We’re all aware that the abuse of statistics seems commonplace in the press and I’m sure we all treat the figures we’re presented with with the degree of scepticism they deserve.
Recently we’ve had Shadow Chancellor, George Osborne, misplace a decimal point to make the suggestion that 54 per cent of girls from deprived areas fell pregnant before their 18th birthday when it should have been 5.4 per cent!! I also couldn’t resist the headline describing some research carried out by Friends Provident which boldly claimed that “27 million Brits confess they’re clueless about private pensions”.
The only reasonable interpretation of the headline is that 27 million “Brits” (which I assume is lazy shorthand for people resident in the United Kingdom, regardless of how they define themselves) have personally declared to Friends Provident that they are clueless about private pensions.
Now, given that the latest population statistics for the UK suggest that there are approximately 61.4 million “Brits”, of whom roughly 20 per cent are under 16, that gives us roughly 49 million adults in the UK – how good at maths am I!!
Given the numbers involved I’m clearly a bit surprised that I didn’t get interviewed and was bit worried that I was either forgetful or unpopular. Clearly, I’m one of the 24 million “Brits” that Friends Provident didn’t manage to speak to. Did they manage to speak to you? No? What a coincidence!
The truth behind the headline is that Friends Provident, or more accurately, a company acting on their behalf, surveyed 2,462 people. So a more accurate headline is “1,428 people who may or may not define themselves as British said in a survey that they were clueless about private pensions”. Not quite as dramatic, I’m sure you’ll agree.
We don’t even know who these people are or how the sample was selected. I remember carrying out similar surveys whilst at university. They highlighted a few flaws in the whole attitude survey process. At least 25 per cent of any survey was generally completed by the students themselves as a result of failing to get their quota of real respondents. Another 50 per cent consisted of people that the students had met in the pub while having lunch. Any remaining balance was made up of people who didn’t have anything better to do in their lives other than to agree to answer inane surveys from spotty students, about topics which they had never really considered, but were still prepared to express a, usually incoherent, opinion on.
There is also the Bradley effect to take account of, which basically states that people lie in opinion polls and surveys all the time. I’ve read all the counter-arguments from pollsters and statisticians about random sampling and techniques to overcome all the flaws inherent in surveys of this kind and, whilst I’m not suggesting that the Friends Provident survey was carried out in the manner of some of my university experiences, I remain unconvinced about the value of simply extrapolating results from a small sample to the UK population as a whole, on any subject.
It’s easy to overlook that there is a real issue behind the headline which needs to be addressed. We are a financially illiterate nation and as the world increases in financial complexity then we need to be ensuring that people are provided with sufficient education and information to make informed decisions about financial transactions.
But keeping matters in a lighter vein on this occasion, the Friends Provident headline represents a pretty spectacular leap from the data. I’d be delighted to hear of your favourite example of the appalling misuse of statistics – why not use the comments box below.
