Friday, 16 May 2014

Lies, damn lies, and statistics (

I am all in favour of more education, and evidence based policy and so on, but when you get an officially endorsed web site like misusing stats like this you have to wonder if there is a hidden agenda.

This stat could mean that 100% of 18 and 19 year olds have seen on-line porn and zero below 18 have seen any. It could mean anything. It depends on the age distribution of the sample, and all sorts. What is worse it is seen porn online, that could be anything from having seen page 3 of the sun on line over the shoulder of an adult at a PC, or some unwanted popup advert because they have not set sensible browser settings. It could mean anything.

As a stat it is completely meaningless. The age of people does matter - if there are 9 year olds actively watching online porn (or any porn) I would be very concerned. If we are talking of a 15 year old boy I would be totally unsurprised. If we are talking an 18 or 19 year old ADULT watching online porn then what is the problem at all?

Quoting that stat at the start of the video ruins any and all credibility of the rest of the video. What are they thinking?

Update: I was thinking more on this awful stat, and a couple more points come up, purely intuitively. 1. The taboo nature of the subject mean that even adults will deny seeing porn even when they have, but it is very unlikely someone would say they have when they have not. 2. The combination of both being taboo, and seen as wrong doing for a minor would mean children are far more likely to deny seeing porn than admit it. This means that the 57% who admit seeing online porn is probably a gross underestimate. 3. Even so, if 57% have done this terrible thing and society is not falling apart, that is a serious indicator that there is no actual problem to solve here. If the stat was 99% or 100% would we really not have to say "clearly this is perfectly normal and nothing to worry about"?

Thanks Alex for pointing it out to me.


  1. This is exactly my attitude to the government's porn filters, which seem to keep being pushed based on the apparently large number of people who watch porn. The government is supposed to act in line with the wishes of the public, so to my mind the more people who do something, the less of an argument the government has for trying to prohibit it!

    Also, whilst protecting kids is laudable, the internet doesn't really change a lot - kids had access to top shelf magazines long before the internet came along. And it seems odd that so much emphasis is placed on stopping people seeing porn whilst violence is apparently ok (watch your average soap opera, news report, etc.) Protecting young kids against stumbling across unsavory material seems sensible, but currently seems to be disproportionately swayed towards taboo material.

  2. The author is correct, and his concerns are valid. I am a statistician and a woman. I have worked for a 100% state-funded organization that supported the well-being of vulnerable children (not in the UK nor eastern hemisphere though).

    Internet pornography and child-inappropriate content is most effectively monitored at a parental and local community level. Using ISP's for this purpose, then invoking meaningless pseudo-statistics to support such actions, is troubling in its implications. I have seen so much less news from the UK since February 2014 (due to the laws put in place at that time?), than in the past. In fact, other than from Germany, it feels like a great beacon of information and human endeavor has been been dialed back to nearly nothing. Just to clarify, I refer to journalism and current events, not to the quantity of internet pornography.

    I worry much more about the harm done by child pornography (and other disturbing trends in Western society wrt children) than the effect of children being inadvertently exposed to adult men and women engaging in activities associated, even obliquely, with human reproduction and harmonious conjugal relations.