Sex and the Tories

Work, the curse of the blogging classes, has temporarily taken over my life at the moment, but I reckon I’ve earned a bit of break for some timely Daily Mail bashing, and your dumb-ass headling for today goes like this:

Sexually charged shows such as Sex And The City and Friends to blame for rise in teenage pregnancy

Wow, not even quotation marks around the ‘to blame’ – the compilers of the daily Idiot’s Guide (© Charlie Brooker) must be feeling confident today.

Now, I should point out here that headlines like this one always generate the same thoughts. What I’d really like, right about now, is for a kindly cricket bat manufacturer to make me a very special bat with the words ‘Correlation Does Not Imply Causation’ embossed on the blade, just so I can nip round to the Mail’s offices and give one their journalists – Barry Wigmore, in this case – a quick lesson in basic logic.

And the lesson, would go something like this:

‘CORRELATION’ [WHACK!]

‘DOES’ [SMACK!]

‘NOT’ [CRACK!]

‘IMPLY’ [SMACK!]

‘FUCKING’ [WHACK! WHACK!]

‘CAUSATION’ [WHACK! SMACK! WHACK-WHACK-WHACK-WHACK-WHACK-WHACK… SMACK!]

Think in terms of Robert De Niro (as Al Capone) in ‘The Untouchables’ and the mental picture that forms is going to be pretty much on the money.

We’ve been at this crap for how long?

Since the 1960’s and the emergence of the doyenne of blue-rinse busybody tendency, Mary Whitehouse?

And in all that time, the amount of solid evidence that TV, video and computer games, rap, hip-hop and heavy metal (and anything else you’d care to name) actually has an adverse influence on teenagers comes to a big fat zero… zip, nada, nowt, nothing and absolutely fuck all.

So even without reading the rest of the article, we can be sure of two things –

1. Whatever the reserach is that the Mail is harping on about, the one thing it absolutely doesn’t prove is the existence of any kind causal link between watching Friends/Sex and the City and teenage pregancy rates, and…

2. The Mail is talking bollocks – as usual.

So, let’s disposing of this crap, and to begin with we’ll deal with the source – the RAND Corporation, which the Daily Mail describes as ‘a respected non-profit research group’.

Let’s fill in a few historical blanks here.

The RAND corporation was set up in the 1940’s by the US Airforce (and Air Force General HH Arnold), around around two-thirds of its research activity is directed towards military and national security issues, with the remaining third branching out into areas such as health, education, civil and criminal justice, labour and population studies, and international economics.

It’s one of a number of inter-related American charitable foundations, including the Ford, Rockefeller and Carnegie Foundations that has, historically, mixed and matched charitable activities with the covert promotion of US foreign policy objectives and an integral component of the US Military-Industrial Complex, all of which is neatly reflected in some of the prominent names associated with it.

Former US Treasury Secretary, Paul O’Neill, was RAND’s chairman from 1997-2000 and returned to its Board of Directors after his resignation from the Bush administration in 2002.

Donald Rumsfeld racked up two stints as RAND’s chairman, from 1981-85 and 1995-6.

Condoleeza Rice is a former RAND intern and served as one of trustees between 1991 and 1997…

And, amongst the other RAND alumni with ties to the outgoing Bush administration you’ll also find Lewis ‘Scooter’ Libby, Zalmay Khalildzad (current US Ambassador to the UN) and David Chu, the current United States Under Secretary of Defense (Personnel and Readiness).

Oh, and I mustn’t forget to mention Andrew W Marshall, who was first appointed to the position of Direction of the United States Department of Defense’s Office of Net Assessment in 1973, by Nixon, and has been reappointed to that position by every US president since. Marshall is not only the big wheel in US foreign policy analysis but has a reputation for developing ‘new talent’, and amongst his former ‘star protégés’ you’ll find not only Donald Rumsfeld, but also Dick Cheney and Paul Wolfowitz.

So, to sum up, RAND is independent is pretty much the same sense that Fox News is ‘fair and balanced’.

Coming to the meat of the research, the basis for the Mail’s claims is reported as follows:

The study of more than 2,000 American youngsters between 12 and 17 is the first to directly link programmes such as Friends and Sex And The City to pregnancy.

It warned: ‘One problem is that these and similar programmes glamorise sex while hardly mentioning its downsides, such as pregnancy and sexually transmitted diseases.’

Lead researcher Dr Anita Chandra said: ‘Sexual content on television has doubled in the last few years, especially during the period of our research. We found a strong association.’

Dr Chandra, a scientist with the RAND Corporation, a respected non-profit research group, said even cartoons with a sexual content can have the same effect.

And Dr Chandra arrived at these conclusions how…?

Well, the abstract to the research paper, published on RAND’s website, gives us the following information:

Does Watching Sex on Television Predict Teen Pregnancy? Findings From a National Longitudinal Survey of Youth – Chandra A, Martino SC, Collins RL, Elliott MN, Berry SH, Kanouse DE, Miu A Pediatrics, Pediatrics 2008;122: 1047-1054

There’s a couple of very obvious points here.

One is simply that a question like ‘Does Watching Sex on Television Predict Teen Pregnancy?’ is hardly the most illuminating question that the researchers could have set themselves – at best its a preliminary/scoping question of the kind you’d use to figure out whether its even worth investigating a subject any further and not a question that’s likely to deliver any real insights.

And the other point rests on the second part of the paper’s title, ‘Findings From A National Longitudinal Survey of Youth’, which tells us that this is a statistical/epidemiological study based on data-mining the results of a general survey, and not a detailed observational study.

So, if you understand the value and limitations of different research methodologies then, right from the outset, the title creates nothing more than a few fairly low expectations as to the significance of the study – at best its going to point to a possible correlation between what young people watch on TV and aspects of their sexual/social behaviour, but the chance of its telling you anything meaningful as to why that association exists and how it operates are near enough zero as makes no difference.

In short, this seems no more than an ideal contender for next year’s Ig Nobel Prize.

Objective
There is increasing evidence that youth exposure to sexual content on television shapes sexual attitudes and behavior in a manner that may influence reproductive health outcomes. To our knowledge, no previous work has empirically examined associations between exposure to television sexual content and adolescent pregnancy.

I do love that last bit about there being no previous work looking at associations between sexual content on TV and teenage pregnancy – all of which sounds quite impressive but is entirely meaningless. There’s actually no shortage, at all, of research looking at (and for) evidence that films, television and music influence the behaviour of children and adolescents in all manner of different ways, and none of that research has ever provided any definitive evidence of a causal relationship. If previous studies have failed to establish a causal link between what adolescents watch on TV and their sexual attitudes and behaviours, generally, then what makes the researcher think that correlating to teenage pregnancy will turn up something that no previous study has ever found?

The truth is that no ones ever bothered to come at this issue from this particular angle because, in research terms, it tells us nothing we don’t already know – the only reason why anyone would look at this issue in this way is for political reasons and a cheap headline, as willingly furnished by the Daily Mail.

Methods
Data from a national longitudinal survey of teens (12-17 years of age, monitored to 15-20 years of age) were used to assess whether exposure to televised sexual content predicted subsequent pregnancy for girls or responsibility for pregnancy for boys. Multivariate logistic regression models controlled for other known correlates of exposure to sexual content and pregnancy. We measured experience of a teen pregnancy during a 3-year period.

Unfortunately for the researchers, this is all to no avail whatsoever because the one thing that this kind of research cannot control for or eliminate is the chicken and egg problem – does watching TV shows that include sexual content actually influence the attitudes of teenagers or do their pre-existing attitudes merely influence their choices of what programmes they do and don’t watch?

You can throw in all the multivariate regression models you like but you still can’t unpick that question from a statistical review of data mined from a longitudinal survey, and without dealing with that question, one way or another, you can’t claim a causal relationship.

FFS – this is first year undergraduate stuff – correlation does not imply causation.

Results
Exposure to sexual content on television predicted teen pregnancy, with adjustment for all covariates. Teens who were exposed to high levels of television sexual content (90th percentile) were twice as likely to experience a pregnancy in the subsequent 3 years, compared with those with lower levels of exposure (10th percentile).

And?

I’ve explain the problem above – there’s no way of sifting out the alleged influence of television programmes from pre-existing attitudes to sex and sexual behaviour and the extent to which they influence viewing choices and any researcher who tries to claim otherwise on the back of statistical study is talking bollocks.

The only way to evaluate such a question would be to use teenagers as guinea pigs in an observational study.

Conclusions

Okay, let’s break this down…

This is the first study to demonstrate a prospective link between exposure to sexual content on television and the experience of a pregnancy before the age of 20.

So what!

You’ll note that the abstract only talks about a ‘prospective link’ rather than suggesting ‘blame’ or ‘causality’, so its clear that the full research paper, at least, recognises the limitations of the research, even if this is not apparent in the Mail’s typically boneheaded coverage.

Limiting adolescent exposure to the sexual content on television and balancing portrayals of sex in the media with information about possible negative consequences might reduce the risk of teen pregnancy.

And it might have no effect at all, if there is no causal relationship – which seems likely from just about every other piece of similar research conducted over the last forty years.

Even if we allow for the possibility that programmes like Sex and the City exerting an influence over the attitudes of some teenagers, ‘limiting their exposure’ to sexual content is easier said than done, unless you’re going to resort to outright censorship – and even if you do take that route then there’s no shortage of alternative sources of information that young people can readily gain access to, some of which (porn) provide an even more stylised and unrealistic portrayal of sex and sexuality than anything you’ll find in Sex and the City.

Parents may be able to mitigate the influence of this sexual content by viewing with their children and discussing these depictions of sex.

Perhaps – although its by no means established that there is any kind of influence to mitigate.

As a final conclusion, encouraging parents to discuss sex with their teenagers is a pretty good idea…

….and providing mandatory, high quality sex education in schools with no parental opt-outs is a even better one.

If research of this kind has any policy implications, for all its manifest limitations, then it shows a clear need for investment in high-quality sex and relationships education in schools.

It can’t be stated often enough that prohibition and censorship just don’t work, and never have done.

In fact, the slightly peculiar mores of television censorship, as it applies to sexual content, could be doing more harm than good, if its genuinely thought a more honest and realistic portrayal of sex and sexuality may help shape attitudes in a more positive direction. When it comes to nudity, while pretty much anything goes for women short of a full-on ‘split beaver’ shot, even the merest hint of a man with a stiffy is strictly off limits, so what teenagers will very rarely, if ever, see on TV, is couple breaking off foreplay in order to hunt franticly through the drawer in the bedside table for a condom, let alone anyone actually putting one on before getting ‘back to business’.

Its often the case that its the censorship of sexual content that makes for the unrealistic depictions that we see on TV.

Of course, no Daily Mail bad science article would be complete without a quote from a mouth-breathing Tory rent-a-gob MP, and as the topic is sex that means yet another chance for Nadine Dorries to gain a little more value for the three grands worth of public money she spent on the services of a PR company earlier this year.

Tory MP Nadine Dorries said last night: ‘It would be interesting to see if a similar study in the UK revealed a trend. Information such as this empowers parents when making difficult decisions as to what they do and don’t allow their young daughters to watch.

‘Last year we saw girls as young as 12 aborting. Any information which could help stop even one child aborting her child has to be welcomed.’

Dorries, who gave us the benefit of this sparkling insight, earlier this year:

“We’ve had 11 years now of liberalised attitudes to sex education in schools. This has to be a contributory factor in the higher rate of abortions.”

…remains entirely oblivious to the fact ( of course) that, across Northern Europe, its the countries that have the most liberal attitudes to sex education in schools (and mandatory sex education, of course) that have dramatically reduced both teenage pregnancies and teenage abortions far more rapidly than the UK while, over the big pond, where the influence of god-botherers and amateur moralists is even more marked than it is over here, and where millions of dollars have been spent on abstinence-based sex education that have now been proven to be ineffective, both the rates of teenage pregnancy and abortion are running at almost double that of the UK.

Last month, Peter Cuthbertson tried to leverage a fairly dull comment posted at Liberal Conspiracy into asking ‘Why the left ignores all data on the fatherless family‘ over at Tory Home’s Centre Right blog, eliciting the usual (if small) run of utterly moronic comments, like these particular gems…

[Martin Jee]

Why?

Because those lefties are the Champagne Socialists and the leafy suburb Lib Dems who out of middle class pretentiousness and a sense of their own superiority believe the political argument on gender roles has not moved on since the 1970s.

The unfortunate fact that they are in their 50s and as such are now in the positions of established power mean that we all have to suffer.

[Man In A Shed]

Johnathan Singh makes the key argument against the left. They are callous because they do not act based on evidence, but on a desire for self gratification.

Its how they can talk of justice and helping people, but show themselves to in fact be such spiteful and despicable charters.

We don’t push this critic nearly enough.

[RichardJ]

The liberal-left are allergic to the idea of judging people or saying some sorts of relationships are “better” than others. That is the crux of the matter. They think it’s “nasty” to point out the blatantly obvious.

Then you have the nutters on the far-left who are opposed to the idea of the family based on Marxist dogma.

I can’t speak for others on the liberal-left, but the main reason why I, personally, switch off when the right when its starts banging on about data and evidence should be self-evident from my comments on this particular piece of research…

It’s because I can read.

Seriously, even on those fairly rare occasions that the right do manage to alight on some fairly interesting evidence, their analysis of the data and the conclusions they draw from it invariably turn out to be nothing more a bunch of dishonest and tendentious crap punctuated only by the distinctive odour of burning straw-men and the kind of basic logical errors that would make a first year philosophy student cringe with embarrassment.

Apples are not oranges (so you can’t compare them). Correlations do not imply causation and you can’t simply ignore or downplay evidence that doesn’t fit in with your pre-conceived ideas about ‘ideal’ family structures.

For example, there’s the evidence relating to children growing up with same-sex (and primarily lesbian) parents, which shows that, all other things being equal, they thrive just as well as children brought up in ‘traditional’ two-parent families (other than that they tend to have slightly better social skills) and show no differences either in adopting conventional gender roles nor in the likelihood of being gay/lesbian themselves.

In IDS’s policy study on family breakdowns, that evidence barely gets a mention and where it does crop up it’s referenced in such a way as to imply [falsely] that two-parent same-sex households fall somewhere in between the so-called ‘traditional’ two-parent family and lone parent families, in terms of outcomes for the child, when – as I’ve already noted – the only significant difference between this ‘alternative’ family structure and the ‘conventional’ family is that lesbians (and there little or no data on gay households as yet) seem to be a little better at turning out children who’re more confident in a social setting.

Why this kind of evidence is routinely either omitted or downplayed by the right should be obvious – what it suggests is that there’s nothing special or magical about fathers and/or male role models. Growing up with two parents is better, statistically, than growing up with only one, but the benefits of the former stem from practical differences, i.e. the stability of the family home, its (typically) improved economic circumstances and the fact that, with two parents to share the load, children growing up in that kind of family will tend to receive more attention and support than children growing up in a lone parent family, particular one where their sole parent also holds down a job.

There’s nothing actually that special about fatherhood but for the fact that having a parent of each gender remains the norm, simply because the vast majority of people are heterosexual and its these people who have most of the families – just because something is biologically the norm doesn’t mean that this automatically invalidates alternatives or prove that the ‘norm’ is somehow superior to those alternatives. The very essence of what it means to be human is that we are more than just the products of our biological inheritence and, as Richard Dawkins (in ‘The Extended Phenotype’) and others have noted, the complex social and cultural environment in which we live is a powerful evolutionary force in itself.

Looked at with a critical eye, much of what the right tries to pass off as evidence-based policy making just doesn’t stand up to scrutiny.

  • Fundamental to the pointlessness of that survey is the inherent contradiction in the stated objectives that you quote. They say that there has been no work to examine the associations between sex on TV adolescent pregnancy, but right before that it says there is “increasing evidence” of such links.

    Where does this evidence come from if there has been no research? Apart from anything I have an aversion to so-called research that seems to pre-empt the results as part of its objectives.

  • Anonymous

    RAND is about as much a part of the Vast Right-Wing Conspiracy as (say) Chatham House here is. Part of the establishment, yes. And by that token, committed to the governing orthodoxy of the day – especially in military and foreign policy matters. The fact that it is a Government contractor means that surely Carter or Clinton would have cut off the lifeline if it were part of the Right.

    The fact that Rumsfeld and Rice passed through (along with, e.g., Walter Mondale) is no more evidence of it being a Right-wing front than their attendance at Bilderberg meetings might be that the Bilderberg Group is a lizard-conspiracy front. IMS, the reason why Herman Kahn and others left to form the Hudson Institute was because RAND was too constrained for the agenda they wished to pursue. (Actually, one might say that the correlation of Right-wing people being involved with RAND should not be taken as evidence of any Right-wing causation on the organisation’s part. But that would be churlish.)

    And so, from that, while the Daily Mail is guilty of jumping to rash conclusions, this post is guilty of jumping to the opposite conclusion. Leaving aside RAND’s biases or otherwise, the article was published in a respectable peer-reviewed journal. The poor researchers might just have been interested to find out whether there’s any evidence for a well-worn theory not often subjected to investigation – surely a good thing, whether to confirm or refute the theory. The criticism of them for using NLSY data rather than primary observational data as somehow ‘data mining’ would basically count out huge swathes of social research.

    Quantitative data is expensive to collect, and that’s why most countries have a number of key sources which are open to other researchers to use (in this country, think of BSA, BHPS, Census, BCS, MCS, and so on). Not all such secondary analysis is good, true. But secondary analysis is not in itself questionable, as the post suggests; indeed, the suggestion borders on the anti-scientific. Data dredging is a risk, but the very fact that the analysis is of secondary data does not prove that it has occurred; normally, you look for them to

    Now, from the abstract bits you’ve quoted, they do not breach the causation-correlation fallacy as such. They say:

    “Exposure to sexual content on television predicted teen pregnancy… Teens who were exposed to high levels of television sexual content … were twice as likely to experience a pregnancy.”

    Well, “predicted” does not say “caused” and “were twice as likely” does not mean “will”… The results are simply articulating the findings from the analysis – which, from the abstract (if only we could see the full article, we could say more) we’ll probably have to take on face vale as being true – it did get published, after all.

    You’ll note they go on to say the study “demonstrate[s] a prospective link” – not a direct and causal relationship, a prospective link. Forgive them for not making the obvious (as you say, first-year undergraduate) point that one does not infer cause from correlation… They probably thought it was obvious, and can be inferred from ther caution in describing their results, especially given that the abstract suggests that they’ve controlled for other variables and still show the relationship.

    As a broader question though, just because the causation can’t be inferred, it doesn’t mean that the antithesis has been demonstrated – if anything, those opposing the “Smut matters” theory now have more to explain as to why it’s not true. The “Smut don’t matter” theory takes a blow (so to speak) here.

    Further, proving causation is terribly difficult in all social science research – and in fact, rarely is proven. Typically, we go from theory to data and back again, to eliminate other possibilities. In this case, there’s a reasonable theory which now has some evidence to support it. That’s probably progress, even if there’s a lot further to go.

    And in case you’re wondering: I don’t believe the Daily Mail “smut matters” theory in its pure sense, although I doubt the researchers do either. My guess is that sexual TV content has an impact on mores, just as it is itself also driven by those changing mores. Which makes causation even more difficult to understand.

  • Anonymous

    The hanging sentence wasn’t intentional – I meant to say that to detect data dredging we’d need to look at the full work, and see if they’d chopped and changed their model so that it had no sound basis in theory, and was just to find the right metrics to do the job for them.

    Incidentally, a quick Google found that the study, although housed by RAND, was funded by a grant from National Institutes of Health in the US. Again, hardly suggestive of the Right-wing fringe.

  • Steve

    Whilst you are bright enough to realize that correlation does not imply causation, (and that apparently disqualifies you for working on a British newspaper) it does not seem to have prevented you for falling for the ‘Guilt by Association’ fallacy.

    A large lump of your rant