Category Archives: Statistics and simplicity

Hard to reach?

A long time ago I worked for a company specialising in researching the ‘hard to reach’, by which we meant the poor, the needy, including the elderly, drug users, asian Muslims, the white working class. Essentially, the kind of people that don’t respond to mail surveys as often as other groups. And in order to talk to these people we went to where they were: the street, bingo halls, community centres, drug treatment centres.

Which is why the headline ‘Church of England eyes £5m of state funds to combat extremism’ (Guardian) made me laugh. The CofE claims it can enable “Mr and Mrs Smith, Mr and Mrs Patel, and Mr and Mrs Hussain” to engage with each other through coffee mornings and so on.

First, they will use money so that vicars and imams can get to know each other. Fair enough, but there’s plenty of that going on already, and I don’t think vicars and imams are failing to get on (unless we’re thinking about the fundamentalists and crazies and they aren’t invited). But once this has happened, then what. In a working-class estate where I’ve worked recently, of around 7,000 residents only 50 or so have any regular involvement in the church. The vast majority of UK adults go to church less than once a year, probably for weddings and funerals (tearfund) and as I expected, it’s the middle classes (AB) and pensioners that are most likely to attend church.

Now forgive me if I’m wrong, but the government isn’t worried about middle-class pensioners starting riots. The kids that fight each other over their backgrounds won’t be reached through the church, and many won’t be reached through the mosque either. Contrary to stereotype, Muslim youth also ‘stop going’, rebel against their parents. If government wants to bring people together why not invest in the truly public sphere: make our parks more appealing, set up sports events, invest in council housing with genuine public spaces where neighbours can get to know each other.

Leave a comment

Filed under government, Statistics and simplicity

What’s in a name?

We’re doomed… or so many people would have it. It seems to be a common thread in newspaper articles and their responses, and blogs too, that Britain is changing demographically at a huge rate, and so in X number of years ‘we’ll be a Muslim country’. The latest versions of this were the stories on baby names – an annual affair – and the recurring story that such and such a city will be majority Muslim, or majority X, or ‘whites will be in a minority’.

The baby names story is interpreted as:

‘Mohammed is top boys name’ (Express), Mohammed, the nation’s (secret) favourite name (Telegraph)

Often, this data is presented in terms of a conspiracy: the ONS is disguising the fact that Mohammed is the most popular boys name by treating all spellings separately. This is, of course, nonsense: the data is available for people to do these calculations, it isn’t buried. If we think about spellings and variants both the boys and girls lists would change. Do we count Harry and Henry together? What about Isabelle and Isabella? Putting these two together would make Issy the 2nd most popular girls name.

This discussion also misses the most important question about trying to translate baby name tables into demographic analysis. How do these name distributions relate to religious distribution? For if we are to look at the boys list, Mohammed is the only Muslim boy’s name in the top 100, and it accounts for 6,535 of 204,494, around 3.2%.

And this needs to be put in the context of how Mohammed is used as a name. Globally, one in five Muslim men have Mohammed as a first name, and I think in the UK it would be even higher. Often, though, it isn’t the name that is used: lots of people have Mohammed as first name, but are referred to by the name after (see the Indie for an example). It’s this convention that means that Mohammed getting to number one in the list does NOT mean that more Muslims were born than anyone else. I’d guess that the 3.2% of boys born being Muslim is probably close to the actual figure.

Leave a comment

Filed under bad social science, News, Statistics and simplicity

Dumb

Now I like small and beautiful as much as the next person, and would use a MacBook Air as a second PC for travelling, but the latest marketing message is just dumb. As the BBC put it, ‘The MacBook Air is 0.11 inches thick at its thinnest point’.

But the important info isn’t that of the thinnest point, but of the thickest point. A really thick laptop could have a thin wedge sticking out; even an equilateral triangular prism has a thinnest point approaching zero.

No, what matters is the thickest point. It’s the thickest point that determines whether it fits in a given laptop bag, goes through your letter box. If the laptop was 0.11 inches thick at the thickest point, then that would impress me.

In fact, that’s what drew me to the story. At first I thought it said simply that the laptop was 0.11 inches thick, and I assumed it was a misprint as it’s impossible. But hey, many people actually have reproduced the story as ‘MacBook Air now 0.11 in thick’, and have fallen for the spin completely: I take it they failed geometry at school.

Leave a comment

Filed under News, Statistics and simplicity

Income distribution

A couple of weeks back I went to a debate about the legacy of the miners’ strike. There was a lot of shouting at Edwina Currie, being the only Conservative there, and a member of the government at the time. There was an element of nostalgia too, with mining jobs being romanticised a bit too much (George Galloway and Ken Loach were there too). However, the fact remains that these dirty and dangerous jobs seem to pay better than the service sector jobs that have replaced them.

One of the more interesting claims was that British workers have really good earnings. Edwina pulled out the ‘creative industries’ argument, like Charlie Leadbeater’s Living on Thin Air, effectively saying we could all be earning good money designing computer games. It is of course true that the average British wage is high, and the creative industries is profitable. But for the worker at the bottom of the pile, it’s the distribution that counts.

For example, if a company makes £1m p.a., after costs, and shares it between 50 workers equally, then they all get £20k each. But if they decide to ‘award’ the 4 managers with £100k salaries, then the remaining 46 only get £13k each. The mean wage in each is the same, so in any analysis we should examine the distribution, not just the minimum, maximum and means.

Thus, on the one hand the government can tell its domestic audience that we’ve never had it so good, and that we’re paid really well. This was Edwina Currie’s line. But when its audience is overseas investment, a different story is told:

‘The UK has a competitive salary structure in the service sector [i.e. cheap], particularly when compared to countries such as Germany, Ireland,Spain, Sweden and Switzerland… hourly compensation costs for production workers in the UK are also lower than in many other countries…’ (UK Invest)

The same document also shows that the UK has the reputation of the most flexible labour market [i.e. best for business, not workers], except for China:

UK has most flexible labour market

But internally, businesses give the impression of being hampered by red tape, unions with too much power, and the minimum wage. The CBI originally said the minimum wage would reduce the number of jobs, then each time there’s due to be a raise they say the same thing.

Perhaps at GDP per head, the UK is doing well, but we also have the most unequal wage structure outside the US, so people can still be badly paid here.

Leave a comment

Filed under bad social science, government, Politicians, Statistics and simplicity

Medical mistakes

Here’s another question of an order of magnitude… Daniel Ubani, a German ‘stand-in doctor’ who was over in the UK doing a shift as an on-call GP and gave a patient a fatal overdose of diamorphine (Guardian again). He gave the patient 10 times the amount – at first I wondered if he’d got millilitres and centilitres mixed up. However, you can now read the inquest material and it makes for interesting reading.

It seems that the mistake came when David Gray’s partner mentioned 100mg and 10mg, pethidine and diamorphine, in the same conversation. Having no pethidine, Ubani chose the diamorphine. Hearing 100mg, he chose the bigger vial. As diamorphine isn’t used in Germany he wouldn’t know the correct dose: he didn’t look at the instructions kept with the drugs, hence ‘manslaughter by gross negligence’. However, the death could have been prevented by proper induction. It seems that no-one showed Ubani round his equipment: he should have known that the 100mg vial was for palliative care and used in slow release with a syringe driver, as opposed to all at once.

At the inquest there was a some discussion about Ubani’s proficiency in understanding English. However, one of his employers said that his English was OK, and it seems unlikely that this was a huge factor. His unfamiliarity with NHS kit and procedures was more to blame. However, the papers seem to have decided that the black African with a German passport was an example of non-English speaking foreigners making a mess of things (see the Mirror). There’s a lot of blame on the profit making agencies too.

But, as noted in the report, this has happened before, and the month before the deaths NHS Cambridgeshire were discussing changing the boxes of drugs to avoid this:

“In attempting to relieve patients in acute pain, doctors in two different situations erroneously selected the 30mg diamorphine vial…and administrated the entire contents to their patients by injection.

“This six-fold overdose caused respiratory depression and collapse; the patients had to be admitted to hospital for resuscitation. If this is repeated and the patient not rescued in time, death could result.” (C4)

I don’t know if these doctors were the stereotyped foreign doctor, unable to speak English, or not. The lack of media interest suggests not as I’m sure a scandal would be being promoted as I write.

Indeed, focusing on this rare occurence misses the wider issues in ‘patient safety’. If society is fixated on this one fatal mistake, then it can ignore the fact that in 2007 there were ‘as many as 860,000 errors or near misses involving medicines’ in the NHS (Guardian). I’m sure that not all these mistakes were down to foreign doctors, but are down to normal human error, the kind of error that can often be avoided with the right systems in place, good management and so on. Dull (and expensive but that’s another story).

Leave a comment

Filed under bad social science, media, News, Statistics and simplicity

An order of magnitude

We know many people are bad at maths. However, most mistakes are easily spotted through experience and common sense. But sometimes, common sense is lacking: here I’m talking about the Conservatives gaff on teenage pregnancy.

So first, the Conservatives. A few days ago they launched a document called Labour’s Two Nations, that was supposed to show how there is great inequality in Britain today (let’s ignore the fact that the rise in inequality happened in the 1980s). What they wanted to point out was that under-18 girls in the most deprived areas are three times more likely to become pregnant than in the least deprived areas (Guardian). It’s not clear what this means with regards to ‘most deprived’ and ‘areas’ – I think it’s top and bottom centiles and districts – and I’m sure I could find a more shocking figure if I chose a harsher definition of most and least deprived. The mistake they did make, though, was to divide 54 by 1000 and come up with 54% not 5.4%. That’s if they did a calculation: some social statistics come as ‘per 1000’ or ‘per 10,000’ and it’s important to notice this.

This matters for two reasons. First, because 54% v 18% is a big difference and much more significant than a difference between 5.4% and 1.8% (significant thought this is). Second, because anyone with any sense would realise that 54%, that is over half, is completely absurd.  Anywhere with 54% of its teenagers pregnant would have babies everywhere. Either the writer and editor just missed this, or they genuinely believed that there could be such a place and they are massively out of touch with normal life.

Leave a comment

Filed under media, News, Politicians, Statistics and simplicity

No good reason

News that caught my eye this week included the rise in fines levied on people who take their children out of school. In the UK, a school can fine parents £50 (rising to £100 if not paid in 28 days), if their children are absent without good reason. The spin on this story is that some people are taking their children on foreign holidays in term-time because it’s out of season and so cheaper, and these people are damaging their children’s education.

However, it’s important to note that there are a number of reasons why a family can take a child out of school and will get permission. These include religious observance, weddings and so on. These absences can, of course, damage a child’s education, but the assumption is that these days can’t be changed, whereas a holiday can. Here is the official guidance.

This privileges certain kinds of events (religious festivals, taking part in ‘approved’ sport, wedddings) over others. It’s difficult to see why the state or headteachers should be able to make a value judgement as to which is worthwhile. Should a child be given an authorised absence for a political conference or a protest march? Who decides which religions count? What about looking after family members? One can see the puritanical hand of government here – work and education good, leisure bad – and also the audit culture: some absences count against the school’s attendance record and some don’t and schools don’t want bad statistics.

And when it comes to justice, it also raises the question, ‘what is freedom?’. Yes, theoretically all parents could choose to take holidays out of term time. But many can’t afford to. Perhaps there is greater harm in a child never having a holiday away from home than missing some school days. If the government really want to ensure people don’t take their kids out in term time, then perhaps they should force holiday companies to alter their pricing, subsidise poor people’s holidays, or, as a simpler solution, increase the minimum wage and benefit rates /introduce a citizens’ wage.

Leave a comment

Filed under government, News, Statistics and simplicity