Unless there’s some piece of information you’re not supplying, the case hasn’t been proven either way yet. Surely someone needs to look at what proportion of the 250k leaflets went to people with cancer. Your assertion that it was a “random mailshot” begs the question.

Unless there’s some piece of information you’re not supplying

You mean apart from the fact that no political party has access to anyone’s personal medical records.

Well, the whole question revolves around whether they did/could have done or not, doesn’t it? Again, your conclusion assumes that your premise is correct. In particular, it would be interesting to know whether they used the Mosaic data that is available to them to cross-reference anonymised hospital statistics with postcodes and aim for a higher hit rate on those with cancer. If you want to prove this either way, look at the proportion of the 250k leaflets that went to cancer patients and check it’s in line with what you expect.

http://dizzythinks.net/2010/04/fix-system-not-speed-to-get-into-it.html

Alix:

Alternatively, the opposition parties could try producing evidence to substantiate their allegations for a change.

I should point out that using area-based anonymous statistics to target area with a higher incidence of cancer would be a perfectly legitimate method of targeting a mailshot and wouldn’t necessarily require the use of Mosaic.

I could pull that trick off just by using publicly available data from the DoH and National Statistics.

The Tory draft manifesto says that they will “remove political interference” in the NHS, and what does Lansley do? He tells the experts of an independent body that they should recommend drugs that are not effective. I think someone should have a look at the drugs Lansley is pushing and see if he’s got any connection with the companies making them. Something does not smell right.

Then, why not make that your main argument, and demand the proof? As it is, it looks slightly like you’ve tried to gloss over the fact that you don’t know what the proportion of cancer sufferers who received the mailshot was either.

I cannot agree altogether with your attitude to postcode/cancer stat targeting either. It would rather depend on what the leaflet said. Targeting as a general rule is fine, but surely it needs to be handled pretty sensitively in the instance of terminal disease. You’ll always upset some people with every output, whatever it is, of course. But it wasn’t exactly unpredictable that some cancer patients would get upset about a political mailshot about cancer, was it?

Alix:

Jeebus, this is getting to be like arguing with a hardcore god-botherer.

Either your party has specific evidence to back up this allegation, in which case you’re perfectly entitled to call for an inquiry, or it doesn’t, in which case it should retract the allegation and apologise.

Put up or shut up – its as simple as that!

Surely the onus is on the ~Party making these claims to prove it? Given that the Tory Party and friends in the press have a well won reputation for being lying scumbags, it is safe to assume that on this occasion they are trying to deflect the debate away from an area where they are notoriously weak.

So, is there evidence of such actions? If so, is it such a bad thing to tell people the truth? Aren’t we demanding that everyone tell the truth? Isn’t cancer screening important enough to be subject of the truth?

My wife got this card. Whether targetted as someone in a cancer screening program or not, it is a nasty document. The upset it caused is entirely predictable and it will probably cause more lost votes than won ones. Makes me wonder about the depth of cynicism and stupidity of whoever decided it was a good idea to send it out. Voters don’t think like political anoraks, this is a very clear example of that..

“Either your party has specific evidence to back up this allegation, in which case you’re perfectly entitled to call for an inquiry, or it doesn’t, in which case it should retract the allegation and apologise.”

Dude. I pay it a membership fee, I’m not married to it. My point is perfectly valid and has got nothing to do with partisanness, otherwise I’d have defended specifically what Vince said. I’m coming to this for the first time, through you, right? What I see is that some people have made an accusation, and you’re attempting to counter it, but you don’t actually have the figure that would prove you right. I’m sure all the linking bits that have now come out in the comments were in your head, but they’re not in your post.

So I’m saying, maybe you should have foregrounded the fact that the onus is on them to supply more detail about the accusation (which I totally agree it is). And not tried to make it your main point that the mailshot is random when you now also concede it could be “legitimately” targeted. You’ve just not set out your piece in a way that makes your response look particularly good, that’s all, and it surprise me because I’m used to you comprehensively demolishing things.

Actually Unity, anybody who watched the Daily Politics yesterday morning would know that Cable was absolutely restrained and careful, repeatedly stressing that IF the allegations were true THEN something needed to be done. he did not go for hyperbole or outrage, and he was very careful not to jump on the opportunistic bandwagon.

For somebody often so sanctimonious about getting on your high horse of accuracy and truth, you’d do well to actually bother to check what people really do say. I mean, the quote you’ve given us is hardly damning in itself, is it? When you watch the video (available on iPlayer) of what Cable said and the context in which he said it, there is absolutely no way he made a “fool” of himself.

TBH I’m getting a bit sick of your hectoring passive/aggressive rants – especially when you’re exhibiting certain degrees of hypocrisy in the process.

Mark @ 10

Whether targetted as someone in a cancer screening program or not, it is a nasty document.

What is nasty about telling the truth? Surely every party has a legitimate right to highlight these types of issues. If the Tories would scrap such a commitment, then isn’t that fair comment?

http://www.bbc.co.uk/iplayer/episode/b00s0z4h/The_Politics_Show_London_11_04_2010/

22 minutes in.

A measured, calm and reasonable response to the question asked.

Cable does not make a “fool” of himself at all.

Unity – the mailing was targeted at individuals within areas, not just at areas. Those people were chosen somehow.

You – and indeed the Labour Party – are welcome to continue pushing the line that statistically some recipients would have had cancer. I accept everything you say on that score, but surely I can simply turn to you and demand – in the way you have of Labour’s critics – that you show that the proportion of those who received this mailing who had had cancer was at the level you suggest.

Labour can shut this down simply by disclosing the selection criteria for the mailing.

So a question for you: why do you think they haven’t?

When you watch the video (available on iPlayer) of what Cable said and the context in which he said it, there is absolutely no way he made a “fool” of himself.

Heaven forbid that anyone should dare to gainsay the beatification of Vince Cable, least of all by expected a slightly higher standard of statistical numeracy from a trained economist.

Alix:

What I’ve clearly demonstrated here is that there are perfectly sound and reasonable methods of accounting for the fact that a number of current/ex cancer patients have received this leaflet, none of which rely in the slightest on the Labour having had access to confidential information taken from people’s medical records.

Not only that, but as should be patently obvious here, random chance and/or limited targeting using publicly available incidence data are both infinitely more plausible explanations than the suggestion that Labour have accessed medical records.

So, unless anyone can produce specific evidence to the contrary, I think I’m perfectly entitled to apply Occam’s razor to this one and conclude that these allegations are bunch of overhyped, overblown bullshit without the need to labour the point.

Matthew:

You appear to be somewhat confused about what constitutes targeting.

What was sent out was a personalised mailshot but not necessarily a targeted one – just because your name’s on letter doesn’t mean its specifically targeted at you.

Last week, I received a personally addressed invitation to apply for a life insurance policy which is targeted specifically at the over 50s.

As I’m 43, that letter was about as much use to me as a chocolate teapot, and a complete waste of time and money for the company that sent it out, but it was, nevertheless, personally addressed to me for my attention.

So what should I do – should I waste a tenner on pulling the data that company has using DPA in order to set them straight, or is it more reasonable to think that they’ve got my name off a marketing database as sent the letter out on the off chance that I might be eligible because they really don’t know how old I am.

Mathew Taylor @ 15

It cannot that be that difficult to select those who are likely to be a target group for such a campaign based on age, sex etc. Surely, if your mailshot was good enough, you would expect to get a high number of ‘hits’ in terms of cancer paitents as well?

The target group of this mailshot should be those whose are at risk of needing a screening, not those on the road treatment though? Why target someone who is already undergoing treatment?

If we where trying to mailshot car owners to hightlight car tax, for example. Let us suppose that one Party were going to tax 4×4 with a higher tax. I could compile a mailshot pretty easily to hit that demographic, without the need to raid the DVLA

Unity: That’s pretty weak, even by the standards of the Labour response on this issue.

Labour sends mailings to 250k specific people, selected from a database covering the majority of people in the UK. Are you seriously suggesting that they just picked 250k random people to receive the mailing?

I keep having to say this to people defending Labour on this issue, but go back and read the specific examples the Times gave in their original article. These mailings were clearly targeted. At best Labour picked specific Mosaic codes; at worst, they used other data from National Canvasse to select recipients. If the latter, there is a valid question about how they were selected.

It’s a data protection issue. Did Labour know that the recipients of the leaflets are or had been cancer sufferers, or did they simply target people in a certain age bracket and this happened to contain a number of cancer sufferers?

If Labour have access to a list of cancer sufferers and have used it to send these people cancer-related literature, there’s a big question of where they got that information from. It’s not clear to me how a political party could get hold of an accurate list of cancer sufferers, so if they have done so then there’s a case to answer. If they just sent this out to the general public, based on anonymised demographic targetting, then there’s no issue (apart from whether you consider the leaflets to be in good taste, but I have no real objection on those grounds, not being a cancer sufferer).

“Experian, the data management company, confirmed that both Labour and the Conservatives use its Mosaic database, which divides voters into 67 groups. The databases can use anonymised hospital statistics, including postcodes and the diagnoses of patients, to identify the likely addresses of those with particular illnesses.

It cannot identify potential breast cancer sufferers because the disease affects adult women of all ages and backgrounds.”

So blaming it on Mosaic is a statistics fail.

“The annual incidence for breast cancer in women is currently around 121 cases per 100,000 female population, while the lifetime risk of developing breast cancer, for women, is 1 in 9.”
Statistics fail. You make those two numbers consistent even if you postulate there is an equal chance of developing breast cancer before puberty. If one assumes a constant risk of 121 in 100,000 from age of 12 to 82 then the lifetime risk comes out at a bit less than one in 12.

“Its a matter of elementary probability – a random mailshot of that size is guaranteed to include at least some women who have had breast cancer – and that’s all there is to it.” NO IT IS NOT guaranteed unless the size of the mailshot exceeds the number of adults who have not had breast cancer. It is highly probable but the study of probability is all about non-guaranteed outcomes.

Other websites have instanced a host of suspicious coincidences that justify an enquiry into whether it was targeted (including the mailshot being sent out by a company that also sends out mailshots for Cancer Research UK). Some of those quoted have a probability of less than 1 in 200 – one of these in a big mailshot is plausible but the probability of three successive 1 in 200 chances is 1 in 8 million which you do not expect from a 250,000 mailsot

Perhaps Unity should LEARN a little elementary probability before slagging off Vince Cable.

Unity:

“Heaven forbid that anyone should dare to gainsay the beatification of Vince Cable, least of all by expected a slightly higher standard of statistical numeracy from a trained economist.”

Heaven forbid that you should actually read what I’ve written and watch that video, and admitt that your characterisation of Cable as an inumerate “fool” is wilfully inaccurate and partisan.

Heaven forbid that you should admit that you’ve over-played your hand and are making an unreasonable attack to which you are not entitled, eh?

John77

If you want to query the incidence and lifetime risk statistics then I suggest you take it with Cancer Research UK, which is where both figures come from.

As a matter of fact, the 1 in 9 figure for lifetime risk is correct as it takes into account both that age is the most significant risk factor breast cancer (i.e. the incidence increases with age) and the longer average life expectancy of women.

As regards guarantees, the probability of randomly selecting 250,000 people out of population in which the lifetime risk of breast cancer is 1 in 9 and not getting at least one person who has had breast cancer is so small as to be reasonably considered negligible.

Other websites have instanced a host of suspicious coincidences that justify an enquiry into whether it was targeted

Perhaps you’d like to point out where some of these other websites are so we look at their calculations – so far I’ve seen nothing that can’t be accounted for by either randomness or by a limited form or targeting that doesn’t rely on access to personal information.

To tell the truth, from the information I’ve seen so far, the only obvious targetting that is apparent here is that many of these leaflets have been delivered to marginal constituencies.

Paul, you’ll pop a vein if you keep that up.

“Paul, you’ll pop a vein if you keep that up.”

Wow. I don’t think I’ve ever seen you talk like that before, Unity.

Unity
Do not try bullshitting me on arithmetic. The maximum value of the product p of a set on n numbers whose sum is fixed at s occurs when all those numbers are the same s/n. So, to be generous to you, I took the risk factor as a constant. If the risk changes with age the lifetime risk factor will be LOWER than the number I quoted. OF COURSE I used female life expectancy: who do you think I am – Stephen Byers?

NO, the 1 in 9 is NOT correct – if you have the name of a statistician in Cancer Research UK who claims it is, please inform me and I shall take it up with him/her personally and his/her university statistics department who deserve a visit from Ofsted (since this is incompetence at GCSE-standard). You then say that the figure is correct which implies that you have checked it, so why should I not take it up with you?

If you want a reference look at Dizzy – but be warned he has had family experience with cancer and does not like ANY party playing politics over this – also Timesonline.

Wow. I don’t think I’ve ever seen you talk like that before, Unity.

Elections do funny things to people…

John77:

Here’s the link to Cancer Research’s section on breast cancer – http://info.cancerresearchuk.org/cancerstats/types/breast/incidence/index.htm – the 1 in 9 lifetime risk comes from their own analysis of incidence/mortality data for 2001-2005.

I don’t think the attack on Cable is warranted TBH, because he is restrained.

But it could also be seen as a dog-whistle. If you have no proof that irregularities took place – then there’s no point implying it.

The burden on proof is on the Tories and Libdems to show if data was accessed illegally and then take it to the press and/or Election Commission. In this case all they’re trying to do is deflect attention from what is actually being said in the leaflet and whether they agree or disagree with it,

@ 28 Unity
Thank you – I shall look at it in detail but the firstr thing I noticed is that it refers to breast cancer rate for females averaged over all ages is 147.3 which is significantly greater than 121 for women.
More than ONE-THIRD greater.
You misquoted the numbers.
Until I have checked I cannot be sure but it looks as if it is quite possible than Cancer Research UK is right but as I said, you were blatantly wrong – and you should not criticise Vince Cable if you are misquoting data and unable to see that the figures you supply are mutually inconsistent.
Further comment to follow when I shall have had time to carry out some basic tests and create a single-variable model (which will overstate the risk as they do not supply data on reoccurrence so a second or third attack on one female will be treated as an attack on a previously unharmed female).

John:

If you check the incidence data on the same page, you’ll see that 121 per 100,000 is the correct figure for 2006.

Oh, and even if we take 1 in 12 for lifetime incidence, we’d still be expecting to turn up just over 20,000 women who have been diagnosed with breast cancer in a random sample of 250,000 UK women.

@Unity: Except this wasn’t a random sample – or are we still playing that hand?

@Sunny: This isn’t a matter for the Electoral Commission, but for the DMA and the ICO, both of whom could tear Labour a new one – since they could make it very difficult for Experian to continue providing data as they have done hitherto. I suspect someone somewhere is looking at that angle anyway; it’s not party political, and is tailor made for one of the privacy campaign groups.

As to the content of the leaflet, I personally think it’s total nonsense. I’ve yet to see a clinical justification for the guarantee that made sense, and Burnham’s appeals to “putting people’s minds at rest” (paraphrase) rather back up that view. I’m not in favour of the guarantee unless it has suitable clinical benefits as it will inevitably involve prioritising that service at the expense of others. It’s a cynical attempt by Labour to distract from the Conservative pledge to increase NHS funding, given the Tories have made it clear they’ll spend that money where it will do most good, even if that means not on one of Labours “pledges”.

Update
Applying the incidence rates for 5-year age-groups from Cancer Research UK to the latest English Life Table gives a lifetime risk factor (on the assumption that there are zero repeat cases) nearer 1 in 8 so their figure of 1 in 9 – on the basis of a 147 per 100,000 females of all ages – significantly very different from 121 per 100,000 women – as a lifetime risk is plausible .
However Cancer Research UK also gives some survival rate figures – the ten year survival rate improved from 47% under Wilson/Callaghan to 58% under Thatcher to 67% under Major to 72% (a much smaller improvement under Blair). They do not quote enough 20-year survival rate figures to make a comparison beyond them being noticeably lower than the ten-year rates
So a 1 in 9 lifetime risk of suffering Breast Cancer does not convert to 30,000 out of 250,000 recipients of the postcards having suffered breast cancer in the past (even if all had been addressed to women, whereas the press reported a smaller number of complaints from men). A random post would be 120,000 men including about 12 survivors of breast cancer and 130,000 women including 7,000 FUTURE sufferers and around 3,000 past sufferers.
Which means that Unity’s estimate of 28,000 of past sufferers is out by a factor of 9 or 10!!

@32 Unity
No we should not! Mortality Rates!!
If there was a 100% survival rate, why should anyone panic over cancer?
Cancer is a killer.
You are bloody irritating. I TOLD you not not to bullshit me on arithmetic.
250,000 leaflets targeted by postcode would give 130,000 women not 250,000
If it was 250,000 women then at 1 in 12 it would be 10,000+ FUTURE sufferers and 5 to 6,000 past ones

@31 Unity
I DID check – I even downloaded the excel file: it says crude rates Females 147.3 per 100,000 population, Males 1.1 per 100,000 population.
The figure you quote is the “European ASR” which you will see, when you learn some elementary statistics, is mathematically inconsistent with a 1 in 9 lifetime risk.
FYI it has a zero rate for females 0-14.

@Unity: Except this wasn’t a random sample – or are we still playing that hand?

MAtthew – in fact I highly doubt it was a random sample. It’s more like they used contact creator to target men and women are most at risk and are likely to be concerned about those services.

That’s the whole point of micro-targeting. Doesn’t mean any privacy laws have been broken though.

Sunny H
I am trying to tackle Unity’s mathematical errors without indulging in any actual politics, but I must point out that the Tories and LibDems did NOT start this issue – the press did and they just reacted (or were faced with questions which Cable answered in a reasonable manner) – the first I saw of it was in a geek blog referring to the Sunday Times – Conhome followed a bit later and LibCon even later.
So your post 29 is divorced from reality – it is not up to the other two parties to take it to the press, it is for the Labour Party to respond to the press (or not).
I have seen references to Experian, but as they said their data could NOT be used to target breast cancer sufferers, that makes no sense.
Cable was quite right – IF there is a case to answer THEN we must investigate.
If there is not, then why waste time and money.
So far we have a lot of upset people, a number of suspicious coincidences, and some replies from Labour party activists such as Unity that merely increase suspicion because their defences are clearly invalid (OK, possibly not to them but clearly to me)
Do you really think that when there was substantial cause for suspicion but there could be no proof (there could be no proof unless the Police had raided Tangent and Labour Party HQ before the story broke whereafter any conspirators would have taken steps to destroy evidence) Vince Cable should have said “oh, they must be innocent – we all know that Labour would never do anything in the least bit dodgy to gain to gain an unfair electoral advantage” instead of providing an honest balanced comment?

“but I must point out that the Tories and LibDems did NOT start this issue – the press did and they just reacted (or were faced with questions which Cable answered in a reasonable manner) ”

Precisely. Any politician posed with this sort of question would answer the same as most did, cautiously condemning without making any sure statement that they believe the media claim that it definitely happened the way they say to cover off all bases.

errrr, this is straight electioneering. The story pervading the MSM is one of Labour targets cancer patients, the cold hearted bastards will do anything to win.
The truth is (to the MSM) irrelevant and will get no coverage while the spun story of labours electioneering is busily making all the delicious swing voters switch to conservative.

@Sunny: Depends on what the targeting was, ultimately – and of course that brings us full circle: Labour could say, but doesn’t want to.

(I think they’ve really missed a trick here: if Andy Burnham’s statement is “true” in the broader sense, as opposed to the narrow semantic sense, they could’ve killed the story yesterday or first thing this morning by specifying the selection criteria.)

Putting that to one side, I think we’d both agree that the privacy law aspect is secondary to public perception: targeting cancer sufferers (or “people who are statistically more likely to have suffered from cancer and/or have had cancer screening”) over cancer care is unacceptable to the general public, where targeting elderly people over personal care, or people with young children over education, is.

I got one of these leaflets.

What a non-story.