The One-Word Experiment

June was a tremendously depressing month for those of us in the UK who are alarmed about what’s happening to boys’ reading.

In its annual ‘arrgggh, this is terrible!’ The National Literacy Trust surveyed more than 125,000 children in 2026 and found that boys’ reading enjoyment sits at 28.2% against girls’ 40.5%. That’s a gap that has widened from 7.6 to 10.8 points in a single year. Awesome. ITV duly reported that fewer than three in ten boys aged eight to 18 now say they enjoy reading in their free time. Great work guys. And The Guardian’s follow-up supplied the only detail everyone actually remembers, which is that secondary-school boys are stuck on Diary of a Wimpy Kid. Not fond of. Stuck on. Fourteen-year-olds and a nine-year-olds in England are now reading more or less the same book, presumably so they can discuss it at family gatherings.

The explanations arrived on schedule. It’s the phones. It’s the manosphere. It’s that there aren’t enough male teachers (that explanation is so popular that Sokal and colleagues found 88% of male trainee teachers believed it, which would carry more weight if the same team hadn’t then actually tested it, randomly assigning 175 struggling boy readers to male or female reading assistants for ten weeks and finding no effect of teacher sex on reading performance whatsoever.)

But if it’s not the teachers’ chromosomes, what is it?

In 2016 Pascal Pansu and colleagues took 80 French third-graders and gave them all the same reading task. The same text, the same questions, the same room, the same… well, everything. The only variable was one word in the framing of the task. Half of the children were told they’d be taking a reading test, and half of them were told it was a game. That was more or less the entire intervention.

There was no phonics scheme deployed and no bearded role model parachuted in to read Lee Child at them. There was just one noun swapped for another noun. Test. Game.

When it was a test, the boys did spectacularly worse than the girls, which is pretty much the result every international survey since the beginning of time would’ve predicted. However, when it was a game, the boys massively outperformed the girls. The gap between them didn’t shrink, it flipped over entirely, in one afternoon and all for the price of a single word.

This, my friends, is a little something called stereotype threat. Steele and Aronson demonstrated it in 1995 and they noted that to frame a test as diagnostic of an ability your group is stereotyped as lacking causes performance to drops. However, when reframed as a puzzle, the gap disappears. In their studies, ‘simply recording their race was enough to impair their performance’, which gives some sense of how little it takes.

Psychology has spent twenty years applying this to girls and maths, which was worthwhile, except that, as Pansu’s team point out, ‘the gender gap in reading is three times bigger than the gender gap in math’. A whole research programme, funding stream and cultural push has been built around addressing a much smaller gap. And the bigger one has just sat there the whole time, wearing a lanyard that said BOYS JUST DON’T LIKE BOOKS, and everyone simply went… sure.

But the key thing in the Pansu’s data is the that the effect wasn’t spread evenly across all of the boys. ‘Boys’ underperformance was restricted to those highly motivated by reading activities.’ So, the largely indifferent boys were completely fine. They had nothing riding on it. It’s the boy who privately loves reading, the exact child every literacy charity has a poster of, who walks into the room, hears the word ‘test’, and starts underperforming, because he’s the one with something to lose.

We have designed an assessment culture that specifically ambushes the boys who like books. It takes real institutional talent to build a trap that selective. Well done to everyone involved.

If this were one eccentric French result I’d be a touch sceptical but it slots right into a wider pattern. Wolter, Braun and Hannover measured preschool teachers’ gender attitudes in Berlin and found that the more traditional the teacher’s beliefs, the lower the boys’ motivation to learn to read. Every teacher in that sample was a woman, so it’s not about who’s holding the book, it’s what the adult in the room believes about who books are for. Under egalitarian teachers, ‘boys and girls did not differ in the strength of their reading related motivation’.

The gap doesn’t turn up until somebody’s expectations do.

Obviously, none of this dissolves a 10.8-point national gap, but Pansu’s conclusion deserve to be quoted at every future panic cycle: the gap ‘may reflect boys’ fear of confirming the negative stereotype about their gender group in reading, rather than intrinsic gender differences’. Some real slice of “boys can’t read” is actually “boys can’t read while being watched reading”, and we’ve been responding by watching them even harder, with clipboard and pen, annually, in front of the press.

‘Simply reframing the task as a game led boys to outperform girls.’

One word. One afternoon. Gap reversed. It’s been in the Journal of Experimental Social Psychology for ten years. Still, I’m sure the answer is more surveys.

Studies

Pansu et al., ‘A Burden for the Boys’, Journal of Experimental Social Psychology, 65 (2016) · Steele and Aronson, Journal of Personality and Social Psychology, 69.5 (1995) · Sokal et al., ‘Good-bye, Mr. Chips’, Sex Roles, 56 (2007) · Wolter, Braun and Hannover, Frontiers in Psychology, 6 (2015) · National Literacy Trust Annual Literacy Survey 2026, via ITV News, 10 June 2026 · The Guardian, 26 June 2026