Different Year, Same Old Shit…

What better way could there be to start a new year at the Ministry than with a typically tawdry tale of statistics abuse featuring Nadine Dorries as supporting rent-a-gob and all brought to you, as ever, by the Daily Mail:

Teenage pregnancies triple in leafy suburbs in damning blow to Labour sex education campaign

Affluent middle-class areas are experiencing the sharpest rises in teenagers giving birth, figures reveal.

The number of teenaged mothers is rising in two out of three constituencies – and has almost tripled in some of the leafiest suburbs.

Teenage maternity is also rising in two-thirds of the areas already worst affected, despite being targeted by Government policies to tackle the problem.

The damning statistics will further-undermine Labour’s claims that is getting to grips with the issue through greater sex education and contraceptive use.

All complete and utter bollocks, I’m afraid, not that that’s anything unusual for the Daily Mail, but even by the Mail’s usual abysmal standards this one is particularly poor effort, all the more so for it appearing to have originated with a Shadow Minister, Tim Loughton MP, the Conservative Minister of State for Children.

The starting point for fact checking any story is, of course, to attempt to identify its source, which turns out to present no difficulties whatsoever in this case as the Mail managed to leave a convenient trails of breadcrumbs back to Loughton:

The Conservatives uncovered the figures from Parliamentary answers.

Tory spokesman on children Tim Loughton said: ‘Despite all the Government’s smoke and mirrors on teenage pregnancy the fact is that in most parts of the country the situation is getting worse.

‘Most of the areas that are already experiencing the biggest problems are seeing the number of teenage pregnancies rise.’

And, sure enough, on 16 July 2008, Loughton put the following written question to the Chancellor of the Duchy of Lancaster who, at the time, was Ed Milliband:

To ask the Chancellor of the Duchy of Lancaster what estimate he has made of (a) the percentage of children living in one parent families, (b) the number of live births per thousand teenage women and (c) the percentage of children born outside marriage in the last 10 years for which figures are available, broken down by (i) social class, (ii) income, (iii) race and (iv) geographic location.

By way of a reply (delivered by Tom Watson) Loughton received a response from the UK Statistics Authority referring him to the Authority’s standard statistical publications, one of which provides a breakdown of teenage conceptions and births by Local Authority/Unitary Authority but not by parliamentary constituency.

So, when the Mail makes the following claims about its source:

Last month official figures showed a 2.7 per cent increase in the rate of under-18 pregnancies in England and Wales last year. Britain already has the highest teenage pregnancy levels in Europe.

The latest figures uncovered by the Tories provide the first seat-by-seat snapshot of the number of teenage girls giving birth.

They show women aged under 20 gave birth to 42,300 babies in England and Wales in 2006

… then we already have several serious problems with the claimed evidence base.

The UK Statistics Authority does not produce a geographical breakdown of teenage births by parliamentary constituency, which means that the ‘seat by seat snapshot’ the Mail refers to is not official statistical data but a breakdown concocted by Loughton (or rather, by one of his bag carriers). It also doesn’t produce a breakdown by Local Authority for births to women under 20, only for under 18s, so its not at all clear exactly which age group(s) this ‘snapshot’ relates to how or how Loughton arrived at his figures.

As for citing last year’s 2.7% increase in the pregnancy rate for under 18s, this has no relevance to the data obtained by Loughton, which runs only to 2006, nor does a figure for a single year increase in the teenage pregnancy rate tell us anything about the overall trend in these rates over over time nor whether this increase is part of that trend or simply an anomalous result single year result – a mere blip in the figures.

Against that, a written answer provided by Beverley Hughes on 15 December 2008, to a question submitted by John Bercow, paints a very different picture of the prevailing situation:

Since the Teenage Pregnancy Strategy was launched in 1999, there has been a steady decline in the under-18 conception rate. Latest annual data (2006) show a rate of 40.6 conceptions per 1,000 females aged 15 to 17, a fall of 12.9 per cent. from the 1998 baseline rate. The 2006 under-18 conception rate is the lowest for over 20 years.

So, other than simple curiosity, why did Loughton get his bag carriers to go to all the time and trouble of generating this ‘snapshot’? Well, this passage from the Mail should make the reason perfectly clear:

The Nottingham suburb of Rushcliffe – whose MP is former Health Secretary Ken Clarke – saw the biggest increase in teenage births.

In 2006, 44 teenagers gave birth compared with only 16 in 2002 – a rise of 175 per cent.

Next on the list was the well-to-do West Yorkshire town of Pudsey, where the number of maternities to mothers aged under 20 rose from 26 to 60, or 130 per cent.

And the leafy Surrey constituency of Epsom and Ewell saw a 113 per cent rise – from 15 to 32.

It’s the same old reason that all politicians cook up spurious sets of statistical data – so they can cherry pick a few anomolous results which, when presented out of context, appear to support their chosen political position even – or perhaps especially – when that position is at odds with reality.

What happens, for example, if we look at the actual data that’s available for Rushcliffe where, unusually, the district council and parliamentary constituency have the same boundary – what does that tell us both about hiw Loughton compiled his statistics and, of course, about teenage preganancy rate in the area. Well, for starters, if we look at the data that is available from the UK Statistics Authority, the first thing we find is that nowhere is there a published breakdown of the number of conceptions in the area for either of the two years cited in the article. What is available is either an annual set of figures, in which the data for Rushcliffe is incorporated into consolidated figures for the county of Nottinghamshire or two sets of figures for the district alone, each of which consolidates three years worth of data into a single set of figures. What we have are either figures for the whole county for 2002 and 2006 or figures for Rushcliffe, alone, for 2001-3 and 2004-2006 and what the latter sets tell use that between 2001 and 2003 there were 98 teenage conceptions recorded in the area while, for 2004 to 2006, there were 117, giving an increase of 19% not 175% claimed by the Mail based on figures supplied by Loughton.

Moreover, we’re dealing with two different types of statistic here.

The Mail is citing births while the data provided by the UK Statistical Authority cites conceptions, which includes preganancies that ended in a termination – and, fortunately, the UKSA dataset also tells us what percentage of preganancies ended in a termination during each of the two periods, which in the case of Rushcliffe in 2001-2003 was actually 67.3%. So, for each birth to a teenager in Rushcliffe between 2001 and 2003 there were another two teenagers who had an abortion, from which we can estimate that if there were 16 births in 2002 there should have been a further 32-33 conceptions which didn’t lead to birth, giving 48-49 conceptions in total.

Now, hang on a second… according to the UKSA there were only 98 conceptions in total in Rushcliffe in the three year period from 2001-2003 of which, if the estimate holds valid, half were actually in 2002, which makes that a pretty bad year for teenage pregnancy in the constituency.

Similarly, if we turn now to 2004-2006, we find that we have 117 conceptions an a lower abortion rate – only 54% of teenage conceptions ended in an abortion over that three year period. So, when we feed in the numbers again we find that we should expect, on 44 births in 2006, a further 51-52 conceptions ending in an abortion, giving us a figure for only one of the three years of 95 conceptions. That leaves a mere 22 conceptions to be shared across the two previous years from which we would expect only 10 births, 6 fewer in two years than the figures cited by the Mail for 2002 alone.

It should be obvious by now that there is something serious screwy going on with Loughton’s figures, and the same problem emerges when we look at Epson and Ewell, where the constituency covers the entire borough council area plus Ashtead, which is part of the neighbouring Mole Valley constituency. Performing same calculations using the abortion figures for the area (71.9% for 2001-3 and 62% for 2004-6) we get estimates for the number of conceptions of 63 for 2002 (15 births, 48 abortions) and 84 for 2006 (32 births, 52 abortions) and yet, according the UKSA’s data, there were only 64 conceptions in the borough between 2001 and 2003 and 79 between 2004 and 2006.

As for Pudsey, forget even trying to figure out where and how Loughton got his data as any conception and birth data for the area is safely buried in the figures for the Metropolitan Borough of Leeds alongside those of six other Parliamentary constituencies, making it nigh on impossible to extract anything approaching a reasonable estimate from the available data.

So, the entire premise of this article is nothing more than a readily debunked fabrication, but we’re not finished yet because there’s more to choice of Rushcliffe and Epsom and Ewell than simply manufacturing a false picture of teenage pregnancy rates in the ‘Middle England’ – what both areas also have in common is that each shows a significant fall in the percentage of teenage pregnancies ending in an abortion over the period covered by Loughton’s dodgy data.

This serves two main purposes in the article. It enables the Mail to introduce a facile celebrity angle into the story (some crap about middle class teenagers getting up the duff in order to emulate Britney Spears) and it conveniently allows Nadine Dorries to throw this delightful piece of bullshit into the mix:

Nadine Dorries, a pro-life Tory MP, said there was ‘no doubt’ that the figures reflected an increasing distaste for abortion among the middle classes.

A large number of the constituencies with the highest increases would traditionally be considered wellheeled areas. A significant proportion have Tory MPs – usually voted in by affluent voters in suburbia or the shires.

Complete and utter rubbish, of course – and to prove the point I generated a few statistics of my own.

Using the government’s deprivation indices I pulled together a list of the 50 least deprived local authority areas in England, most of which you won’t be surprised to find are located in the South-East and Home Counties but for a couple of outliers (Harrogate, Rushcliffe). The logic here is simple, the least deprived areas are at the very least the most middle class if not the most affluent parts of England.

Next, I used this list to extract the corresponding data from the UKSA’s tables on teenage conceptions and abortions for 2001-3 and 2004-6 and calculated averages for the conception rates and numbers of conceptions ending in abortion amongst teenagers living in these affluent areas.

What I found was that, for 2001-2003, the conception rate amongst teenagers in these 50 districts came in at 23.6 conceptions per 1,000 women, which is substantially lower than the national rate of the time, which stood at around 42 per 1,000. Moving on to 2004-6, the data showed a very small fall in the conception rate, down to 22.5 conceptions per 1,000, which is about what you’d expect given that the national figures also fell by a similar amount. So, while the teenage conception rates are much lower in affluent areas, any recent downward trend is no better or worse than that of England as a whole.

But, when we come to look at the figures for the percentage of conceptions ending in an abortion, things start to get rather more interesting.

For 2001-3, the percentage of teenagers choosing an abortion rather than parenthood in these affluent areas stood at 56.3% but by 2004-6 this actually rose by 1.6% to 57.9%. By comparison the overall figure for England was, during both periods, stable at a little over 48%. The teenage scions of Middle England may well be about half as likely to get pregnant in their teens but when it come to choosing between early parenthood and an abortion they’re actually 20% more likely to take the latter route than the average for England.

Of these 50 districts, 37 saw a fall in their teenage conception rate between 2001-3 and 2004-6, but then the same is true of 217 of the 371 Local Authority areas in England and Wales for which data was available for both periods. That said, the number teenagers falling pregnant did rise over the same period, even though the conception rate fell slightly, which tell us nothing more than that the overall numbers of teenagers also rose over that period and at a faster rate than they were conceiving babies. When we comes to data on abortions, however, we find that the percentage of teenager choosing an abortion was on the rise in 28 of the 50 affluent districts, and East Hampshire (3rd in the least disadvantaged list) had the largest increase in England, rising from 36% to 57%.

On the full list of local authorities a mere 8 of the 50 areas with the highest abortion rates as a percentage of conceptions are actually deprived urban inner city areas, all of which are in London with Camden highest on the list in 23rd place. That’s only 12 places ahead of Dorries’ own constituency area (Mid Bedfordshire), where the number of pregnant teenagers choosing abortion rather than parenthood rose by 7.9% between 2001-3 and 2004-6 to more than 62%.

The middle classes may well find abortion distasteful, but any distaste they may feel seems rather easily put aside when its their own offspring up the duff, even in Dorries’ own backyard. After all, of the 50 least deprived local authorities areas in England, all of which have conception rates amongst teenagers well below the national average, only four have an abortion rate amongst teenagers (as a percentage of all conceptions) at or below the national average and this list includes the three districts with the highest rate of abortions against conceptions in England, the highest of which is South Bucks (which covers the Beaconsfield constituency of Dominic Grieve) at 77.9%.

Somewhere in all this there is almost certainly a really important series of debate to be had about everything from sex education to the evolving role of family and a whole bunch of other stuff besides…

…I only wish people like Loughton, Dorries and the Daily Mail would stop bullshitting for long enough to allow us to have those debates.

6 thoughts on “Different Year, Same Old Shit…

  1. Well written. I am unfortunate to live in Nadine Dorries constituency, and nothing makes me more annoyed than her grandstanding on the issue of abortion. I would say she did it purely for cheap political gain, but I feel that she is also positioning herself as a ‘woman on the Right’ in the event that Cameron gets elected and brings in some fresh faces. I have to put up with stupid lectures by her (based on dodgy stats) all the time in the local Bedford paper.

  2. Not taking issue with the wider point of the post (although politicians misusing statistics is hardly news… er, knife crime?), but the suggestion that the data are made up rests upon a big inference you’ve made. Just because the data are published by LA/UA, it doesn’t mean that they’re only available that way. ONS used to, and I assume UK Statistics Authority still do, offer a bespoke service, and assuming the data are recorded by postcode (“births are assigned to an area according to the usual residence of the mother, as stated at registration,” as per the answer) then it would be fairly easy for them to tabulate by Parliamentary constituency. Just because Tom Watson pointed them to the published source, it doesn’t mean that they didn’t make a further request.

    The statistical oddities could owe themselves to the process of retabulation – if the geography data isn’t available for all records, some would drop out. This would be a simpler explanation than an outright fabrication – cock-up rather than conspiracy. On the whole, politicians prepare to twist the truth rather than make it up (plausible deniability etc.).

    It’s odd that you don’t (maybe because of the wish to prove it’s all a fabrication) mention the simplest criticism of the story, which you obviously do appreciate by the way you did your own analysis – that year-to-year percentage changes when the numbers are in low tens are a highly unreliable guide, especially when the geographic areas are quite small (a girl may live in one constituency and go to school in another, where she meets a boy who’s from a very different area).

    By the way, more linking to data sources would add to the post.

    By the way 2, you cite uncritically Beverley Hughes’ defence of the Teenage Pregnancy Strategy. This isn’t unreasonable, but it’s worth mentioning that the U18 conception rate have fallen but by less than 10% (i.e., 4 percentage points); by the look of the data (4.1 in Pop Trends) the 1999-2006 fall in U18 births is split between reduced conception rates and increased termination rates. It’s also worth mentioning that it seems likely that the good economic times will have greatly affected this; at the margin, good prospects will surely encourage more U18 girls to delay conception… The true test of the Teenage Pregnancy Strategy will be whether U18 conception rates continue to fall or at least do not increase during a drab economy likely to last for some years; if not, then the Government will have simply been claiming credit for trends beyond their control. Not that that would happen.

Leave a Reply to Anonymous Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.