Just how harmful is porn?
A protest in 1979. Credit: Colin Davey/Evening Standard/Hulton Archive/Getty   

Exactly what percentage of men watch porn isn’t clear, but it’s most of them. In 2009, a University of Montreal researcher tried to do a study on how pornography affects men, and couldn’t find any men in their 20s who hadn’t used it to use as a control group. This study found 98% of men (and 73% of women) had used it in the last six months. This one found 87% of men report using it for sexual purposes, along with 31% of women.

That context is important in the light of the weirdly under-reported fact that, some time in the next few months, the Government will require all pornographic websites to demand age verification of their users.

Full disclosure: I was ready to write this piece as a full-throated blast against the idiocy of the whole thing. Rowland Manthorpe, Sky’s technology correspondent, has done exactly that, and it’s pretty convincing. In the end, I’m going to be a little more circumspect, but only a little.

There are good reasons to be sceptical. One, the verification process screams “bad idea”. It will mean giving a third-party verification provider your credit card details, or your driving licence or passport. Some of these providers will require you to buy a “porn pass” from a local newsagents, which I suppose will bring back a nostalgic thrill for the over-50s who remember hiding a copy of Razzle underneath their Radio Times as they approached the till. One is run by MindGeek, which, incidentally, also owns PornHub, YouPorn and RedTube. Fox in charge of the henhouse, and all that.

This runs a big risk of creating a database of sexual preferences and porn habits, in the hands of providers with varying abilities to protect it. This sort of thing has happened before: the “dating site for married people”, Ashley Madison, was hacked in 2015. It became something of a feeding frenzy for extortionists, and there were reports of suicides linked to the leak. A lot more people use porn than use skeevy hookup sites for affairs. And it’d be OK for those of us with vanilla proclivities, but for people with unusual or hidden kinks, it could be awful.

More importantly, though, I was – and am – concerned that the evidence for the harms of porn is enormously overstated. Those fears were far from allayed when I looked at the impact assessment the Department for Digital, Culture, Media and Sport put out to justify the legislation.

For one thing, it quotes a shoddy NSPCC/Childline poll which was, not to put too fine a point on it, utter bullshit. It’s offline now, but it made frankly ludicrous claims, such as that 10% of 12- to 13-year-olds were “addicted to pornography” (“pornography addiction” is not a recognised disorder and it’s hotly disputed whether it exists at all). The very next day, Sajid Javid announced the age-verification legislation.

The whole impact assessment, in fact, is seriously dodgy. At one point, it says the Government believes that “there is sufficient expert opinion that pornographic content can lead to harm to people under 18, whether or not this relationship is causal or correlational”. Which is … odd. If a relationship is correlational but not causal, then it doesn’t lead to harm. That’s what the words “causal” and “correlational” mean.

There’s more stuff like this. It says, in its second sentence, “One in ten UK visitors to adult sites in May 2015 were children.” There’s no reference given, but its first footnote is a (broken) link to something called “For Adults Only? Underage access to online porn”, a report for ATVOD, the TV-on-demand watchdog. When you track down the working report, it says that, in December 2013, one in 20 visitors was “underage”.

There is a DCMS report that looks at ComScore stats for May 2015, but that stat is nowhere to be seen. I honestly don’t know where they got it from. Later, they say that “In other words, one in 10 (13%) of children aged 6-14 visited a porn site in May 2015” – a totally different stat – so I think they just got confused. If nothing else, it’s an object lesson in the value of clear referencing.

I want to be charitable, though. The Government’s impact assessment is embarrassingly bad, but that doesn’t mean that there’s no impact to assess. I spoke to Dr Victoria Nash of the Oxford Internet Institute: she took part in a major review of the evidence in 2015. She said that they found that establishing the evidence of harm is extremely difficult. “There’s a really limited evidence base,” she said. “There are many studies looking at impact on young people but it’s very hard to draw strong conclusions from them.”

You can’t do controlled experiments on the effects of porn on children – take a thousand nine-year-olds, give half of them access to XHamster, and see if it warps their tiny minds – for obvious ethical reasons. So it’s all correlational research, from self-reported surveys. If you find that kids who watch porn are more likely to, say, have negative attitudes to women, you don’t know if the porn causes the attitudes, or kids with those attitudes are more likely to watch porn.

She says that there is better evidence that some young people, when you ask them, express distress at seeing inappropriate images online. But, again, she says, it’s hard to know whether those images are actual porn, or whether they’ve seen “an episode of Game of Thrones”, because you have to ask them age-appropriate questions. But, she says, “that is the area where the evidence for harm is greater”.

Nash thinks that the age-verification system isn’t obviously terrible. “Insofar as it is already illegal” for under-18s to view this stuff, she says, “it’s a proportionate response. It should make it much less likely that kids will stumble across porn.” The number of kids who do stumble across porn, according to her research, is much lower than the shock-horror numbers the Government impact assessment was touting – quoting ATVOD, it says “around 6% of UK 6-15-year-olds”. But it’s not negligible.

You may think that adults shouldn’t be allowed to view porn either, of course. It’s a legitimate viewpoint, although I will point out that there is no evidence that I am aware of that, as was the fear a few years ago, porn use leads to sexual violence. If anything the correlation is the other way around – countries and states with higher porn use experience lower levels of rape and sexual assault. It may be causal, but the author of the study I just linked to says that it’s probably just because countries with liberal porn laws also tend to be more progressive generally. There was also a recently released study which found that once you control for confounding factors, there’s no link between porn consumption and adolescents’ psychological wellbeing, even in girls.

And the trouble is, when you try to prevent people from doing things they want to do, is that you don’t really stop them, you just drive the behaviour underground. You can see this with drugs, where countries with stringent anti-drug laws do not seem to have any less drug use than the rest, but have more problems associated with drugs. Nash is worried that the age-verification law will push older teenagers – the ones who really do want to look at porn, and for whom it is entirely developmentally normal to want to do so – into weird dark-web sites, or using VPNs, or just grim subreddits.

Josh Grubbs, a professor of clinical psychology at Bowling Green University in Ohio, who studies internet porn use, told me that “banning people from doing a behaviour that they are going to do anyway sets them up for failure. This is the equivalent of abstinence-only education, which we know is associated with more harm than good.”

So this block may prevent a small but non-negligible number of young children seeing images that upset them. It’s hard to tell if it will do any good beyond that. But it will definitely cause a very large number of adults to have to jump through inconvenient and embarrassing hoops, and put their data in somewhat risky places, to do something legal. And it may push people towards parts of the internet we’d rather they avoid. If we’re going to do that, I’d rather it was done on the back of an impact assessment that understood what the word “correlation” means.