Don’t worry, Bolsover: the problem is not you but how happiness statistics are understood

The Office for National Statistics released their latest report last week on National Well-Being. They showed many interesting things – for example, that since this measurement series began in 2011, each of the four measures they track (satisfaction, happiness, feeling life is worthwhile, and anxiety) have shown improvement on average across the UK. The ONS quite rightly point out that this is likely to result from improvements to the economic situation since then. Things aren’t perfect now, but unemployment has come down from its 2012 peak and household income is up at about where it was pre-recession.

While national averages are all well and good, the ONS also provide a very detailed breakdown by geographical area within the UK. Everyone seems to enjoy this type of data – news items quickly followed on the best and worst areas of the UK to live in. Particularly striking perhaps are stories on the happiest place in the UK (Fermanagh and Omagh) versus the most miserable (Bolsover), or similarly on the best and worst places to live in London. People were invited onto news programmes to explain why their areas are so happy or miserable (see here, for example, about 38 minutes in) – economic equality, festivals, beautiful scenery are given as reasons for the high average; an expression of surprise was the response from the Bolsover resident. What are you supposed to do when faced with such evidence that the place you live in is wonderful utopia or a miserable hovel? My advice is not to be smug (happy places) and not to set about defending your home town (unhappy places), but rather: ignore it.

There are two main reasons to be wary of this kind of statistic. The first is statistical significance. I took a look at the 95% confidence intervals that the ONS provided and after a quick sorting exercise was left with a long list of places that have a happiness level the same (or at least not different, statistically speaking) as Bolsover. The list includes some nice-sounding places like mid-Suffolk and the Forest of Dean. Similarly, looking at the confidence intervals shows that joy is not confined to Fermanagh and Omagh – there are places equal in happiness, statistically speaking, scattered all over the country including Ribble Valley, Fylde, areas of Devon, Dorset, Oxfordshire and Scottish Islands (to name but a few – See Fig 1).

Average happiness, with confidence intervals, in several happy places in the UK. Author’s illustration based on ONS data.

The second reason is compositional effects. The scientific evidence on the causes of happiness confirms in many cases what people already know: that it is no fun being ill, being unemployed or living in poverty, for example. If we look at an area with high unemployment and poverty levels, or more older people (who are more likely to be ill) therefore, it will not be surprising if happiness there was a little lower than average. It is a question of causality. If you move to Bolsover you won’t suddenly become miserable, nor will moving to Fermanagh miraculously lift all your worries.

This is not to say that well-being measures are not important at all; people who are unhappy and the places they live in deserve the attention of our politicians and policy makers. But let’s not get too excited about these news stories about happy places – just because they are based on ‘statistics’ it doesn’t mean we should take them at face value without some awareness of what the numbers really mean (which might not be a lot).


Post by Dr Lindsay Richards