with the campaign's usual framing removed. A UN study is saying that a gender stereotype produces a worse outcome for boys, and naming the mechanism: adults in the system assume boys need less protecting. That is a finding about how a stereotype harms the sex it is supposed to flatter, written down by the United Nations, in 2019.

One paragraph later, a second finding, this one about fathers. Most states allow a convicted mother to keep her young child with her in prison. The study reports that only eight states explicitly permit a father to do the same, and that even where it is permitted there are almost no father and child units, so in practice it does not happen.

Now the part this site is obliged to print, because the same document says it and leaving it out would make this article the thing it is criticising. Paragraph 38 is about girls, and it is not a footnote.

Paragraph 39 does the same for LGBTI children, who it reports are overrepresented in child justice facilities, placed in unsuitable facilities and particularly vulnerable to violence. Both paragraphs are true, both matter, and neither is in tension with the 94 per cent. A group can be the overwhelming majority of a detained population and other groups can still face distinct harms inside it. The study holds both. So does this article.

Except that it does not hold them the same way, and the reason took me to a second document. Behind the twenty-three page note there is the study itself: 804 pages, published in 2019 under Manfred Nowak's name. Its chapter 8 is called Gender Dimension, and it opens like this.

Boys often face discriminatory treatment. That is not this campaign's phrasing. It is the study's, in the opening line of the chapter where it examines sex. It names the mechanism too: stereotypical views about how violent boys are.

Now search the note that was sent to member states for that finding. Discriminatory treatment does not appear in it. Neither does stereotypical views, nor propensity, nor the spotlight the chapter says it is shining. What survived the journey from the study to the General Assembly is the number, 94 per cent, and the word that explains it, chivalrous. The finding that boys face discriminatory treatment did not make the summary.

The chapter does not bury it, either. Its second section carries a heading of its own, listed in the study's contents: Discrimination against Boys. Under that heading the study writes a sentence that almost nobody in this field writes down.

That is the argument this campaign exists to make, printed inside a UN study, under a heading, with footnotes. The research addresses discrimination against girls, not against boys. It is not written as an accusation. It is written as a description of the literature the study had just finished reading. The chapter goes further and reproduces, at length, an academic's judgement that the children's rights movement itself avoids the question, a passage the study chose to quote rather than to answer.

The same attenuation shows up in what each document says about violence. In the note there are twenty sentences mentioning violence: eighteen about children generally, two naming girls, none naming a boy, a male, a man, him or his. In the full volume there are 450, and nine of them do name boys. So the study is not silent. It records a boy beaten with a wooden stick until he nearly lost an eye, and it records this:

Read the first clause again. The study is not only reporting a harm to boys, it is reporting that less is known about it, which is the entire argument of this website written by the people who did the counting. And that sentence, like the discriminatory treatment sentence, is not in the note either.

So paragraph 38 is not the balance it reads as, and the reason is now visible in three stages of one study. The 804-page volume says boys face discriminatory treatment and that less is known about violence against them. The 23-page note to member states keeps the 94 per cent and drops both. The recommendations chapter inside that note drops the word boys altogether. Each step is shorter than the one before, which is what a summary is for. But at every step it is the same half that goes.

I want to be careful about what this does and does not show. Nothing here says girls are not at particular risk in detention; the volume documents that at length and in more detail than it documents boys, and some of that gap is the study telling you honestly that the research does not exist. Nothing here says anyone dropped anything on purpose. I do not know why the sentence went, and neither does anyone who has not seen the drafting. What is checkable is that it went.

Where the finding goes, which is nowhere.

A study like this ends by telling states what to do. In A/74/136 that is chapter VIII, running from paragraph 98 to the end of the document, about sixteen thousand characters of general and situation-specific recommendations. I extracted it and counted whole words in it.

The chapter is not silent on sex. It contains the word girls once, in a recommendation about children associated with armed forces and armed groups, asking that assistance take account of their specific situation and needs and provide tailored measures. That recommendation is good. It should be there. It is also, precisely, the thing that did not get written for the 94 per cent.

The number is in the study. The number is not in the ask.

There is one subject where the two documents can be laid directly on top of each other, because both examine it and both arrive at a recommendation about it. Diversion is the decision to deal with a child outside the courts, and it is the difference between a criminal record and none. Paragraph 36 of the note reports the disparity itself.

The full study then says what should be done about it, in one sentence, naming both.

The note’s recommendation on diversion reads, in full: States should prioritize restorative justice, diversion from judicial proceedings and non-custodial solutions. The word equally is gone, and so are boys and girls. The word equal does survive into that chapter, once, in a recommendation that girls associated with armed forces be guaranteed equal access to rehabilitation and reintegration. That recommendation belongs there. So did the other one.

I want to be careful about what this does and does not show, because the easy version of this paragraph is wrong. The recommendations are largely written for all children, and boys are children, so they are covered by construction. Nobody is excluded. That is a fair reading and I am not going to pretend otherwise. But universal coverage is what the rest of this site keeps running into. A recommendation that covers everyone assigns nothing to anyone. When the study wanted a group's needs taken into account specifically, it knew exactly how to say so, and it said so for girls in armed conflict. It had the same sentence available for the group that is 94 per cent of the population it had just measured, and it did not write it.

This is not a failure of noticing. The noticing is the best thing in the document. It is a failure of the step after noticing, which is the step this campaign is about.