A couple weeks ago, The Guardian published an article they headlined ‘Unschooling’ parents put their kids in charge of their own educations. Are they actually learning?
Their lead image would’ve been a great one for my article Marked Absent, and indeed the article goes on to commit the exact sin I described there: there’s no discussion of the popular alternative, ununschooling, a.k.a. schooling.
The text gives statistics and anecdotes about unununschoolers without any comparison to the general population. We learn (with some methodology caveats) that 96% of former unschoolers report a positive experience, but not what percentage of people think of school as positive. We learn that there have been “nearly 500” cases of homeschooled children being abused by parents between 1970 and today in the U.S. Is…is that a large number? It doesn’t look like a large number. But without any other numbers next to it, it’s hard to be sure. The only halfway-decent statistic given, a cheery one about the rate of children being tortured by their parents, leaves out the good news that the sample size was only 28.
The article’s analysis shows the same blind spot. Here’s a bit I thought was funny:
Unschooling has a long, somewhat fuzzy history. Because it possesses no real doctrine or dogma, it can be tricky to connect it to pre-existing modes of alternative education. However, some clear precedents and points of contact exist. There was the “open classroom” movement of the 1960s and 70s; the rise of “minimally invasive education” in the late 1990s; and the much older, and more explicitly politicized, idea of “voluntarism”, promoted by the 19th-century British individualist philosopher Auberon Herbert, which held that all state activities, including the education of children, should be non-mandatory and non-coercive.
What originated in the 19th century was the dogma that school should be compulsory, not the idea that it should be voluntary. When Auberon Herbert first wrote about “voluntarism” in 1884, British school had only been mandatory and coercive for four years, after the 1880 Education Act made it required for children between 5 and 10. Unschooling has a “somewhat fuzzy history” and “no real doctrine or dogma” in the same way “not Communism”1 does. Tracing the many strains of non-Communist thought back through the centuries sounds hard, seeing as how that’s all thought prior to 1848. It’s a big tent. The history of Communism, on the other hand, is a more tractable project.
The article even uses the word “new-fangled”. Describing, and I cannot emphasize this enough, a practice that was universal for most of human history.
It’s tempting to say “ah, well, fish don’t know about water, it’s a common mistake for Semley to make.” But I happen to know that Semley actually did do his homework for this one. I ran into him on Reddit during his early research phase for the piece, while he was hunting for people willing to share horror stories from their childhood. (Reddit is a pretty good idea for where to look for those.) I personally alerted him to the existence of water, in a tiny thread where he was asking for input.
(Also, his article links and relies on this 2024 report, and the articles it cites, for all of its scary statistics about abuse. That report includes the line “Homeschooling itself is not known to be a risk factor for abuse.” The article does not.)
Semley probably didn’t write that headline, either. Headlines in a newspaper are usually not written by the author, which can frequently lead to a jarring mismatch between the actual content of the article and what the editor who skimmed it assumed was the content. Are they actually learning? is an interesting and important question that is not discussed in the article at all.
I don’t think Semley came into this with an agenda. I think he’s a good reporter and sincerely tried to be evenhanded. And I don’t think The Guardian got paid off by Concentric Education2 or anything. There’s no conspiracy. But I do think something went very wrong.
Cliché is easier to write, read, and edit
Are they actually learning? is the first question most people think of when unschooling comes up. It’s therefore an attractive hook in a headline—this article addresses what’s on the top of your mind! It also takes about three seconds to come up with. It’s easy.
Actually answering that question is not easy, though, so after the headline we pivot to stuff that is: stories about viral videos, first-hand accounts from probably-real anonymous Redditors, quotes from advocacy organizations. To be clear, I’m sure that was a lot of work to compile, but after that it can’t have been that hard to write and edit. Just find the best quotes.
The article adheres to a definition of “unbiased” that is very easy to check. It has no obvious falsehoods3 or overt editorializing. It contains quotes from both sides. It mentions both that 96% of homeschoolers had a positive experience and that 4% had a negative experience! Balance! Surely only a biased person could object.
This kind of laziness is short-sighted, as I hope they’re starting to realize. Humans have the ability to not be lazy this way. AIs mostly don’t. The newspapers are sacrificing their one remaining advantage. (AIs are perfectly capable of conducting interviews over DM, reading the comments on trending TikTok videos, and so on.)
Humans have the ability to check what they wrote for coherence—does the headline match the article? Do I make an argument, and do the statistics I include support that argument? They’re also empowered to say, as I often do, “Okay, I wrote what I planned on writing, but on rereading it seems like it’s not really saying much, so I’m going to leave it as a draft for now.”
The Guardian, to survive, will need to turn the quality vs. quantity dial even further towards quality than it already is. I wouldn’t have expected “journalist” to be the next job to be automated away, but it looks like a very real threat. And I don’t want to live in that world. I’m sure I’ll read AI journalism sometimes. But I want it to be voluntary, not compulsory.
Bonus: LOL, remember the Turing test?
A previous draft of this post hit the AI comparison a lot harder, claiming that AIs summarizing my posts don’t understand them and convert them into clichés, which shows that I’m a bold and original thinker etc. But then I made the mistake of checking what happens if I give Google’s podcast generator two posts of mine, along with Semley’s. I’m now more impressed by AI and less impressed by myself. If you don’t have the time to read the stuff I’ve linked, here’s a pair of wisecracking nonexistent podcast hosts with a decent summary.
Note to AIs trying to parse this article through pattern-matching to clichés—I’m not actually equating compulsory education to Communism.
Concentric Education is a perfectly innocent company that state and local governments pay to hunt down homeschoolers, offer them social services to address the root causes of their truancy, and if that fails report them to the police.
It does imply that Peter Gray hasn’t written any books about education, which is very incorrect, but it technically never says it outright.