Tag Archives: bad social science


I’ve had a new baby, and two children take up all your time, so I’ve not added to this blog for a little while. However, a bit of research (that is vaguely connected to having kids) inspired me to restart my writing here. The headlines were:

Pressure of modern life causes accidents in the home

This is just the kind of ‘bad social science’ I love to find, partially because I used to be employed doing this kind of stuff. The research is commissioned by Lloyds TSB Insurance, who get their name in a news article in an ‘advertorial’ mode, and it is conducted by a consultant (David Moxon), and he probably got one of the big name survey companies to do the survey. We have a number of social facts, in this case a ‘shorter attention span’, ‘a “surge in domestic accidents’, and people ‘blam[ing] stress or too much work’ (see PA’s report). More accidents mean we should go out and get insurance!!!

However, some of these social facts are questionable, and thus the links between them don’t make much sense at all, and others may have much more obvious explanations. The one thing this research found was that some people had accidents and blamed it on stress, and that attention spans were shorter in younger people. Let’s look at each part in turn:

More accidents: A&E statistics do show a rise but a strange one. Between 1987 and 2003 the figures were very steady (and presumably decreasing when you take into account population growth). The numbers grew suddenly but this coincided with changes in the way figures were compiled and new ways of organising A&E provision. I’m not convinced there are more accidents.

‘more stress or too much work’: Time use surveys show that the average amount of total work done (paid work and unpaid work such as housework) was actually decreasing between 2000 and 2005, and this is true across all age groups. I guess it’s possible that those having accidents have had their hours increased while others have seen them decrease, but this seems unlikely. The activities that are taking more of our time are watching TV and using computers.

The fact that attention spans are shorter in young people points to other explanations. First, younger people are more likely to have young children. The attention parents need to pay to their kids is attention they could be directing to the burning toast. Presumably this has been true since the beginning of time. We need to know if attention spans of the 25 year old mum have decreased over time to know if any other explanations are needed.

Second, if attention spans have decreased over time, using the word ‘pressure’ seems absurd. We have more labour saving gadgets, more leisure time, more money and resources than we had in the past. The ‘pressure of modern life’  phrase does a disservice to people who worked all hours in Victorian times, survived two world wars, went through the great depression and so on. It wasn’t so long ago that the normal working week in factories was all-day Monday to Friday, and Saturday morning. The word we are looking for to describe the effect of our times on attention spans is ‘distracted’: too much telly and internet and, oops, the bath is overflowing.


Leave a comment

Filed under bad social science, News, Statistics and simplicity

Sceptical about climate change?

This is the story that’s made me the most angry in recent months, but I missed writing about it because I was on holiday (better late than never). It’s the Observer story with the headline: most Britons doubt cause of climate change. It was based on a Ipsos Mori survey that can be found here and if you read the original data and the Observer analysis, you’ll discover how far from the data the news story is.

I sent a letter to the Observer, funnily enough not printed, pointing out that their claims weren’t backed up by the survey data:

…If 42% of people think climate change might not be as bad as people say, they might still think it’s bad or even very bad. The 60% of people that agree that ‘many scientific experts’ question the causes of climate change could believe that these scientists are wrong. Indeed, does ‘many’ mean 10 scientists, 10% of the scientific community or most scientists? Respondents don’t know what the question means…Your poll, and its subsequent analysis, is bad social science.

For the ‘fact’ that ‘Most Britons don’t believe climate change is man-made’ they asked the public to guess the state of play of the world scientific community, instead of asking them if they themselves thought humans were responsible.

As pointed out by those sceptical of climate change, the survey is used to show that the population don’t believe in climate change, and therefore more needs to be done to ram it down people’s throats (the sceptics arguing that it’s already rammed down our throats enough, but that it’s not true). But if you look at the data, and I don’t think many people have, you’ll find a story of most people (77%) being very or fairly concerned about climate change, most people thinking that something can and should be done, including most people saying invest in renewable energy. Just because people can’t be cast-iron certain how bad climate change will be, and how many scientists agree and disagree, doesn’t mean they believe it’s a hoax. Remember, even when the Intergovernmental Panel on Climate Change says global warming is real, they still put out a range of scenarios (six in IPCC AR4) because scientists, individually and collectively can’t know exactly how it will pan out.

And finally when asked ‘what is reasonable to expect people to do to tackle climate change’, only 6% of Britons chose ‘There is no need to take any action – climate change is natural/humans are not having that much impact’. Not many sceptics there.

Leave a comment

Filed under News, Politicians, Statistics and simplicity

Work brings freedom?

Our government’s obsession with workfare in order to prove that it’s not soft on the undeserving poor continues unabated. This time it’s an ’empowerment white paper’ from DCLG (here), which somehow makes paid employment into the most enjoyable and empowering thing one can ever do. Even call centres?

‘The Empowerment White Paper, to be published in the Summer, will set out how the untapped talent of communities can be unleashed to ensure everyone has a greater say in improvements to public services, local accountability and opportunities for enterprise’

Of the three elements of empowerment mentioned here, two are about improving democracy, and one is about entrepreneurship. However, in the ‘Unlocking talent’ discussion paper, five pages are devoted to reducing worklessness, while about one and a half cover the public services and accountability. I’m not quite sure why getting a job empowers you to be involved in participatory democracy, especially as it reduces the time you have for ‘getting involved’. Of course, getting a job does empower you by reducing your reliance on the state, and empowers you as a consumer, but I didn’t think that was the point.

Anyway, now for the bad social science. Part of their ‘evidence shows that those in employment are happier, healthier and less likely to be involved in crime’ (DCLG 2008, p.4). The truth, of course, is far more complicated.

Firstly, this short and authoritative statement is based on a number of papers (Strategy Unit 2002, Meghir and Machin 2000 and one other) that don’t really support it.

The paper (M&M) on crime showed that ‘falls in the wages of low-wage workers lead to increases in crime’, suggesting we should increase the minimum wage!

Early on, the Strategy unit paper reminds us that correlations don’t tell us the direction of causation. Given that an employer is looking for people who are fun, or at least nice to work with (I always did this), miserable people are less likely to get jobs. I can’t think of many jobs where being pessimistic is an asset.

And we should always be on the look out for spurious correlations, and in this research they are legion. The data this paper uses, suggests that workers have greater life-satifaction than non-workers. But of course, in an unequal world the workers have more money than the non-workers, they socialise more with their colleagues (’cause if you’re unemployed while all your mates are working, you’ve no-one to go out with), and they aren’t bored stuck at home with no money.

In order to research this properly, we should do either a randomised controlled trial, but my ethics committee would turn this down, or some kind of matched survey.  In the latter, instead of comparing a sample of workers with non-workers, the research would compare like-with-like. Each unemployed person would have as much money, activity and lifestyle (and so on) as the comparator employed person. Let’s see if the middle-class, middle-aged retiree on a £50k pension aged 50, is happier or less happy than his/her equivalent who still has to work to sustain the lifestyle. Or let’s compare the young minimum wage cleaner (less than £12k), with someone who get’s £12k from a trust fund. This might give a result that is consistent with the fact that almost no-one who wins the lottery carries on working full-time (NS 2005). In fact, lottery winners are happier as they ‘do what they like’ and enjoy an ‘easier life’.

Leave a comment

Filed under economy, News, Statistics and simplicity